1 Guile NEWS --- history of user-visible changes.
2 Copyright (C) 1996-2014 Free Software Foundation, Inc.
3 See the end for copying conditions.
5 Please send Guile bug reports to bug-guile@gnu.org.
9 Changes in 2.1.1 (changes since the 2.0.x series):
15 The biggest change in Guile 2.2 is a complete rewrite of its virtual
16 machine and compiler internals. The result is faster startup time,
17 better memory usage, and faster execution of user code. See the
18 "Performance improvements" section below for more details.
20 ** Better thread-safety
22 This new release series takes the ABI-break opportunity to fix some
23 interfaces that were difficult to use correctly from multiple threads.
24 Notably, weak hash tables are now transparently thread-safe. Ports are
25 also thread-safe; see "New interfaces" below for details on the changes
28 ** Better space-safety
30 It used to be the case that, when calling a Scheme procedure, the
31 procedure and arguments were always preserved against garbage
32 collection. This is no longer the case; Guile is free to collect the
33 procedure and arguments if they become unreachable, or to re-use their
34 slots for other local variables. Guile still offers good-quality
35 backtraces by determining the procedure being called from the
36 instruction pointer instead of from the value in slot 0 of an
37 application frame, and by using a live variable map that allows the
38 debugger to know which locals are live at all points in a frame.
40 ** Off-main-thread finalization
42 Following Guile 2.0.6's change to invoke finalizers via asyncs, Guile
43 2.2 takes the additional step of invoking finalizers from a dedicated
44 finalizer thread, if threads are enabled. This avoids concurrency
45 issues between finalizers and application code, and also speeds up
46 finalization. If your application's finalizers are not robust to the
47 presence of threads, see "Foreign Objects" in the manual for information
48 on how to disable automatic finalization and instead run finalizers
51 ** Better locale support in Guile scripts
53 When Guile is invoked directly, either from the command line or via a
54 hash-bang line (e.g. "#!/usr/bin/guile"), it now installs the current
55 locale via a call to `(setlocale LC_ALL "")'. For users with a unicode
56 locale, this makes all ports unicode-capable by default, without the
57 need to call `setlocale' in your program. This behavior may be
58 controlled via the GUILE_INSTALL_LOCALE environment variable; see the
61 ** Complete Emacs-compatible Elisp implementation
63 Thanks to the work of BT Templeton, Guile's Elisp implementation is now
64 fully Emacs-compatible, implementing all of Elisp's features and quirks
65 in the same way as the editor we know and love.
67 ** Dynamically expandable stacks
69 Instead of allocating fixed stack sizes for running Scheme code, Guile
70 now starts off each thread with only one page of stack, and expands and
71 shrinks it dynamically as needed. Guile will throw an exception for
72 stack overflows if growing the stack fails. It is also possible to
73 impose a stack limit during the extent of a function call. See "Stack
74 Overflow" in the manual, for more.
76 This change allows users to write programs that use the stack as a data
77 structure for pending computations, as it was meant to be, without
78 reifying that data out to the heap. Where you would previously make a
79 loop that collect its results in reverse order only to re-reverse them
80 at the end, now you can just recurse without worrying about stack
83 ** Out-of-memory improvements
85 Instead of aborting, failures to allocate memory will now raise an
86 unwind-only `out-of-memory' exception, and cause the corresponding
87 `catch' expression to run garbage collection in order to free up memory.
89 * Performance improvements
91 ** Faster programs via new virtual machine
93 Guile's new virtual machine compiles programs to instructions for a new
94 virtual machine. The new virtual machine's instructions can address
95 their source and destination operands by "name" (slot). This makes
96 access to named temporary values much faster, and removes a lot of
97 value-shuffling that the old virtual machine had to do. The end result
98 is that loop-heavy code can be two or three times as fast with Guile 2.2
99 as in 2.0. Your mileage may vary, of course; see "A Virtual Machine for
100 Guile" in the manual for the nitties and the gritties.
102 ** Better startup time, memory usage with ELF object file format
104 Guile now uses the standard ELF format for its compiled code. (Guile
105 has its own loader and linker, so this does not imply a dependency on
106 any particular platform's ELF toolchain.) The benefit is that Guile is
107 now able to statically allocate more data in the object files. ELF also
108 enables more sharing of data between processes, and decreases startup
109 time (about 40% faster than the already fast startup of the Guile 2.0
110 series). Guile also uses DWARF for some of its debugging information.
111 Much of the debugging information can be stripped from the object files
112 as well. See "Object File Format" in the manual, for full details.
114 ** Better optimizations via compiler rewrite
116 Guile's compiler now uses a Continuation-Passing Style (CPS)
117 intermediate language, allowing it to reason easily about temporary
118 values and control flow. Examples of optimizations that this permits
119 are optimal contification, optimal common subexpression elimination,
120 dead code elimination, parallel moves with at most one temporary,
121 allocation of stack slots using precise liveness information, and
122 closure optimization. For more, see "Continuation-Passing Style" in the
125 ** Faster interpreter
127 Combined with a number of optimizations to the interpreter itself,
128 simply compiling `eval.scm' with the new compiler yields an interpreter
129 that is consistently two or three times faster than the one in Guile
132 ** Allocation-free dynamic stack
134 Guile now implements the dynamic stack with an actual stack instead of a
135 list of heap objects, avoiding most allocation. This speeds up prompts,
136 the `scm_dynwind_*' family of functions, fluids, and `dynamic-wind'.
138 ** Optimized UTF-8 and Latin-1 ports, symbols, and strings
140 Guile 2.2 is faster at reading and writing UTF-8 and Latin-1 strings
141 from ports, and at converting symbols and strings to and from these
144 ** Optimized hash functions
146 Guile 2.2 now uses Bob Jenkins' `hashword2' (from his `lookup3.c') for
147 its string hash, and Thomas Wang's integer hash function for `hashq' and
148 `hashv'. These functions produce much better hash values across all
149 available fixnum bits.
151 ** Optimized generic array facility
153 Thanks to work by Daniel Llorens, the generic array facility is much
154 faster now, as it is internally better able to dispatch on the type of
155 the underlying backing store.
159 ** New `cond-expand' feature: `guile-2.2'
161 Use this feature if you need to check for Guile 2.2 from Scheme code.
163 ** New predicate: `nil?'
165 See "Nil" in the manual.
167 ** New compiler modules
169 Since the compiler was rewritten, there are new modules for the back-end
170 of the compiler and the low-level loader and introspection interfaces.
171 See the "Guile Implementation" chapter in the manual for all details.
173 ** New functions: `scm_to_intptr_t', `scm_from_intptr_t'
174 ** New functions: `scm_to_uintptr_t', `scm_from_uintptr_t'
176 See "Integers" in the manual, for more.
178 ** New thread-safe port API
180 For details on `scm_c_make_port', `scm_c_make_port_with_encoding',
181 `scm_c_lock_port', `scm_c_try_lock_port', `scm_c_unlock_port',
182 `scm_c_port_type_ref', `scm_c_port_type_add_x', `SCM_PORT_DESCRIPTOR',
183 and `scm_dynwind_lock_port', see XXX.
185 There is now a routine to atomically adjust port "revealed counts". See
186 XXX for more on `scm_adjust_port_revealed_x' and
187 `adjust-port-revealed!',
189 All other port API now takes the lock on the port if needed. There are
190 some C interfaces if you know that you don't need to take a lock; see
191 XXX for details on `scm_get_byte_or_eof_unlocked',
192 `scm_peek_byte_or_eof_unlocked' `scm_c_read_unlocked',
193 `scm_getc_unlocked' `scm_unget_byte_unlocked', `scm_ungetc_unlocked',
194 `scm_ungets_unlocked', `scm_fill_input_unlocked' `scm_putc_unlocked',
195 `scm_puts_unlocked', and `scm_lfwrite_unlocked'.
197 ** New inline functions: `scm_new_smob', `scm_new_double_smob'
199 These can replace many uses of SCM_NEWSMOB, SCM_RETURN_NEWSMOB2, and the
200 like. See XXX in the manual, for more.
202 ** New low-level type accessors
204 For more on `SCM_HAS_TYP7', `SCM_HAS_TYP7S', `SCM_HAS_TYP16', see XXX.
206 `SCM_HEAP_OBJECT_P' is now an alias for the inscrutable `SCM_NIMP'.
208 `SCM_UNPACK_POINTER' and `SCM_PACK_POINTER' are better-named versions of
209 the old `SCM2PTR' and `PTR2SCM'. Also, `SCM_UNPACK_POINTER' yields a
212 ** <standard-vtable>, standard-vtable-fields
214 See "Structures" in the manual for more on these
216 ** Convenience utilities for ports and strings.
218 See XXX for more on `scm_from_port_string', `scm_from_port_stringn',
219 `scm_to_port_string', and `scm_to_port_stringn'.
221 ** New expressive PEG parser
223 See "PEG Parsing" in the manual for more. Thanks to Michael Lucy for
224 originally writing these, and to Noah Lavine for integration work.
226 * Incompatible changes
228 ** ASCII is not ISO-8859-1
230 In Guile 2.0, if a user set "ASCII" or "ANSI_X3.4-1968" as the encoding
231 of a port, Guile would treat it as ISO-8859-1. While these encodings
232 are the same for codepoints 0 to 127, ASCII does not extend past that
233 range, whereas ISO-8859-1 goes up to 255. Guile 2.2 no longer treats
234 ASCII as ISO-8859-1. This is likely to be a problem only if the user's
235 locale is set to ASCII, and the user or a program writes non-ASCII
236 codepoints to a port.
238 ** String ports default to UTF-8
240 Guile 2.0 would use the `%default-port-encoding' when creating string
241 ports. This resulted in ports that could only accept a subset of valid
242 characters, which was surprising to users. Now string ports default to
243 the UTF-8 encoding. Sneaky users can still play encoding conversion
244 games with string ports by explicitly setting the encoding of a port
245 after it is open. See "Ports" in the manual for more.
247 ** `scm_from_stringn' and `scm_to_stringn' encoding arguments are never NULL
249 These functions now require a valid `encoding' argument, and will abort
252 ** All r6rs ports are both textual and binary
254 Because R6RS ports are a thin layer on top of Guile's ports, and Guile's
255 ports are both textual and binary, Guile's R6RS ports are also both
256 textual and binary, and thus both kinds have port transcoders. This is
257 an incompatibility with respect to R6RS.
259 ** Vtable hierarchy changes
261 In an attempt to make Guile's structure and record types integrate
262 better with GOOPS by unifying the vtable hierarchy, `make-vtable-vtable'
263 is now deprecated. Instead, users should just use `make-vtable' with
264 appropriate arguments. See "Structures" in the manual for all of the
265 details. As such, `record-type-vtable' and `%condition-type-vtable' now
266 have a parent vtable and are no longer roots of the vtable hierarchy.
268 ** Syntax parameters are a distinct type
270 Guile 2.0's transitional implementation of `syntax-parameterize' was
271 based on the `fluid-let-syntax' interface inherited from the psyntax
272 expander. This interface allowed any binding to be dynamically rebound
273 -- even bindings like `lambda'. This is no longer the case in Guile
274 2.2. Syntax parameters must be defined via `define-syntax-parameter',
275 and only such bindings may be parameterized. See "Syntax Parameters" in
278 ** Defined identifiers scoped in the current module
280 Sometimes Guile's expander would attach incorrect module scoping
281 information for top-level bindings made by an expansion. For example,
282 given the following R6RS library:
287 (define-syntax-rule (defconst name val)
290 (define-syntax-rule (name) t))))
292 Attempting to use it would produce an error:
297 =| Unbound variable: t
299 It wasn't clear that we could fix this in Guile 2.0 without breaking
300 someone's delicate macros, so the fix is only coming out now.
302 ** Pseudo-hygienically rename macro-introduced bindings
304 Bindings introduced by macros, like `t' in the `defconst' example above,
305 are now given pseudo-fresh names. This allows
310 to introduce different bindings for `t'. These pseudo-fresh names are
311 made in such a way that if the macro is expanded again, for example as
312 part of a simple recompilation, the introduced identifiers get the same
313 pseudo-fresh names. See "Hygiene and the Top-Level" in the manual, for
316 ** Fix literal matching for module-bound literals
318 `syntax-rules' and `syntax-case' macros can take a set of "literals":
319 bound or unbound keywords that the syntax matcher treats specially.
320 Before, literals were always matched symbolically (by name). Now they
321 are matched by binding. This allows literals to be reliably bound to
322 values, renamed by imports or exports, et cetera. See "Syntax-rules
323 Macros" in the manual for more on literals.
325 ** `dynamic-wind' doesn't check that guards are thunks
327 Checking that the dynamic-wind out-guard procedure was actually a thunk
328 before doing the wind was slow, unreliable, and not strictly needed.
330 ** All deprecated code removed
332 All code deprecated in Guile 2.0 has been removed. See older NEWS, and
333 check that your programs can compile without linker warnings and run
334 without runtime warnings. See "Deprecation" in the manual.
336 ** Remove miscellaneous unused interfaces
338 We have removed accidentally public, undocumented interfaces that we
339 think are not used, and not useful. This includes `scm_markstream',
340 `SCM_FLUSH_REGISTER_WINDOWS', `SCM_THREAD_SWITCHING_CODE', `SCM_FENCE',
341 `scm_call_generic_0', `scm_call_generic_1', `scm_call_generic_2'
342 `scm_call_generic_3', `scm_apply_generic', and `scm_program_source'.
343 `scm_async_click' was renamed to `scm_async_tick', and `SCM_ASYNC_TICK'
344 was made private (use `SCM_TICK' instead).
346 ** Many internal compiler / VM changes
348 As the compiler and virtual machine were re-written, there are many
349 changes in the back-end of Guile to interfaces that were introduced in
350 Guile 2.0. These changes are only only of interest if you wrote a
351 language on Guile 2.0 or a tool using Guile 2.0 internals. If this is
352 the case, drop by the IRC channel to discuss the changes.
354 ** Defining a SMOB or port type no longer mucks exports of `(oop goops)'
356 It used to be that defining a SMOB or port type added an export to
357 GOOPS, for the wrapper class of the smob type. This violated
358 modularity, though, so we have removed this behavior.
360 ** Bytecode replaces objcode as a target language
362 One way in which people may have used details of Guile's runtime in
363 Guile 2.0 is in compiling code to thunks for later invocation. Instead
364 of compiling to objcode and then calling `make-program', now the way to
365 do it is to compile to `bytecode' and then call `load-thunk-from-memory'
366 from `(system vm loader)'.
368 ** Weak pairs removed
370 Weak pairs were not safe to access with `car' and `cdr', and so were
373 ** Weak alist vectors removed
375 Use weak hash tables instead.
377 ** Weak vectors may no longer be accessed via `vector-ref' et al
379 Weak vectors may no longer be accessed with the vector interface. This
380 was a source of bugs in the 2.0 Guile implementation, and a limitation
381 on using vectors as building blocks for other abstractions. Vectors in
382 Guile are now a concrete type; for an abstract interface, use the
383 generic array facility (`array-ref' et al).
385 ** scm_t_array_implementation removed
387 This interface was introduced in 2.0 but never documented. It was a
388 failed attempt to layer the array implementation that actually
389 introduced too many layers, as it prevented the "vref" and "vset"
390 members of scm_t_array_handle (called "ref" and "set" in 1.8, not
391 present in 2.0) from specializing on array backing stores.
393 Notably, the definition of scm_t_array_handle has now changed, to not
394 include the (undocumented) "impl" member. We are sorry for any
395 inconvenience this may cause.
399 ** SCM_WTA_DISPATCH_0, SCM_WTA_DISPATCH_1, SCM_WTA_DISPATCH_2, SCM_WTA_DISPATCH_N
400 ** SCM_GASSERT0, SCM_GASSERT1, SCM_GASSERT2, SCM_GASSERTn
401 ** SCM_WTA_DISPATCH_1_SUBR
403 These macros were used in dispatching primitive generics. They can be
404 replaced by using C functions (the same name but in lower case), if
405 needed, but this is a hairy part of Guile that perhaps you shouldn't be
408 * Changes to the distribution
412 The "effective version" of Guile is now 2.2, which allows parallel
413 installation with other effective versions (for example, the older Guile
414 2.0). See "Parallel Installations" in the manual for full details.
415 Notably, the `pkg-config' file is now `guile-2.2'.
417 ** Bump required libgc version to 7.2, released March 2012.
419 ** The readline extension is now installed in the extensionsdir
421 The shared library that implements Guile's readline extension is no
422 longer installed to the libdir. This change should be transparent to
423 users, but packagers may be interested.
427 Changes in 2.0.11 (since 2.0.10):
429 This release fixes an embarrassing regression introduced in the C
430 interface to SRFI-4 vectors. See
431 <https://lists.gnu.org/archive/html/guile-devel/2014-03/msg00047.html>
435 Changes in 2.0.10 (since 2.0.9):
439 ** New GDB extension to support Guile
441 Guile now comes with an extension for GDB 7.8 or later (unreleased at
442 the time of writing) that simplifies debugging of C code that uses
443 Guile. See "GDB Support" in the manual.
445 ** Improved integration between R6RS and native Guile exceptions
447 R6RS exception handlers, established using 'with-exception-handler' or
448 'guard', are now able to catch native Guile exceptions, which are
449 automatically converted into appropriate R6RS condition objects.
451 ** Support for HTTP proxies
453 Guile's built-in web client now honors the 'http_proxy' environment
454 variable, as well as the new 'current-http-proxy' parameter. See
455 "Web Client" in the manual for details.
457 ** Lexical syntax improvements
459 *** Support |...| symbol notation.
461 Guile's core reader and printer now support the R7RS |...| notation
462 for writing symbols with arbitrary characters, as a more portable and
463 attractive alternative to Guile's native #{...}# notation. To enable
464 this notation by default, put one or both of the following in your
467 (read-enable 'r7rs-symbols)
468 (print-enable 'r7rs-symbols)
470 *** Support '#true' and '#false' notation for booleans.
472 The booleans '#t' and '#f' may now be written as '#true' and '#false'
473 for improved readability, per R7RS.
475 *** Recognize '#\escape' character name.
477 The escape character '#\esc' may now be written as '#\escape', per R7RS.
479 *** Accept "\|" in string literals.
481 The pipe character may now be preceded by a backslash, per R7RS.
483 ** Custom binary input ports now support 'setvbuf'.
485 Until now, ports returned by 'make-custom-binary-input-port' were always
486 full-buffered. Now, their buffering mode can be changed using 'setvbuf'.
488 ** SRFI-4 predicates and length accessors no longer accept arrays.
490 Given that the SRFI-4 accessors don't work for arrays, the fact that the
491 predicates and length accessors returned true for arrays was a bug.
493 ** GUILE_PROGS now supports specifying a minimum required version.
495 The 'GUILE_PROGS' autoconf macro in guile.m4 now allows an optional
496 argument to specify a minimum required Guile version. By default, it
497 requires Guile >= 2.0. A micro version can also be specified, e.g.:
498 GUILE_PROGS([2.0.10])
500 ** Error reporting improvements
502 *** Improved run-time error reporting in (ice-9 match).
504 If no pattern matches in a 'match' form, the datum that failed to match
505 is printed along with the location of the failed 'match' invocation.
507 *** Print the faulty object upon invalid-keyword errors.
508 *** Improved error reporting of procedures defined by define-inlinable.
509 *** Improved error reporting for misplaced ellipses in macro definitions.
510 *** Improved error checking in 'define-public' and 'module-add!'.
511 *** Improved error when 'include' form with relative path is not in a file.
513 ** Speed improvements
515 *** 'scm_c_read' on ISO-8859-1 (e.g. binary) unbuffered ports is faster.
516 *** New inline asm for VM fixnum multiply, for faster overflow checking.
517 *** New inline asm for VM fixnum operations on ARM and 32-bit x86.
518 *** 'positive?' and 'negative?' are now compiled to VM primitives.
519 *** Numerical comparisons with more than 2 arguments are compiled to VM code.
520 *** Several R6RS bitwise operators have been optimized.
524 *** Web: 'content-disposition' headers are now supported.
525 *** Web: 'uri-encode' hexadecimal percent-encoding is now uppercase.
526 *** Size argument to 'make-doubly-weak-hash-table' is now optional.
527 *** Timeout for 'unlock-mutex' and SRFI-18 'mutex-unlock!' may now be #f.
531 Guile's copy of Gnulib was updated to v0.1-92-g546ff82. The following
532 modules were imported from Gnulib: copysign, fsync, isfinite, link,
533 lstat, mkdir, mkstemp, readlink, rename, rmdir, and unistd.
537 ** Cooperative REPL servers
539 This new facility supports REPLs that run at specified times within an
540 existing thread, for example in programs utilizing an event loop or in
541 single-threaded programs. This allows for safe access and mutation of
542 a program's data structures from the REPL without concern for thread
543 synchronization. See "Cooperative REPL Servers" in the manual for
546 ** SRFI-43 (Vector Library)
548 Guile now includes SRFI-43, a comprehensive library of vector operations
549 analogous to the SRFI-1 list library. See "SRFI-43" in the manual for
552 ** SRFI-64 (A Scheme API for test suites)
554 Guile now includes SRFI-64, a flexible framework for creating test
555 suites. The reference implementation of SRFI-64 has also been updated
556 to fully support earlier versions of Guile.
560 See "SRFI-111" in the manual.
564 See "Binding multiple return values" in the manual.
566 ** Custom ellipsis identifiers using 'with-ellipsis' or SRFI-46.
568 Guile now allows macro definitions to use identifiers other than '...'
569 as the ellipsis. This is convenient when writing macros that generate
570 macro definitions. The desired ellipsis identifier can be given as the
571 first operand to 'syntax-rules', as specified in SRFI-46 and R7RS, or by
572 using the new 'with-ellipsis' special form in procedural macros. With
573 this addition, Guile now fully supports SRFI-46.
575 See "Specifying a Custom Ellipsis Identifier" and "Custom Ellipsis
576 Identifiers for syntax-case Macros" in the manual for details.
578 ** R7RS 'syntax-error'
580 Guile now supports 'syntax-error', as specified by R7RS, allowing for
581 improved compile-time error reporting from 'syntax-rules' macros. See
582 "Reporting Syntax Errors in Macros" in the manual for details.
584 ** New procedures to convert association lists into hash tables
586 Guile now includes the convenience procedures 'alist->hash-table',
587 'alist->hashq-table', 'alist->hashv-table', and 'alist->hashx-table'.
588 See "Hash Table Reference" in the manual.
590 ** New predicates: 'exact-integer?' and 'scm_is_exact_integer'
592 See "Integers" in the manual.
594 ** 'weak-vector-length', 'weak-vector-ref', and 'weak-vector-set!'
596 These should now be used to access weak vectors, instead of
597 'vector-length', 'vector-ref', and 'vector-set!'.
601 ** Improve docs for 'eval-when'.
603 Each 'eval-when' condition is now explained in detail, including
604 'expand' which was previously undocumented. (expand load eval) is now
605 the recommended set of conditions, instead of (compile load eval).
606 See "Eval When" in the manual, for details.
608 ** Update the section on SMOBs and memory management.
610 See "Defining New Types (Smobs)" in the manual.
614 *** GOOPS: #:dsupers is the init keyword for the dsupers slot.
615 *** 'unfold-right' takes a tail, not a tail generator.
616 *** Clarify that 'append!' and 'reverse!' might not mutate.
617 *** Fix doc that incorrectly claimed (integer? +inf.0) => #t.
618 (http://bugs.gnu.org/16356)
619 *** Document that we support SRFI-62 (S-expression comments).
620 *** Document that we support SRFI-87 (=> in case clauses).
621 *** Document 'equal?' in the list of R6RS incompatibilities.
622 *** Remove outdated documentation of LTDL_LIBRARY_PATH.
623 *** Fix 'weak-vector?' doc: Weak hash tables are not weak vectors.
624 *** Fix 'my-or' examples to use let-bound variable.
625 (http://bugs.gnu.org/14203)
629 ** General 'uniform-vector' interface
631 This interface lacked both generality and specificity. The general
632 replacements are 'array-length', 'array-ref', and friends on the scheme
633 side, and the array handle interface on the C side. On the specific
634 side of things, there are the specific bytevector, SRFI-4, and bitvector
637 ** Use of the vector interface on arrays
638 ** 'vector-length', 'vector-ref', and 'vector-set!' on weak vectors
639 ** 'vector-length', 'vector-ref', and 'vector-set!' as primitive-generics
641 Making the vector interface operate only on a single representation will
642 allow future versions of Guile to compile loops involving vectors to
643 more efficient native code.
645 ** 'htons', 'htonl', 'ntohs', 'ntohl'
647 These procedures, like their C counterpart, were used to convert numbers
648 to/from network byte order, typically in conjunction with the
649 now-deprecated uniform vector API.
651 This functionality is now covered by the bytevector and binary I/O APIs.
652 See "Interpreting Bytevector Contents as Integers" in the manual.
654 ** 'gc-live-object-stats'
656 It hasn't worked in the whole 2.0 series. There is no replacement,
659 ** 'scm_c_program_source'
661 This internal VM function was not meant to be public. Use
662 'scm_procedure_source' instead.
666 ** Fix build with Clang 3.4.
669 *** Do not add $(EXEEXT) to guild or guile-tools.
670 *** tests: Use double quotes around shell arguments, for Windows.
671 *** tests: Don't rely on $TMPDIR and /tmp on Windows.
672 *** tests: Skip FFI tests that use `qsort' when it's not accessible.
673 *** tests: Remove symlink only when it exists.
674 *** tests: Don't rely on `scm_call_2' being visible.
676 ** Fix computation of LIBLOBJS so dependencies work properly.
677 (http://bugs.gnu.org/14193)
681 ** Web: Fix web client with methods other than GET.
682 (http://bugs.gnu.org/15908)
683 ** Web: Add Content-Length header for empty bodies.
684 ** Web: Accept "UTC" as the zone offset in date headers.
685 (http://bugs.gnu.org/14128)
686 ** Web: Don't throw if a response is longer than its Content-Length says.
687 ** Web: Write out HTTP Basic auth headers correctly.
688 (http://bugs.gnu.org/14370)
689 ** Web: Always print a path component in 'write-request-line'.
690 ** Fix 'define-public' from (ice-9 curried-definitions).
691 ** psyntax: toplevel variable definitions discard previous syntactic binding.
692 (http://bugs.gnu.org/11988)
693 ** Fix thread-unsafe lazy initializations.
694 ** Make (ice-9 popen) thread-safe.
695 (http://bugs.gnu.org/15683)
696 ** Make guardians thread-safe.
697 ** Make regexp_exec thread-safe.
698 (http://bugs.gnu.org/14404)
699 ** vm: Gracefully handle stack overflows.
700 (http://bugs.gnu.org/15065)
701 ** Fix 'rationalize'.
702 (http://bugs.gnu.org/14905)
703 ** Fix inline asm for VM fixnum operations on x32.
704 ** Fix 'SCM_SYSCALL' to really swallow EINTR.
705 ** Hide EINTR returns from 'accept'.
706 ** SRFI-19: Update the table of leap seconds.
707 ** Add missing files to the test-suite Makefile.
708 ** Make sure 'ftw' allows directory traversal when running as root.
709 ** Fix 'hash-for-each' for weak hash tables.
710 ** SRFI-18: Export 'current-thread'.
711 (http://bugs.gnu.org/16890)
712 ** Fix inlining of tail list to apply.
713 (http://bugs.gnu.org/15533)
714 ** Fix bug in remqueue in threads.c when removing last element.
715 ** Fix build when '>>' on negative integers is not arithmetic.
716 ** Fix 'bitwise-bit-count' for negative arguments.
717 (http://bugs.gnu.org/14864)
718 ** Fix VM 'ash' for right shifts by large amounts.
719 (http://bugs.gnu.org/14864)
720 ** Fix rounding in scm_i_divide2double for negative arguments.
721 ** Avoid lossy conversion from inum to double in numerical comparisons.
722 ** Fix numerical comparison of fractions to infinities.
723 ** Allow fl+ and fl* to accept zero arguments.
724 (http://bugs.gnu.org/14869)
725 ** flonum? returns false for complex number objects.
726 (http://bugs.gnu.org/14866)
727 ** flfinite? applied to a NaN returns false.
728 (http://bugs.gnu.org/14868)
729 ** Flonum operations always return flonums.
730 (http://bugs.gnu.org/14871)
731 ** min and max: NaNs beat infinities, per R6RS errata.
732 (http://bugs.gnu.org/14865)
733 ** Fix 'fxbit-count' for negative arguments.
734 ** 'gcd' and 'lcm' support inexact integer arguments.
735 (http://bugs.gnu.org/14870)
736 ** Fix R6RS 'fixnum-width'.
737 (http://bugs.gnu.org/14879)
738 ** tests: Use shell constructs that /bin/sh on Solaris 10 can understand.
739 (http://bugs.gnu.org/14042)
740 ** Fix display of symbols containing backslashes.
741 (http://bugs.gnu.org/15033)
742 ** Fix truncated-print for uniform vectors.
743 ** Define `AF_UNIX' only when Unix-domain sockets are supported.
744 ** Decompiler: fix handling of empty 'case-lambda' expressions.
745 ** Fix handling of signed zeroes and infinities in 'numerator' and 'denominator'.
746 ** dereference-pointer: check for null pointer.
747 ** Optimizer: Numerical comparisons are not negatable, for correct NaN handling.
748 ** Compiler: Evaluate '-' and '/' in left-to-right order.
749 (for more robust floating-point arithmetic)
750 ** snarf.h: Declare static const function name vars as SCM_UNUSED.
751 ** chars.c: Remove duplicate 'const' specifiers.
752 ** Modify SCM_UNPACK type check to avoid warnings in clang.
753 ** Arrange so that 'file-encoding' does not truncate the encoding name.
754 (http://bugs.gnu.org/16463)
755 ** Improve error checking in bytevector->uint-list and bytevector->sint-list.
756 (http://bugs.gnu.org/15100)
757 ** Fix (ash -1 SCM_I_FIXNUM_BIT-1) to return a fixnum instead of a bignum.
758 ** i18n: Fix null pointer dereference when locale info is missing.
759 ** Fix 'string-copy!' to work properly with overlapping src/dest.
760 ** Fix hashing of vectors to run in bounded time.
761 ** 'port-position' works on CBIPs that do not support 'set-port-position!'.
762 ** Custom binary input ports sanity-check the return value of 'read!'.
763 ** bdw-gc.h: Check SCM_USE_PTHREAD_THREADS using #if not #ifdef.
764 ** REPL Server: Don't establish a SIGINT handler.
765 ** REPL Server: Redirect warnings to client socket.
766 ** REPL Server: Improve robustness of 'stop-server-and-clients!'.
767 ** Add srfi-16, srfi-30, srfi-46, srfi-62, srfi-87 to %cond-expand-features.
768 ** Fix trap handlers to handle applicable structs.
769 (http://bugs.gnu.org/15691)
770 ** Fix optional end argument in `uniform-vector-read!'.
771 (http://bugs.gnu.org/15370)
772 ** Fix brainfuck->scheme compiler.
773 ** texinfo: Fix newline preservation in @example with lines beginning with @
775 ** C standards conformance improvements
777 Improvements and bug fixes were made to the C part of Guile's run-time
780 *** Don't use the identifier 'noreturn'.
781 (http://bugs.gnu.org/15798)
782 *** Rewrite SCM_I_INUM to avoid unspecified behavior when not using GNU C.
783 *** Improve fallback implemention of SCM_SRS to avoid unspecified behavior.
784 *** SRFI-60: Reimplement 'rotate-bit-field' on inums to be more portable.
785 *** Improve compliance with C standards regarding signed integer shifts.
786 *** Avoid signed overflow in random.c.
787 *** VM: Avoid signed overflows in 'add1' and 'sub1'.
788 *** VM: Avoid overflow in ASM_ADD when the result is most-positive-fixnum.
789 *** read: Avoid signed integer overflow in 'read_decimal_integer'.
793 Changes in 2.0.9 (since 2.0.7):
795 Note: 2.0.8 was a brown paper bag release that was never announced, but
796 some mirrors may have picked it up. Please do not use it.
800 ** New keyword arguments for procedures that open files
802 The following procedures that open files now support keyword arguments
803 to request binary I/O or to specify the character encoding for text
804 files: `open-file', `open-input-file', `open-output-file',
805 `call-with-input-file', `call-with-output-file', `with-input-from-file',
806 `with-output-to-file', and `with-error-to-file'.
808 It is also now possible to specify whether Guile should scan files for
809 Emacs-style coding declarations. This scan was done by default in
810 versions 2.0.0 through 2.0.7, but now must be explicitly requested.
812 See "File Ports" in the manual for details.
814 ** Rewritten guile.m4
816 The `guile.m4' autoconf macros have been rewritten to use `guild' and
817 `pkg-config' instead of the deprecated `guile-config' (which itself
820 There is also a new macro, `GUILE_PKG', which allows packages to select
821 the version of Guile that they want to compile against. See "Autoconf
822 Macros" in the manual, for more information.
824 ** Better Windows support
826 Guile now correctly identifies absolute paths on Windows (MinGW), and
827 creates files on that platform according to its path conventions. See
828 "File System" in the manual, for all details.
830 In addition, the new Gnulib imports provide `select' and `poll' on
833 As an incompatible change, systems that are missing <sys/select.h> were
834 previously provided a public `scm_std_select' C function that defined a
835 version of `select', but unhappily it also provided its own incompatible
836 definitions for FD_SET, FD_ZERO, and other system interfaces. Guile
837 should not be setting these macros in public API, so this interface was
838 removed on those plaforms (basically only MinGW).
840 ** Numerics improvements
842 `number->string' now reliably outputs enough digits to produce the same
843 number when read back in. Previously, it mishandled subnormal numbers
844 (printing them as "#.#"), and failed to distinguish between some
845 distinct inexact numbers, e.g. 1.0 and (+ 1.0 (expt 2.0 -52)). These
846 problems had far-reaching implications, since the compiler uses
847 `number->string' to serialize numeric constants into .go files.
849 `sqrt' now produces exact rational results when possible, and handles
850 very large or very small numbers more robustly.
852 A number (ahem) of operations involving exact rationals have been
853 optimized, most notably `integer-expt' and `expt'.
855 `exact->inexact' now performs correct IEEE rounding.
859 There were a number of improvements to the partial evaluator, allowing
860 complete reduction of forms such as:
862 ((let ((_ 10)) (lambda () _)))
866 (apply (lambda _ _) 1 2 3 '(4))
868 (call-with-values (lambda () (values 1 2)) (lambda _ _))
870 `string-join' now handles huge lists efficiently.
872 `get-bytevector-some' now uses buffered input, which is much faster.
874 Finally, `array-ref', `array-set!' on arrays of rank 1 or 2 is now
875 faster, because it avoids building a rest list. Similarly, the
876 one-argument case of `array-for-each' and `array-map!' has been
877 optimized, and `array-copy!' and `array-fill!' are faster.
879 ** `peek-char' no longer consumes EOF
881 As required by the R5RS, if `peek-char' returns EOF, then the next read
882 will also return EOF. Previously `peek-char' would consume the EOF.
883 This makes a difference for terminal devices where it is possible to
888 Guile's copy of Gnulib was updated to v0.0-7865-ga828bb2. The following
889 modules were imported from Gnulib: select, times, pipe-posix, fstat,
890 getlogin, poll, and c-strcase.
892 ** `include' resolves relative file names relative to including file
894 Given a relative file name, `include' will look for it relative to the
895 directory of the including file. This harmonizes the behavior of
896 `include' with that of `load'.
898 ** SLIB compatibility restored
900 Guile 2.0.8 is now compatible with SLIB. You will have to use a
901 development version of SLIB, however, until a new version of SLIB is
904 ** Better ,trace REPL command
906 Sometimes the ,trace output for nested function calls could overflow the
907 terminal width, which wasn't useful. Now there is a limit to the amount
908 of space the prefix will take. See the documentation for ",trace" for
911 ** Better docstring syntax supported for `case-lambda'
913 Docstrings can now be placed immediately after the `case-lambda' or
914 `case-lambda*' keyword. See "Case-lambda" in the manual.
916 ** Improved handling of Unicode byte order marks
918 See "BOM Handling" in the manual for details.
920 ** Update predefined character sets to Unicode 6.2
922 ** GMP 4.2 or later required
924 Guile used to require GMP at least version 4.1 (released in May 2002),
925 and now requires at least version 4.2 (released in March 2006).
929 ** Better SXML documentation
931 The documentation for SXML modules was much improved, though there is
932 still far to go. See "SXML" in manual.
936 Use of "iff" was replaced with standard English. Keyword arguments are
937 now documented consistently, along with their default values.
939 ** An end to the generated-documentation experiment
941 When Guile 2.0 imported some modules from Guile-Lib, they came with a
942 system that generated documentation from docstrings and module
943 commentaries. This produced terrible documentation. We finally bit the
944 bullet and incorporated these modules into the main text, and will be
945 improving them manually over time, as is the case with SXML. Help is
950 There is now documentation for `scm_array_type', and `scm_array_ref', as
951 well as for the new `array-length' / 'scm_c_array_length' /
952 `scm_array_length' functions. `array-in-bounds?' has better
953 documentation as well. The `program-arguments-alist' and
954 `program-lambda-list' functions are now documented, as well as `and=>',
955 `exit', and `quit'. The (system repl server) module is now documented
956 (see REPL Servers). Finally, the GOOPS class hierarchy diagram has been
957 regenerated for the web and print output formats.
961 ** Deprecate generalized vector interface
963 The generalized vector interface, introduced in 1.8.0, is simply a
964 redundant, verbose interface to arrays of rank 1. `array-ref' and
965 similar functions are entirely sufficient. Thus,
966 `scm_generalized_vector_p', `scm_generalized_vector_length',
967 `scm_generalized_vector_ref', `scm_generalized_vector_set_x', and
968 `scm_generalized_vector_to_list' are now deprecated.
970 ** Deprecate SCM_CHAR_CODE_LIMIT and char-code-limit
972 These constants were defined to 256, which is not the highest codepoint
973 supported by Guile. Given that they were useless and incorrect, they
974 have been deprecated.
976 ** Deprecate `http-get*'
978 The new `#:streaming?' argument to `http-get' subsumes the functionality
979 of `http-get*' (introduced in 2.0.7). Also, the `#:extra-headers'
980 argument is deprecated in favor of `#:headers'.
982 ** Deprecate (ice-9 mapping)
984 This module, present in Guile since 1996 but never used or documented,
985 has never worked in Guile 2.0. It has now been deprecated and will be
986 removed in Guile 2.2.
988 ** Deprecate undocumented array-related C functions
990 These are `scm_array_fill_int', `scm_ra_eqp', `scm_ra_lessp',
991 `scm_ra_leqp', `scm_ra_grp', `scm_ra_greqp', `scm_ra_sum',
992 `scm_ra_product', `scm_ra_difference', `scm_ra_divide', and
993 `scm_array_identity'.
999 See "SRFI-41" in the manual.
1001 ** SRFI-45 exports `promise?'
1003 SRFI-45 now exports a `promise?' procedure that works with its promises.
1004 Also, its promises now print more nicely.
1006 ** New HTTP client procedures
1008 See "Web Client" for documentation on the new `http-head', `http-post',
1009 `http-put', `http-delete', `http-trace', and `http-options' procedures,
1010 and also for more options to `http-get'.
1012 ** Much more capable `xml->sxml'
1014 See "Reading and Writing XML" for information on how the `xml->sxml'
1015 parser deals with namespaces, processed entities, doctypes, and literal
1016 strings. Incidentally, `current-ssax-error-port' is now a parameter
1019 ** New procedures for converting strings to and from bytevectors
1021 See "Representing Strings as Bytes" for documention on the new `(ice-9
1022 iconv)' module and its `bytevector->string' and `string->bytevector'
1025 ** Escape continuations with `call/ec' and `let/ec'
1027 See "Prompt Primitives".
1029 ** New procedures to read all characters from a port
1031 See "Line/Delimited" in the manual for documentation on `read-string'
1034 ** New procedure `sendfile'
1038 ** New procedure `unget-bytevector'
1040 See "R6RS Binary Input".
1042 ** New C helper: `scm_c_bind_keyword_arguments'
1044 See "Keyword Procedures".
1046 ** New command-line arguments: `--language' and `-C'
1048 See "Command-line Options" in the manual.
1050 ** New environment variables: `GUILE_STACK_SIZE', `GUILE_INSTALL_LOCALE'
1052 See "Environment Variables".
1054 ** New procedures for dealing with file names
1056 See "File System" for documentation on `system-file-name-convention',
1057 `file-name-separator?', `absolute-file-name?', and
1058 `file-name-separator-string'.
1060 ** `array-length', an array's first dimension
1062 See "Array Procedures".
1064 ** `hash-count', for hash tables
1068 ** `round-ash', a bit-shifting operator that rounds on right-shift
1070 See "Bitwise Operations".
1072 ** New foreign types: `ssize_t', `ptrdiff_t'
1074 See "Foreign Types".
1076 ** New C helpers: `scm_from_ptrdiff_t', `scm_to_ptrdiff_t'
1080 ** Socket option `SO_REUSEPORT' now available from Scheme
1082 If supported on the platform, `SO_REUSEPORT' is now available from
1083 Scheme as well. See "Network Sockets and Communication".
1085 ** `current-language' in default environment
1087 Previously defined only in `(system base language)', `current-language'
1088 is now defined in the default environment, and is used to determine the
1089 language for the REPL, and for `compile-and-load'.
1091 ** New procedure: `fluid->parameter'
1093 See "Parameters", for information on how to convert a fluid to a
1096 ** New `print' REPL option
1098 See "REPL Commands" in the manual for information on the new
1099 user-customizable REPL printer.
1101 ** New variable: %site-ccache-dir
1103 The "Installing Site Packages" and "Build Config" manual sections now
1104 refer to this variable to describe where users should install their
1109 ** Fix compilation against libgc 7.3.
1110 ** Fix cross-compilation of `c-tokenize.o'.
1111 ** Fix warning when compiling against glibc 2.17.
1112 ** Fix documentation build against Texinfo 5.0.
1113 ** Fix building Guile from a directory with non-ASCII characters.
1114 ** Fix native MinGW build.
1115 ** Fix --disable-posix build.
1116 ** Fix MinGW builds with networking, POSIX, and thread support.
1120 ** Fix inexact number printer.
1121 (http://bugs.gnu.org/13757)
1122 ** Fix infinite loop when parsing optional-argument short options (SRFI-37).
1123 (http://bugs.gnu.org/13176)
1124 ** web: Support non-GMT date headers in the HTTP client.
1125 (http://bugs.gnu.org/13544)
1126 ** web: support IP-literal (IPv6 address) in Host header.
1127 ** Avoid stack overflows with `par-map' and nested futures in general.
1128 (http://bugs.gnu.org/13188)
1129 ** Peek-char no longer consumes EOF.
1130 (http://bugs.gnu.org/12216)
1131 ** Avoid swallowing multiple EOFs in R6RS binary-input procedures.
1132 ** A fork when multiple threads are running will now print a warning.
1133 ** Allow for spurious wakeups from pthread_cond_wait.
1134 (http://bugs.gnu.org/10641)
1135 ** Warn and ignore module autoload failures.
1136 (http://bugs.gnu.org/12202)
1137 ** Use chmod portably in (system base compile).
1138 (http://bugs.gnu.org/10474)
1139 ** Fix response-body-port for HTTP responses without content-length.
1140 (http://bugs.gnu.org/13857)
1141 ** Allow case-lambda expressions with no clauses.
1142 (http://bugs.gnu.org/9776)
1143 ** Improve standards conformance of string->number.
1144 (http://bugs.gnu.org/11887)
1145 ** Support calls and tail-calls with more than 255 formals.
1146 ** ,option evaluates its right-hand-side.
1147 (http://bugs.gnu.org/13076)
1148 ** Structs with tail arrays are not simple.
1149 (http://bugs.gnu.org/12808)
1150 ** Make `SCM_LONG_BIT' usable in preprocessor conditionals.
1151 (http://bugs.gnu.org/13848)
1152 ** Fix thread-unsafe lazy initializations.
1153 ** Allow SMOB mark procedures to be called from parallel markers.
1154 (http://bugs.gnu.org/13611)
1155 ** Fix later-bindings-win logic in with-fluids.
1156 (http://bugs.gnu.org/13843)
1157 ** Fix duplicate removal of with-fluids.
1158 (http://bugs.gnu.org/13838)
1159 ** Support calling foreign functions of 10 arguments or more.
1160 (http://bugs.gnu.org/13809)
1161 ** Let reverse! accept arbitrary types as second argument.
1162 (http://bugs.gnu.org/13835)
1163 ** Recognize the `x86_64.*-gnux32' triplet.
1164 ** Check whether a triplet's OS part specifies an ABI.
1165 ** Recognize mips64* as having 32-bit pointers by default.
1166 ** Use portable sed constructs.
1167 (http://bugs.gnu.org/14042)
1168 ** Remove language/glil/decompile-assembly.scm.
1169 (http://bugs.gnu.org/10622)
1170 ** Use O_BINARY in `copy-file', `load-objcode', `mkstemp'.
1171 ** Use byte-oriented functions in `get-bytevector*'.
1172 ** Fix abort when iconv swallows BOM from UTF-16 or UTF-32 stream.
1173 ** Fix compilation of functions with more than 255 local variables.
1174 ** Fix `getgroups' for when zero supplementary group IDs exist.
1175 ** Allow (define-macro name (lambda ...)).
1176 ** Various fixes to the (texinfo) modules.
1177 ** guild: Gracefully handle failures to install the locale.
1178 ** Fix format string warnings for ~!, ~|, ~/, ~q, ~Q, and ~^.
1179 (http://bugs.gnu.org/13485)
1180 ** Fix source annotation bug in psyntax 'expand-body'.
1181 ** Ecmascript: Fix conversion to boolean for non-numbers.
1182 ** Use case-insensitive comparisons for encoding names.
1183 ** Add missing cond-expand feature identifiers.
1184 ** A failure to find a module's file does not prevent future loading.
1185 ** Many (oop goops save) fixes.
1186 ** `http-get': don't shutdown write end of socket.
1187 (http://bugs.gnu.org/13095)
1188 ** Avoid signed integer overflow in scm_product.
1189 ** http: read-response-body always returns bytevector or #f, never EOF.
1190 ** web: Correctly detect "No route to host" conditions.
1191 ** `system*': failure to execvp no longer leaks dangling processes.
1192 (http://bugs.gnu.org/13166)
1193 ** More sensible case-lambda* dispatch.
1194 (http://bugs.gnu.org/12929)
1195 ** Do not defer expansion of internal define-syntax forms.
1196 (http://bugs.gnu.org/13509)
1200 Changes in 2.0.7 (since 2.0.6):
1204 ** SRFI-105 curly infix expressions are supported
1206 Curly infix expressions as described at
1207 http://srfi.schemers.org/srfi-105/srfi-105.html are now supported by
1208 Guile's reader. This allows users to write things like {a * {b + c}}
1209 instead of (* a (+ b c)). SRFI-105 support is enabled by using the
1210 `#!curly-infix' directive in source code, or the `curly-infix' reader
1211 option. See the manual for details.
1213 ** Reader options may now be per-port
1215 Historically, `read-options' and related procedures would manipulate
1216 global options, affecting the `read' procedure for all threads, and all
1217 current uses of `read'.
1219 Guile can now associate `read' options with specific ports, allowing
1220 different ports to use different options. For instance, the
1221 `#!fold-case' and `#!no-fold-case' reader directives have been
1222 implemented, and their effect is to modify the current read options of
1223 the current port only; similarly for `#!curly-infix'. Thus, it is
1224 possible, for instance, to have one port reading case-sensitive code,
1225 while another port reads case-insensitive code.
1227 ** Futures may now be nested
1229 Futures may now be nested: a future can itself spawn and then `touch'
1230 other futures. In addition, any thread that touches a future that has
1231 not completed now processes other futures while waiting for the touched
1232 future to completed. This allows all threads to be kept busy, and was
1233 made possible by the use of delimited continuations (see the manual for
1236 Consequently, `par-map' and `par-for-each' have been rewritten and can
1239 ** `GUILE_LOAD_PATH' et al can now add directories to the end of the path
1241 `GUILE_LOAD_PATH' and `GUILE_LOAD_COMPILED_PATH' can now be used to add
1242 directories to both ends of the load path. If the special path
1243 component `...' (ellipsis) is present in these environment variables,
1244 then the default path is put in place of the ellipsis, otherwise the
1245 default path is placed at the end. See "Environment Variables" in the
1248 ** `load-in-vicinity' search for `.go' files in `%load-compiled-path'
1250 Previously, `load-in-vicinity' would look for compiled files in the
1251 auto-compilation cache, but not in `%load-compiled-path'. This is now
1252 fixed. This affects `load', and the `-l' command-line flag. See
1253 <http://bugs.gnu.org/12519> for details.
1255 ** Extension search order fixed, and LD_LIBRARY_PATH preserved
1257 Up to 2.0.6, Guile would modify the `LD_LIBRARY_PATH' environment
1258 variable (or whichever is relevant for the host OS) to insert its own
1259 default extension directories in the search path (using GNU libltdl
1260 facilities was not possible here.) This approach was problematic in two
1263 First, the `LD_LIBRARY_PATH' modification would be visible to
1264 sub-processes, and would also affect future calls to `dlopen', which
1265 could lead to subtle bugs in the application or sub-processes. Second,
1266 when the installation prefix is /usr, the `LD_LIBRARY_PATH' modification
1267 would typically end up inserting /usr/lib before /usr/local/lib in the
1268 search path, which is often the opposite of system-wide settings such as
1271 Both issues have now been fixed.
1273 ** `make-vtable-vtable' is now deprecated
1275 Programs should instead use `make-vtable' and `<standard-vtable>'.
1277 ** The `-Wduplicate-case-datum' and `-Wbad-case-datum' are enabled
1279 These recently introduced warnings have been documented and are now
1280 enabled by default when auto-compiling.
1282 ** Optimize calls to `equal?' or `eqv?' with a constant argument
1284 The compiler simplifies calls to `equal?' or `eqv?' with a constant
1285 argument to use `eq?' instead, when applicable.
1289 ** SRFI-9 records now documented under "Compound Data Types"
1291 The documentation of SRFI-9 record types has been moved in the "Compound
1292 Data Types", next to Guile's other record APIs. A new section
1293 introduces the various record APIs, and describes the trade-offs they
1294 make. These changes were made in an attempt to better guide users
1295 through the maze of records API, and to recommend SRFI-9 as the main
1298 The documentation of Guile's raw `struct' API has also been improved.
1300 ** (ice-9 and-let-star) and (ice-9 curried-definitions) now documented
1302 These modules were missing from the manual.
1306 ** New "functional record setters" as a GNU extension of SRFI-9
1308 The (srfi srfi-9 gnu) module now provides three new macros to deal with
1309 "updates" of immutable records: `define-immutable-record-type',
1310 `set-field', and `set-fields'.
1312 The first one allows record type "functional setters" to be defined;
1313 such setters keep the record unchanged, and instead return a new record
1314 with only one different field. The remaining macros provide the same
1315 functionality, and also optimize updates of multiple or nested fields.
1316 See the manual for details.
1318 ** web: New `http-get*', `response-body-port', and `text-content-type?'
1321 These procedures return a port from which to read the response's body.
1322 Unlike `http-get' and `read-response-body', they allow the body to be
1323 processed incrementally instead of being stored entirely in memory.
1325 The `text-content-type?' predicate allows users to determine whether the
1326 content type of a response is textual.
1328 See the manual for details.
1330 ** `string-split' accepts character sets and predicates
1332 The `string-split' procedure can now be given a SRFI-14 character set or
1333 a predicate, instead of just a character.
1335 ** R6RS SRFI support
1337 Previously, in R6RS modules, Guile incorrectly ignored components of
1338 SRFI module names after the SRFI number, making it impossible to specify
1339 sub-libraries. This release corrects this, bringing us into accordance
1342 ** `define-public' is no a longer curried definition by default
1344 The (ice-9 curried-definitions) should be used for such uses. See the
1349 ** Remove reference to `scm_init_popen' when `fork' is unavailable
1351 This fixes a MinGW build issue (http://bugs.gnu.org/12477).
1353 ** Fix race between installing `guild' and the `guile-tools' symlink
1357 ** Procedures returned by `eval' now have docstrings
1358 (http://bugs.gnu.org/12173)
1359 ** web client: correctly handle uri-query, etc. in relative URI headers
1360 (http://bugs.gnu.org/12827)
1361 ** Fix docs for R6RS `hashtable-copy'
1362 ** R6RS `string-for-each' now accepts multiple string arguments
1363 ** Fix out-of-range error in the compiler's CSE pass
1364 (http://bugs.gnu.org/12883)
1365 ** Add missing R6RS `open-file-input/output-port' procedure
1366 ** Futures: Avoid creating the worker pool more than once
1367 ** Fix invalid assertion about mutex ownership in threads.c
1368 (http://bugs.gnu.org/12719)
1369 ** Have `SCM_NUM2FLOAT' and `SCM_NUM2DOUBLE' use `scm_to_double'
1370 ** The `scandir' procedure now uses `lstat' instead of `stat'
1371 ** Fix `generalized-vector->list' indexing bug with shared arrays
1372 (http://bugs.gnu.org/12465)
1373 ** web: Change `http-get' to try all the addresses for the given URI
1374 ** Implement `hash' for structs
1375 (http://lists.gnu.org/archive/html/guile-devel/2012-10/msg00031.html)
1376 ** `read' now adds source properties for data types beyond pairs
1377 ** Improve error reporting in `append!'
1378 ** In fold-matches, set regexp/notbol unless matching string start
1379 ** Don't stat(2) and access(2) the .go location before using it
1380 ** SRFI-19: use zero padding for hours in ISO 8601 format, not blanks
1381 ** web: Fix uri-encoding for strings with no unreserved chars, and octets 0-15
1382 ** More robust texinfo alias handling
1383 ** Optimize `format' and `simple-format'
1384 (http://bugs.gnu.org/12033)
1385 ** Angle of -0.0 is pi, not zero
1388 Changes in 2.0.6 (since 2.0.5):
1392 ** New optimization pass: common subexpression elimination (CSE)
1394 Guile's optimizer will now run a CSE pass after partial evaluation.
1395 This pass propagates static information about branches taken, bound
1396 lexicals, and effects from an expression's dominators. It can replace
1397 common subexpressions with their boolean values (potentially enabling
1398 dead code elimination), equivalent bound lexicals, or it can elide them
1399 entirely, depending on the context in which they are executed. This
1400 pass is especially useful in removing duplicate type checks, such as
1401 those produced by SRFI-9 record accessors.
1403 ** Improvements to the partial evaluator
1405 Peval can now hoist tests that are common to both branches of a
1406 conditional into the test. This can help with long chains of
1407 conditionals, such as those generated by the `match' macro. Peval can
1408 now do simple beta-reductions of procedures with rest arguments. It
1409 also avoids residualizing degenerate lexical aliases, even when full
1410 inlining is not possible. Finally, peval now uses the effects analysis
1411 introduced for the CSE pass. More precise effects analysis allows peval
1414 ** Run finalizers asynchronously in asyncs
1416 Finalizers are now run asynchronously, via an async. See Asyncs in the
1417 manual. This allows Guile and user code to safely allocate memory while
1420 ** Update SRFI-14 character sets to Unicode 6.1
1422 Note that this update causes the Latin-1 characters `§' and `¶' to be
1423 reclassified as punctuation. They were previously considered to be part
1424 of `char-set:symbol'.
1426 ** Better source information for datums
1428 When the `positions' reader option is on, as it is by default, Guile's
1429 reader will record source information for more kinds of datums.
1431 ** Improved error and warning messages
1433 `syntax-violation' errors now prefer `subform' for source info, with
1434 `form' as fallback. Syntactic errors in `cond' and `case' now produce
1435 better errors. `case' can now warn on duplicate datums, or datums that
1436 cannot be usefully compared with `eqv?'. `-Warity-mismatch' now handles
1437 applicable structs. `-Wformat' is more robust in the presence of
1438 `gettext'. Finally, various exceptions thrown by the Web modules now
1439 define appropriate exception printers.
1441 ** A few important bug fixes in the HTTP modules.
1443 Guile's web server framework now checks if an application returns a body
1444 where it is not permitted, for example in response to a HEAD request,
1445 and warn or truncate the response as appropriate. Bad requests now
1446 cause a 400 Bad Request response to be printed before closing the port.
1447 Finally, some date-printing and URL-parsing bugs were fixed.
1449 ** Pretty-print improvements
1451 When Guile needs to pretty-print Tree-IL, it will try to reconstruct
1452 `cond', `or`, and other derived syntax forms from the primitive tree-IL
1453 forms. It also uses the original names instead of the fresh unique
1454 names, when it is unambiguous to do so. This can be seen in the output
1455 of REPL commands like `,optimize'.
1457 Also, the `pretty-print' procedure has a new keyword argument,
1460 ** Fix memory leak involving applicable SMOBs
1462 At some point in the 1.9.x series, Guile began leaking any applicable
1463 SMOB that was actually applied. (There was a weak-key map from SMOB to
1464 trampoline functions, where the value had a strong reference on the
1465 key.) This has been fixed. There was much rejoicing!
1467 ** Support for HTTP/1.1 chunked transfer coding
1469 See "Transfer Codings" in the manual, for more.
1471 ** Micro-optimizations
1473 A pile of micro-optimizations: the `string-trim' function when called
1474 with `char-set:whitespace'; the `(web http)' parsers; SMOB application;
1475 conversion of raw UTF-8 and UTF-32 data to and from SCM strings; vlists
1476 and vhashes; `read' when processing string literals.
1478 ** Incompatible change to `scandir'
1480 As was the original intention, `scandir' now runs the `select?'
1481 procedure on all items, including subdirectories and the `.' and `..'
1482 entries. It receives the basename of the file in question instead of
1483 the full name. We apologize for this incompatible change to this
1484 function introduced in the 2.0.4 release.
1488 The manual has been made much more consistent in its naming conventions
1489 with regards to formal parameters of functions. Thanks to Bake Timmons.
1493 ** New C function: `scm_to_pointer'
1494 ** New C inline functions: `scm_new_smob', `scm_new_double_smob'
1495 ** (ice-9 format): Add ~h specifier for localized number output.
1496 ** (web response): New procedure: `response-must-not-include-body?'
1497 ** New predicate: 'supports-source-properties?'
1498 ** New C helpers: `scm_c_values', `scm_c_nvalues'
1499 ** Newly public inline C function: `scm_unget_byte'
1500 ** (language tree-il): New functions: `tree-il=?', `tree-il-hash'
1501 ** New fluid: `%default-port-conversion-strategy'
1502 ** New syntax: `=>' within `case'
1503 ** (web http): `make-chunked-input-port', `make-chunked-output-port'
1504 ** (web http): `declare-opaque-header!'
1506 Search the manual for these identifiers, for more information.
1510 ** `close-io-port' deprecated
1514 ** `scm_sym2var' deprecated
1516 In most cases, replace with `scm_lookup' or `scm_module_variable'. Use
1517 `scm_define' or `scm_module_ensure_local_variable' if the second
1518 argument is nonzero. See "Accessing Modules from C" in the manual, for
1521 ** Lookup closures deprecated
1523 These were never documented. See "Module System Reflection" in the
1524 manual for replacements.
1528 ** Fix compilation against uninstalled Guile on non-GNU platforms.
1529 ** Fix `SCM_I_ERROR' definition for MinGW without networking.
1530 ** Fix compilation with the Sun C compiler.
1531 ** Fix check for `clock_gettime' on OpenBSD and some other systems.
1532 ** Fix build with --enable-debug-malloc.
1533 ** Honor $(program_transform_name) for the `guile-tools' symlink.
1534 ** Fix cross-compilation of GOOPS-using code.
1538 ** Fix use of unitialized stat buffer in search-path of absolute paths.
1539 ** Avoid calling `freelocale' with a NULL argument.
1540 ** Work around erroneous tr_TR locale in Darwin 8 in tests.
1541 ** Fix `getaddrinfo' test for Darwin 8.
1542 ** Use Gnulib's `regex' module for better regex portability.
1543 ** `source-properties' and friends work on any object
1544 ** Rewrite open-process in C, for robustness related to threads and fork
1545 ** Fix <TAG>vector-length when applied to other uniform vector types
1546 ** Fix escape-only prompt optimization (was disabled previously)
1547 ** Fix a segfault when /dev/urandom is not accessible
1548 ** Fix flush on soft ports, so that it actually runs.
1549 ** Better compatibility of SRFI-9 records with core records
1550 ** Fix and clarify documentation of `sorted?'.
1551 ** Fix IEEE-754 endianness conversion in bytevectors.
1552 ** Correct thunk check in the `wind' instruction.
1553 ** Add @acronym support to texinfo modules
1554 ** Fix docbook->texi for <ulink> without URL
1555 ** Fix `setvbuf' to leave the line/column number unchanged.
1556 ** Add missing public declaration for `scm_take_from_input_buffers'.
1557 ** Fix relative file name canonicalization with empty %LOAD-PATH entries.
1558 ** Import newer (ice-9 match) from Chibi-Scheme.
1559 ** Fix unbound variables and unbound values in ECMAScript runtime.
1560 ** Make SRFI-6 string ports Unicode-capable.
1563 Changes in 2.0.5 (since 2.0.4):
1565 This release fixes the binary interface information (SONAME) of
1566 libguile, which was incorrect in 2.0.4. It does not contain other
1570 Changes in 2.0.4 (since 2.0.3):
1574 ** Better debuggability for interpreted procedures.
1576 Guile 2.0 came with a great debugging experience for compiled
1577 procedures, but the story for interpreted procedures was terrible. Now,
1578 at least, interpreted procedures have names, and the `arity' procedure
1579 property is always correct (or, as correct as it can be, in the presence
1582 ** Support for cross-compilation.
1584 One can now use a native Guile to cross-compile `.go' files for a
1585 different architecture. See the documentation for `--target' in the
1586 "Compilation" section of the manual, for information on how to use the
1587 cross-compiler. See the "Cross building Guile" section of the README,
1588 for more on how to cross-compile Guile itself.
1590 ** The return of `local-eval'.
1592 Back by popular demand, `the-environment' and `local-eval' allow the
1593 user to capture a lexical environment, and then evaluate arbitrary
1594 expressions in that context. There is also a new `local-compile'
1595 command. See "Local Evaluation" in the manual, for more. Special
1596 thanks to Mark Weaver for an initial implementation of this feature.
1598 ** Fluids can now have default values.
1600 Fluids are used for dynamic and thread-local binding. They have always
1601 inherited their values from the context or thread that created them.
1602 However, there was a case in which a new thread would enter Guile, and
1603 the default values of all the fluids would be `#f' for that thread.
1605 This has now been fixed so that `make-fluid' has an optional default
1606 value for fluids in unrelated dynamic roots, which defaults to `#f'.
1608 ** Garbage collector tuning.
1610 The garbage collector has now been tuned to run more often under some
1613 *** Unmanaged allocation
1615 The new `scm_gc_register_allocation' function will notify the collector
1616 of unmanaged allocation. This will cause the collector to run sooner.
1617 Guile's `scm_malloc', `scm_calloc', and `scm_realloc' unmanaged
1618 allocators eventually call this function. This leads to better
1619 performance under steady-state unmanaged allocation.
1621 *** Transient allocation
1623 When the collector runs, it will try to record the total memory
1624 footprint of a process, if the platform supports this information. If
1625 the memory footprint is growing, the collector will run more frequently.
1626 This reduces the increase of the resident size of a process in response
1627 to a transient increase in allocation.
1629 *** Management of threads, bignums
1631 Creating a thread will allocate a fair amount of memory. Guile now does
1632 some GC work (using `GC_collect_a_little') when allocating a thread.
1633 This leads to a better memory footprint when creating many short-lived
1636 Similarly, bignums can occupy a lot of memory. Guile now offers hooks
1637 to enable custom GMP allocators that end up calling
1638 `scm_gc_register_allocation'. These allocators are enabled by default
1639 when running Guile from the command-line. To enable them in libraries,
1640 set the `scm_install_gmp_memory_functions' variable to a nonzero value
1641 before loading Guile.
1643 ** SRFI-39 parameters are available by default.
1645 Guile now includes support for parameters, as defined by SRFI-39, in the
1646 default environment. See "Parameters" in the manual, for more
1647 information. `current-input-port', `current-output-port', and
1648 `current-error-port' are now parameters.
1650 ** Add `current-warning-port'.
1652 Guile now outputs warnings on a separate port, `current-warning-port',
1653 initialized to the value that `current-error-port' has on startup.
1655 ** Syntax parameters.
1657 Following Racket's lead, Guile now supports syntax parameters. See
1658 "Syntax parameters" in the manual, for more.
1660 Also see Barzilay, Culpepper, and Flatt's 2011 SFP workshop paper,
1661 "Keeping it Clean with syntax-parameterize".
1663 ** Parse command-line arguments from the locale encoding.
1665 Guile now attempts to parse command-line arguments using the user's
1666 locale. However for backwards compatibility with other 2.0.x releases,
1667 it does so without actually calling `setlocale'. Please report any bugs
1668 in this facility to bug-guile@gnu.org.
1670 ** One-armed conditionals: `when' and `unless'
1672 Guile finally has `when' and `unless' in the default environment. Use
1673 them whenever you would use an `if' with only one branch. See
1674 "Conditionals" in the manual, for more.
1676 ** `current-filename', `add-to-load-path'
1678 There is a new form, `(current-filename)', which expands out to the
1679 source file in which it occurs. Combined with the new
1680 `add-to-load-path', this allows simple scripts to easily add nearby
1681 directories to the load path. See "Load Paths" in the manual, for more.
1683 ** `random-state-from-platform'
1685 This procedure initializes a random seed using good random sources
1686 available on your platform, such as /dev/urandom. See "Random Number
1687 Generation" in the manual, for more.
1689 ** Warn about unsupported `simple-format' options.
1691 The `-Wformat' compilation option now reports unsupported format options
1692 passed to `simple-format'.
1696 Besides the sections already mentioned, the following manual sections
1697 are new in this release: "Modules and the File System", "Module System
1698 Reflection", "Syntax Transformer Helpers", and "Local Inclusion".
1702 ** (ice-9 session): `apropos-hook'
1703 ** New print option: `escape-newlines', defaults to #t.
1704 ** (ice-9 ftw): `file-system-fold', `file-system-tree', `scandir'
1705 ** `scm_c_value_ref': access to multiple returned values from C
1706 ** scm_call (a varargs version), scm_call_7, scm_call_8, scm_call_9
1707 ** Some new syntax helpers in (system syntax)
1709 Search the manual for these identifiers and modules, for more.
1713 ** FreeBSD build fixes.
1714 ** OpenBSD compilation fixes.
1715 ** Solaris 2.10 test suite fixes.
1716 ** IA64 compilation fix.
1717 ** MinGW build fixes.
1718 ** Work around instruction reordering on SPARC and HPPA in the VM.
1719 ** Gnulib updates: added `dirfd', `setenv' modules.
1723 ** Add a deprecated alias for $expt.
1724 ** Add an exception printer for `getaddrinfo-error'.
1725 ** Add deprecated shim for `scm_display_error' with stack as first argument.
1726 ** Add warnings for unsupported `simple-format' options.
1727 ** Allow overlapping regions to be passed to `bytevector-copy!'.
1728 ** Better function prologue disassembly
1729 ** Compiler: fix miscompilation of (values foo ...) in some contexts.
1730 ** Compiler: fix serialization of #nil-terminated lists.
1731 ** Compiler: allow values bound in non-tail let expressions to be collected.
1732 ** Deprecate SCM_ASRTGO.
1733 ** Document invalidity of (begin) as expression; add back-compat shim.
1734 ** Don't leak file descriptors when mmaping objcode.
1735 ** Empty substrings no longer reference the original stringbuf.
1736 ** FFI: Fix `set-pointer-finalizer!' to leave the type cell unchanged.
1737 ** FFI: Hold a weak reference to the CIF made by `procedure->pointer'.
1738 ** FFI: Hold a weak reference to the procedure passed to `procedure->pointer'.
1739 ** FFI: Properly unpack small integer return values in closure call.
1740 ** Fix R6RS `fold-left' so the accumulator is the first argument.
1741 ** Fix bit-set*! bug from 2005.
1742 ** Fix bug in `make-repl' when `lang' is actually a <language>.
1743 ** Fix bugs related to mutation, the null string, and shared substrings.
1744 ** Fix <dynwind> serialization.
1745 ** Fix erroneous check in `set-procedure-properties!'.
1746 ** Fix generalized-vector-{ref,set!} for slices.
1747 ** Fix error messages involving definition forms.
1748 ** Fix primitive-eval to return #<unspecified> for definitions.
1749 ** HTTP: Extend handling of "Cache-Control" header.
1750 ** HTTP: Fix qstring writing of cache-extension values
1751 ** HTTP: Fix validators for various list-style headers.
1752 ** HTTP: Permit non-date values for Expires header.
1753 ** HTTP: `write-request-line' writes absolute paths, not absolute URIs.
1754 ** Hack the port-column of current-output-port after printing a prompt.
1755 ** Make sure `regexp-quote' tests use Unicode-capable string ports.
1756 ** Peval: Fix bugs in the new optimizer.
1757 ** Statistically unique marks and labels, for robust hygiene across sessions.
1758 ** Web: Allow URIs with empty authorities, like "file:///etc/hosts".
1759 ** `,language' at REPL sets the current-language fluid.
1760 ** `primitive-load' returns the value(s) of the last expression.
1761 ** `scm_from_stringn' always returns unique strings.
1762 ** `scm_i_substring_copy' tries to narrow the substring.
1763 ** i18n: Fix gc_malloc/free mismatch on non-GNU systems.
1766 Changes in 2.0.3 (since 2.0.2):
1768 * Speed improvements
1770 ** Guile has a new optimizer, `peval'.
1772 `Peval' is a partial evaluator that performs constant folding, dead code
1773 elimination, copy propagation, and inlining. By default it runs on
1774 every piece of code that Guile compiles, to fold computations that can
1775 happen at compile-time, so they don't have to happen at runtime.
1777 If we did our job right, the only impact you would see would be your
1778 programs getting faster. But if you notice slowdowns or bloated code,
1779 please send a mail to bug-guile@gnu.org with details.
1781 Thanks to William R. Cook, Oscar Waddell, and Kent Dybvig for inspiring
1782 peval and its implementation.
1784 You can see what peval does on a given piece of code by running the new
1785 `,optimize' REPL meta-command, and comparing it to the output of
1786 `,expand'. See "Compile Commands" in the manual, for more.
1788 ** Fewer calls to `stat'.
1790 Guile now stats only the .go file and the .scm file when loading a fresh
1795 ** New module: `(web client)', a simple synchronous web client.
1797 See "Web Client" in the manual, for more.
1799 ** Users can now install compiled `.go' files.
1801 See "Installing Site Packages" in the manual.
1803 ** Remove Front-Cover and Back-Cover text from the manual.
1805 The manual is still under the GNU Free Documentation License, but no
1806 longer has any invariant sections.
1808 ** More helpful `guild help'.
1810 `guild' is Guile's multi-tool, for use in shell scripting. Now it has a
1811 nicer interface for querying the set of existing commands, and getting
1812 help on those commands. Try it out and see!
1814 ** New macro: `define-syntax-rule'
1816 `define-syntax-rule' is a shorthand to make a `syntax-rules' macro with
1817 one clause. See "Syntax Rules" in the manual, for more.
1819 ** The `,time' REPL meta-command now has more precision.
1821 The output of this command now has microsecond precision, instead of
1822 10-millisecond precision.
1824 ** `(ice-9 match)' can now match records.
1826 See "Pattern Matching" in the manual, for more on matching records.
1828 ** New module: `(language tree-il debug)'.
1830 This module provides a tree-il verifier. This is useful for people that
1831 generate tree-il, usually as part of a language compiler.
1833 ** New functions: `scm_is_exact', `scm_is_inexact'.
1835 These provide a nice C interface for Scheme's `exact?' and `inexact?',
1840 See the git log (or the ChangeLog) for more details on these bugs.
1842 ** Fix order of importing modules and resolving duplicates handlers.
1843 ** Fix a number of bugs involving extended (merged) generics.
1844 ** Fix invocation of merge-generics duplicate handler.
1845 ** Fix write beyond array end in arrays.c.
1846 ** Fix read beyond end of hashtable size array in hashtab.c.
1847 ** (web http): Locale-independent parsing and serialization of dates.
1848 ** Ensure presence of Host header in HTTP/1.1 requests.
1849 ** Fix take-right and drop-right for improper lists.
1850 ** Fix leak in get_current_locale().
1851 ** Fix recursive define-inlinable expansions.
1852 ** Check that srfi-1 procedure arguments are procedures.
1853 ** Fix r6rs `map' for multiple returns.
1854 ** Fix scm_tmpfile leak on POSIX platforms.
1855 ** Fix a couple of leaks (objcode->bytecode, make-boot-program).
1856 ** Fix guile-lib back-compatibility for module-stexi-documentation.
1857 ** Fix --listen option to allow other ports.
1858 ** Fix scm_to_latin1_stringn for substrings.
1859 ** Fix compilation of untyped arrays of rank not 1.
1860 ** Fix unparse-tree-il of <dynset>.
1861 ** Fix reading of #||||#.
1862 ** Fix segfault in GOOPS when class fields are redefined.
1863 ** Prefer poll(2) over select(2) to allow file descriptors above FD_SETSIZE.
1866 Changes in 2.0.2 (since 2.0.1):
1870 ** `guile-tools' renamed to `guild'
1872 The new name is shorter. Its intended future use is for a CPAN-like
1873 system for Guile wizards and journeyfolk to band together to share code;
1874 hence the name. `guile-tools' is provided as a backward-compatible
1875 symbolic link. See "Using Guile Tools" in the manual, for more.
1877 ** New control operators: `shift' and `reset'
1879 See "Shift and Reset" in the manual, for more information.
1881 ** `while' as an expression
1883 Previously the return value of `while' was unspecified. Now its
1884 values are specified both in the case of normal termination, and via
1885 termination by invoking `break', possibly with arguments. See "while
1886 do" in the manual for more.
1888 ** Disallow access to handles of weak hash tables
1890 `hash-get-handle' and `hash-create-handle!' are no longer permitted to
1891 be called on weak hash tables, because the fields in a weak handle could
1892 be nulled out by the garbage collector at any time, but yet they are
1893 otherwise indistinguishable from pairs. Use `hash-ref' and `hash-set!'
1896 ** More precision for `get-internal-run-time', `get-internal-real-time'
1898 On 64-bit systems which support POSIX clocks, Guile's internal timing
1899 procedures offer nanosecond resolution instead of the 10-millisecond
1900 resolution previously available. 32-bit systems now use 1-millisecond
1903 ** Guile now measures time spent in GC
1905 `gc-stats' now returns a meaningful value for `gc-time-taken'.
1909 The statprof profiler now exports a `gcprof' procedure, driven by the
1910 `after-gc-hook', to see which parts of your program are causing GC. Let
1911 us know if you find it useful.
1913 ** `map', `for-each' and some others now implemented in Scheme
1915 We would not mention this in NEWS, as it is not a user-visible change,
1916 if it were not for one thing: `map' and `for-each' are no longer
1917 primitive generics. Instead they are normal bindings, which can be
1918 wrapped by normal generics. This fixes some modularity issues between
1919 core `map', SRFI-1 `map', and GOOPS.
1921 Also it's pretty cool that we can do this without a performance impact.
1923 ** Add `scm_peek_byte_or_eof'.
1925 This helper is like `scm_peek_char_or_eof', but for bytes instead of
1928 ** Implement #:stop-at-first-non-option option for getopt-long
1930 See "getopt-long Reference" in the manual, for more information.
1932 ** Improve R6RS conformance for conditions in the I/O libraries
1934 The `(rnrs io simple)' module now raises the correct R6RS conditions in
1935 error cases. `(rnrs io ports)' is also more correct now, though it is
1936 still a work in progress.
1938 ** All deprecated routines emit warnings
1940 A few deprecated routines were lacking deprecation warnings. This has
1943 * Speed improvements
1945 ** Constants in compiled code now share state better
1947 Constants with shared state, like `("foo")' and `"foo"', now share state
1948 as much as possible, in the entire compilation unit. This cuts compiled
1949 `.go' file sizes in half, generally, and speeds startup.
1951 ** VLists: optimize `vlist-fold-right', and add `vhash-fold-right'
1953 These procedures are now twice as fast as they were.
1955 ** UTF-8 ports to bypass `iconv' entirely
1957 This reduces memory usage in a very common case.
1959 ** Compiler speedups
1961 The compiler is now about 40% faster. (Note that this is only the case
1962 once the compiler is itself compiled, so the build still takes as long
1967 Some assertions that were mostly useful for sanity-checks on the
1968 bytecode compiler are now off for both "regular" and "debug" engines.
1969 This together with a fix to cache a TLS access and some other tweaks
1970 improve the VM's performance by about 20%.
1972 ** SRFI-1 list-set optimizations
1974 lset-adjoin and lset-union now have fast paths for eq? sets.
1976 ** `memq', `memv' optimizations
1978 These procedures are now at least twice as fast than in 2.0.1.
1982 ** Deprecate scm_whash API
1984 `scm_whash_get_handle', `SCM_WHASHFOUNDP', `SCM_WHASHREF',
1985 `SCM_WHASHSET', `scm_whash_create_handle', `scm_whash_lookup', and
1986 `scm_whash_insert' are now deprecated. Use the normal hash table API
1989 ** Deprecate scm_struct_table
1991 `SCM_STRUCT_TABLE_NAME', `SCM_SET_STRUCT_TABLE_NAME',
1992 `SCM_STRUCT_TABLE_CLASS', `SCM_SET_STRUCT_TABLE_CLASS',
1993 `scm_struct_table', and `scm_struct_create_handle' are now deprecated.
1994 These routines formed part of the internals of the map between structs
1997 ** Deprecate scm_internal_dynamic_wind
1999 The `scm_t_inner' type and `scm_internal_dynamic_wind' are deprecated,
2000 as the `scm_dynwind' API is better, and this API encourages users to
2001 stuff SCM values into pointers.
2003 ** Deprecate scm_immutable_cell, scm_immutable_double_cell
2005 These routines are deprecated, as the GC_STUBBORN API doesn't do
2010 Andreas Rottman kindly transcribed the missing parts of the `(rnrs io
2011 ports)' documentation from the R6RS documentation. Thanks Andreas!
2015 ** Fix double-loading of script in -ds case
2016 ** -x error message fix
2017 ** iconveh-related cross-compilation fixes
2018 ** Fix small integer return value packing on big endian machines.
2019 ** Fix hash-set! in weak-value table from non-immediate to immediate
2020 ** Fix call-with-input-file & relatives for multiple values
2021 ** Fix `hash' for inf and nan
2022 ** Fix libguile internal type errors caught by typing-strictness==2
2023 ** Fix compile error in MinGW fstat socket detection
2024 ** Fix generation of auto-compiled file names on MinGW
2025 ** Fix multithreaded access to internal hash tables
2026 ** Emit a 1-based line number in error messages
2027 ** Fix define-module ordering
2028 ** Fix several POSIX functions to use the locale encoding
2029 ** Add type and range checks to the complex generalized vector accessors
2030 ** Fix unaligned accesses for bytevectors of complex numbers
2032 ** Fix erroneous VM stack overflow for canceled threads
2035 Changes in 2.0.1 (since 2.0.0):
2039 ** guile.m4 supports linking with rpath
2041 The GUILE_FLAGS macro now sets GUILE_LIBS and GUILE_LTLIBS, which
2042 include appropriate directives to the linker to include libguile-2.0.so
2043 in the runtime library lookup path.
2045 ** `begin' expands macros in its body before other expressions
2047 This enables support for programs like the following:
2052 (or (= x 0) (odd? (- x 1)))))
2055 ((odd? x) (not (even? x)))))
2058 ** REPL reader usability enhancements
2060 The REPL now flushes input after a read error, which should prevent one
2061 error from causing other errors. The REPL also now interprets comments
2064 ** REPL output has configurable width
2066 The REPL now defaults to output with the current terminal's width, in
2067 columns. See "Debug Commands" in the manual for more information on
2070 ** Better C access to the module system
2072 Guile now has convenient C accessors to look up variables or values in
2073 modules and their public interfaces. See `scm_c_public_ref' and friends
2074 in "Accessing Modules from C" in the manual.
2076 ** Added `scm_call_5', `scm_call_6'
2078 See "Fly Evaluation" in the manual.
2080 ** Added `scm_from_latin1_keyword', `scm_from_utf8_keyword'
2082 See "Keyword Procedures" in the manual, for more. Note that
2083 `scm_from_locale_keyword' should not be used when the name is a C string
2086 ** R6RS unicode and string I/O work
2088 Added efficient implementations of `get-string-n' and `get-string-n!'
2089 for binary ports. Exported `current-input-port', `current-output-port'
2090 and `current-error-port' from `(rnrs io ports)', and enhanced support
2093 ** Added `pointer->scm', `scm->pointer' to `(system foreign)'
2095 These procedure are useful if one needs to pass and receive SCM values
2096 to and from foreign functions. See "Foreign Variables" in the manual,
2099 ** Added `heap-allocated-since-gc' to `(gc-stats)'
2101 Also fixed the long-standing bug in the REPL `,stat' command.
2103 ** Add `on-error' REPL option
2105 This option controls what happens when an error occurs at the REPL, and
2106 defaults to `debug', indicating that Guile should enter the debugger.
2107 Other values include `report', which will simply print a backtrace
2108 without entering the debugger. See "System Commands" in the manual.
2110 ** Enforce immutability of string literals
2112 Attempting to mutate a string literal now causes a runtime error.
2114 ** Fix pthread redirection
2116 Guile 2.0.0 shipped with headers that, if configured with pthread
2117 support, would re-define `pthread_create', `pthread_join', and other API
2118 to redirect to the BDW-GC wrappers, `GC_pthread_create', etc. This was
2119 unintended, and not necessary: because threads must enter Guile with
2120 `scm_with_guile', Guile can handle thread registration itself, without
2121 needing to make the GC aware of all threads. This oversight has been
2124 ** `with-continuation-barrier' now unwinds on `quit'
2126 A throw to `quit' in a continuation barrier will cause Guile to exit.
2127 Before, it would do so before unwinding to the barrier, which would
2128 prevent cleanup handlers from running. This has been fixed so that it
2129 exits only after unwinding.
2131 ** `string->pointer' and `pointer->string' have optional encoding arg
2133 This allows users of the FFI to more easily deal in strings with
2134 particular (non-locale) encodings, like "utf-8". See "Void Pointers and
2135 Byte Access" in the manual, for more.
2137 ** R6RS fixnum arithmetic optimizations
2139 R6RS fixnum operations are are still slower than generic arithmetic,
2142 ** New procedure: `define-inlinable'
2144 See "Inlinable Procedures" in the manual, for more.
2146 ** New procedure: `exact-integer-sqrt'
2148 See "Integer Operations" in the manual, for more.
2150 ** "Extended read syntax" for symbols parses better
2152 In #{foo}# symbols, backslashes are now treated as escapes, as the
2153 symbol-printing code intended. Additionally, "\x" within #{foo}# is now
2154 interpreted as starting an R6RS hex escape. This is backward compatible
2155 because the symbol printer would never produce a "\x" before. The
2156 printer also works better too.
2158 ** Added `--fresh-auto-compile' option
2160 This allows a user to invalidate the auto-compilation cache. It's
2161 usually not needed. See "Compilation" in the manual, for a discussion.
2165 ** GOOPS documentation updates
2169 Thanks to Mark Harig for improvements to guile.1.
2171 ** SRFI-23 documented
2173 The humble `error' SRFI now has an entry in the manual.
2177 ** `(ice-9 binary-ports)': "R6RS I/O Ports", in the manual
2178 ** `(ice-9 eval-string)': "Fly Evaluation", in the manual
2179 ** `(ice-9 command-line)', not documented yet
2183 ** Fixed `iconv_t' memory leak on close-port
2184 ** Fixed some leaks with weak hash tables
2185 ** Export `vhash-delq' and `vhash-delv' from `(ice-9 vlist)'
2186 ** `after-gc-hook' works again
2187 ** `define-record-type' now allowed in nested contexts
2188 ** `exact-integer-sqrt' now handles large integers correctly
2189 ** Fixed C extension examples in manual
2190 ** `vhash-delete' honors HASH argument
2191 ** Make `locale-digit-grouping' more robust
2192 ** Default exception printer robustness fixes
2193 ** Fix presence of non-I CPPFLAGS in `guile-2.0.pc'
2194 ** `read' updates line/column numbers when reading SCSH block comments
2195 ** Fix imports of multiple custom interfaces of same module
2196 ** Fix encoding scanning for non-seekable ports
2197 ** Fix `setter' when called with a non-setter generic
2198 ** Fix f32 and f64 bytevectors to not accept rationals
2199 ** Fix description of the R6RS `finite?' in manual
2200 ** Quotient, remainder and modulo accept inexact integers again
2201 ** Fix `continue' within `while' to take zero arguments
2202 ** Fix alignment for structures in FFI
2203 ** Fix port-filename of stdin, stdout, stderr to match the docs
2204 ** Fix weak hash table-related bug in `define-wrapped-pointer-type'
2205 ** Fix partial continuation application with pending procedure calls
2206 ** scm_{to,from}_locale_string use current locale, not current ports
2207 ** Fix thread cleanup, by using a pthread_key destructor
2208 ** Fix `quit' at the REPL
2209 ** Fix a failure to sync regs in vm bytevector ops
2210 ** Fix (texinfo reflection) to handle nested structures like syntax patterns
2211 ** Fix stexi->html double translation
2212 ** Fix tree-il->scheme fix for <prompt>
2213 ** Fix compilation of <prompt> in <fix> in single-value context
2214 ** Fix race condition in ensure-writable-dir
2215 ** Fix error message on ,disassemble "non-procedure"
2216 ** Fix prompt and abort with the boot evaluator
2217 ** Fix `procedure->pointer' for functions returning `void'
2218 ** Fix error reporting in dynamic-pointer
2219 ** Fix problems detecting coding: in block comments
2220 ** Fix duplicate load-path and load-compiled-path in compilation environment
2221 ** Add fallback read(2) suppport for .go files if mmap(2) unavailable
2222 ** Fix c32vector-set!, c64vector-set!
2223 ** Fix mistakenly deprecated read syntax for uniform complex vectors
2224 ** Fix parsing of exact numbers with negative exponents
2225 ** Ignore SIGPIPE in (system repl server)
2226 ** Fix optional second arg to R6RS log function
2227 ** Fix R6RS `assert' to return true value.
2228 ** Fix fencepost error when seeking in bytevector input ports
2229 ** Gracefully handle `setlocale' errors when starting the REPL
2230 ** Improve support of the `--disable-posix' configure option
2231 ** Make sure R6RS binary ports pass `binary-port?' regardless of the locale
2232 ** Gracefully handle unterminated UTF-8 sequences instead of hitting an `assert'
2236 Changes in 2.0.0 (changes since the 1.8.x series):
2238 * New modules (see the manual for details)
2240 ** `(srfi srfi-18)', more sophisticated multithreading support
2241 ** `(srfi srfi-27)', sources of random bits
2242 ** `(srfi srfi-38)', External Representation for Data With Shared Structure
2243 ** `(srfi srfi-42)', eager comprehensions
2244 ** `(srfi srfi-45)', primitives for expressing iterative lazy algorithms
2245 ** `(srfi srfi-67)', compare procedures
2246 ** `(ice-9 i18n)', internationalization support
2247 ** `(ice-9 futures)', fine-grain parallelism
2248 ** `(rnrs bytevectors)', the R6RS bytevector API
2249 ** `(rnrs io ports)', a subset of the R6RS I/O port API
2250 ** `(system xref)', a cross-referencing facility (FIXME undocumented)
2251 ** `(ice-9 vlist)', lists with constant-time random access; hash lists
2252 ** `(system foreign)', foreign function interface
2253 ** `(sxml match)', a pattern matcher for SXML
2254 ** `(srfi srfi-9 gnu)', extensions to the SRFI-9 record library
2255 ** `(system vm coverage)', a line-by-line code coverage library
2256 ** `(web uri)', URI data type, parser, and unparser
2257 ** `(web http)', HTTP header parsers and unparsers
2258 ** `(web request)', HTTP request data type, reader, and writer
2259 ** `(web response)', HTTP response data type, reader, and writer
2260 ** `(web server)', Generic HTTP server
2261 ** `(ice-9 poll)', a poll wrapper
2262 ** `(web server http)', HTTP-over-TCP web server implementation
2264 ** Replaced `(ice-9 match)' with Alex Shinn's compatible, hygienic matcher.
2266 Guile's copy of Andrew K. Wright's `match' library has been replaced by
2267 a compatible hygienic implementation by Alex Shinn. It is now
2268 documented, see "Pattern Matching" in the manual.
2270 Compared to Andrew K. Wright's `match', the new `match' lacks
2271 `match-define', `match:error-control', `match:set-error-control',
2272 `match:error', `match:set-error', and all structure-related procedures.
2274 ** Imported statprof, SSAX, and texinfo modules from Guile-Lib
2276 The statprof statistical profiler, the SSAX XML toolkit, and the texinfo
2277 toolkit from Guile-Lib have been imported into Guile proper. See
2278 "Standard Library" in the manual for more details.
2280 ** Integration of lalr-scm, a parser generator
2282 Guile has included Dominique Boucher's fine `lalr-scm' parser generator
2283 as `(system base lalr)'. See "LALR(1) Parsing" in the manual, for more
2286 * Changes to the stand-alone interpreter
2288 ** Guile now can compile Scheme to bytecode for a custom virtual machine.
2290 Compiled code loads much faster than Scheme source code, and runs around
2291 3 or 4 times as fast, generating much less garbage in the process.
2293 ** Evaluating Scheme code does not use the C stack.
2295 Besides when compiling Guile itself, Guile no longer uses a recursive C
2296 function as an evaluator. This obviates the need to check the C stack
2297 pointer for overflow. Continuations still capture the C stack, however.
2299 ** New environment variables: GUILE_LOAD_COMPILED_PATH,
2300 GUILE_SYSTEM_LOAD_COMPILED_PATH
2302 GUILE_LOAD_COMPILED_PATH is for compiled files what GUILE_LOAD_PATH is
2303 for source files. It is a different path, however, because compiled
2304 files are architecture-specific. GUILE_SYSTEM_LOAD_COMPILED_PATH is like
2307 ** New read-eval-print loop (REPL) implementation
2309 Running Guile with no arguments drops the user into the new REPL. See
2310 "Using Guile Interactively" in the manual, for more information.
2312 ** Remove old Emacs interface
2314 Guile had an unused `--emacs' command line argument that was supposed to
2315 help when running Guile inside Emacs. This option has been removed, and
2316 the helper functions `named-module-use!' and `load-emacs-interface' have
2319 ** Add `(system repl server)' module and `--listen' command-line argument
2321 The `(system repl server)' module exposes procedures to listen on
2322 sockets for connections, and serve REPLs to those clients. The --listen
2323 command-line argument allows any Guile program to thus be remotely
2326 See "Invoking Guile" for more information on `--listen'.
2328 ** Command line additions
2330 The guile binary now supports a new switch "-x", which can be used to
2331 extend the list of filename extensions tried when loading files
2334 ** New reader options: `square-brackets', `r6rs-hex-escapes',
2335 `hungry-eol-escapes'
2337 The reader supports a new option (changeable via `read-options'),
2338 `square-brackets', which instructs it to interpret square brackets as
2339 parentheses. This option is on by default.
2341 When the new `r6rs-hex-escapes' reader option is enabled, the reader
2342 will recognize string escape sequences as defined in R6RS. R6RS string
2343 escape sequences are incompatible with Guile's existing escapes, though,
2344 so this option is off by default.
2346 Additionally, Guile follows the R6RS newline escaping rules when the
2347 `hungry-eol-escapes' option is enabled.
2349 See "String Syntax" in the manual, for more information.
2351 ** Function profiling and tracing at the REPL
2353 The `,profile FORM' REPL meta-command can now be used to statistically
2354 profile execution of a form, to see which functions are taking the most
2355 time. See `,help profile' for more information.
2357 Similarly, `,trace FORM' traces all function applications that occur
2358 during the execution of `FORM'. See `,help trace' for more information.
2360 ** Recursive debugging REPL on error
2362 When Guile sees an error at the REPL, instead of saving the stack, Guile
2363 will directly enter a recursive REPL in the dynamic context of the
2364 error. See "Error Handling" in the manual, for more information.
2366 A recursive REPL is the same as any other REPL, except that it
2367 has been augmented with debugging information, so that one can inspect
2368 the context of the error. The debugger has been integrated with the REPL
2369 via a set of debugging meta-commands.
2371 For example, one may access a backtrace with `,backtrace' (or
2372 `,bt'). See "Interactive Debugging" in the manual, for more
2375 ** New `guile-tools' commands: `compile', `disassemble'
2377 Pass the `--help' command-line option to these commands for more
2380 ** Guile now adds its install prefix to the LTDL_LIBRARY_PATH
2382 Users may now install Guile to nonstandard prefixes and just run
2383 `/path/to/bin/guile', instead of also having to set LTDL_LIBRARY_PATH to
2384 include `/path/to/lib'.
2386 ** Guile's Emacs integration is now more keyboard-friendly
2388 Backtraces may now be disclosed with the keyboard in addition to the
2391 ** Load path change: search in version-specific paths before site paths
2393 When looking for a module, Guile now searches first in Guile's
2394 version-specific path (the library path), *then* in the site dir. This
2395 allows Guile's copy of SSAX to override any Guile-Lib copy the user has
2396 installed. Also it should cut the number of `stat' system calls by half,
2399 ** Value history in the REPL on by default
2401 By default, the REPL will save computed values in variables like `$1',
2402 `$2', and the like. There are programmatic and interactive interfaces to
2403 control this. See "Value History" in the manual, for more information.
2405 ** Readline tab completion for arguments
2407 When readline is enabled, tab completion works for arguments too, not
2408 just for the operator position.
2410 ** Expression-oriented readline history
2412 Guile's readline history now tries to operate on expressions instead of
2413 input lines. Let us know what you think!
2415 ** Interactive Guile follows GNU conventions
2417 As recommended by the GPL, Guile now shows a brief copyright and
2418 warranty disclaimer on startup, along with pointers to more information.
2420 * Changes to Scheme functions and syntax
2422 ** Support for R6RS libraries
2424 The `library' and `import' forms from the latest Scheme report have been
2425 added to Guile, in such a way that R6RS libraries share a namespace with
2426 Guile modules. R6RS modules may import Guile modules, and are available
2427 for Guile modules to import via use-modules and all the rest. See "R6RS
2428 Libraries" in the manual for more information.
2430 ** Implementations of R6RS libraries
2432 Guile now has implementations for all of the libraries defined in the
2433 R6RS. Thanks to Julian Graham for this excellent hack. See "R6RS
2434 Standard Libraries" in the manual for a full list of libraries.
2436 ** Partial R6RS compatibility
2438 Guile now has enough support for R6RS to run a reasonably large subset
2441 Guile is not fully R6RS compatible. Many incompatibilities are simply
2442 bugs, though some parts of Guile will remain R6RS-incompatible for the
2443 foreseeable future. See "R6RS Incompatibilities" in the manual, for more
2446 Please contact bug-guile@gnu.org if you have found an issue not
2447 mentioned in that compatibility list.
2449 ** New implementation of `primitive-eval'
2451 Guile's `primitive-eval' is now implemented in Scheme. Actually there is
2452 still a C evaluator, used when building a fresh Guile to interpret the
2453 compiler, so we can compile eval.scm. Thereafter all calls to
2454 primitive-eval are implemented by VM-compiled code.
2456 This allows all of Guile's procedures, be they interpreted or compiled,
2457 to execute on the same stack, unifying multiple-value return semantics,
2458 providing for proper tail recursion between interpreted and compiled
2459 code, and simplifying debugging.
2461 As part of this change, the evaluator no longer mutates the internal
2462 representation of the code being evaluated in a thread-unsafe manner.
2464 There are two negative aspects of this change, however. First, Guile
2465 takes a lot longer to compile now. Also, there is less debugging
2466 information available for debugging interpreted code. We hope to improve
2467 both of these situations.
2469 There are many changes to the internal C evalator interface, but all
2470 public interfaces should be the same. See the ChangeLog for details. If
2471 we have inadvertantly changed an interface that you were using, please
2472 contact bug-guile@gnu.org.
2474 ** Procedure removed: `the-environment'
2476 This procedure was part of the interpreter's execution model, and does
2477 not apply to the compiler.
2479 ** No more `local-eval'
2481 `local-eval' used to exist so that one could evaluate code in the
2482 lexical context of a function. Since there is no way to get the lexical
2483 environment any more, as that concept has no meaning for the compiler,
2484 and a different meaning for the interpreter, we have removed the
2487 If you think you need `local-eval', you should probably implement your
2488 own metacircular evaluator. It will probably be as fast as Guile's
2491 ** Scheme source files will now be compiled automatically.
2493 If a compiled .go file corresponding to a .scm file is not found or is
2494 not fresh, the .scm file will be compiled on the fly, and the resulting
2495 .go file stored away. An advisory note will be printed on the console.
2497 Note that this mechanism depends on the timestamp of the .go file being
2498 newer than that of the .scm file; if the .scm or .go files are moved
2499 after installation, care should be taken to preserve their original
2502 Auto-compiled files will be stored in the $XDG_CACHE_HOME/guile/ccache
2503 directory, where $XDG_CACHE_HOME defaults to ~/.cache. This directory
2504 will be created if needed.
2506 To inhibit automatic compilation, set the GUILE_AUTO_COMPILE environment
2507 variable to 0, or pass --no-auto-compile on the Guile command line.
2509 ** New POSIX procedures: `getrlimit' and `setrlimit'
2511 Note however that the interface of these functions is likely to change
2512 in the next prerelease.
2514 ** New POSIX procedure: `getsid'
2516 Scheme binding for the `getsid' C library call.
2518 ** New POSIX procedure: `getaddrinfo'
2520 Scheme binding for the `getaddrinfo' C library function.
2522 ** Multicast socket options
2524 Support was added for the IP_MULTICAST_TTL and IP_MULTICAST_IF socket
2525 options. See "Network Sockets and Communication" in the manual, for
2528 ** `recv!', `recvfrom!', `send', `sendto' now deal in bytevectors
2530 These socket procedures now take bytevectors as arguments, instead of
2531 strings. There is some deprecated string support, however.
2533 ** New GNU procedures: `setaffinity' and `getaffinity'.
2535 See "Processes" in the manual, for more information.
2537 ** New procedures: `compose', `negate', and `const'
2539 See "Higher-Order Functions" in the manual, for more information.
2541 ** New procedure in `(oops goops)': `method-formals'
2543 ** New procedures in (ice-9 session): `add-value-help-handler!',
2544 `remove-value-help-handler!', `add-name-help-handler!'
2545 `remove-name-help-handler!', `procedure-arguments'
2547 The value and name help handlers provide some minimal extensibility to
2548 the help interface. Guile-lib's `(texinfo reflection)' uses them, for
2549 example, to make stexinfo help documentation available. See those
2550 procedures' docstrings for more information.
2552 `procedure-arguments' describes the arguments that a procedure can take,
2553 combining arity and formals. For example:
2555 (procedure-arguments resolve-interface)
2556 => ((required . (name)) (rest . args))
2558 Additionally, `module-commentary' is now publically exported from
2561 ** Removed: `procedure->memoizing-macro', `procedure->syntax'
2563 These procedures created primitive fexprs for the old evaluator, and are
2564 no longer supported. If you feel that you need these functions, you
2565 probably need to write your own metacircular evaluator (which will
2566 probably be as fast as Guile's, anyway).
2568 ** New language: ECMAScript
2570 Guile now ships with one other high-level language supported,
2571 ECMAScript. The goal is to support all of version 3.1 of the standard,
2572 but not all of the libraries are there yet. This support is not yet
2573 documented; ask on the mailing list if you are interested.
2575 ** New language: Brainfuck
2577 Brainfuck is a toy language that closely models Turing machines. Guile's
2578 brainfuck compiler is meant to be an example of implementing other
2579 languages. See the manual for details, or
2580 http://en.wikipedia.org/wiki/Brainfuck for more information about the
2581 Brainfuck language itself.
2583 ** New language: Elisp
2585 Guile now has an experimental Emacs Lisp compiler and runtime. You can
2586 now switch to Elisp at the repl: `,language elisp'. All kudos to Daniel
2587 Kraft and Brian Templeton, and all bugs to bug-guile@gnu.org.
2589 ** Better documentation infrastructure for macros
2591 It is now possible to introspect on the type of a macro, e.g.
2592 syntax-rules, identifier-syntax, etc, and extract information about that
2593 macro, such as the syntax-rules patterns or the defmacro arguments.
2594 `(texinfo reflection)' takes advantage of this to give better macro
2597 ** Support for arbitrary procedure metadata
2599 Building on its support for docstrings, Guile now supports multiple
2600 docstrings, adding them to the tail of a compiled procedure's
2601 properties. For example:
2607 (procedure-properties foo)
2608 => ((name . foo) (documentation . "one") (documentation . "two"))
2610 Also, vectors of pairs are now treated as additional metadata entries:
2613 #((quz . #f) (docstring . "xyzzy"))
2615 (procedure-properties bar)
2616 => ((name . bar) (quz . #f) (docstring . "xyzzy"))
2618 This allows arbitrary literals to be embedded as metadata in a compiled
2621 ** The psyntax expander now knows how to interpret the @ and @@ special
2624 ** The psyntax expander is now hygienic with respect to modules.
2626 Free variables in a macro are scoped in the module that the macro was
2627 defined in, not in the module the macro is used in. For example, code
2628 like this works now:
2630 (define-module (foo) #:export (bar))
2631 (define (helper x) ...)
2633 (syntax-rules () ((_ x) (helper x))))
2635 (define-module (baz) #:use-module (foo))
2638 It used to be you had to export `helper' from `(foo)' as well.
2639 Thankfully, this has been fixed.
2641 ** Support for version information in Guile's `module' form
2643 Guile modules now have a `#:version' field. See "R6RS Version
2644 References", "General Information about Modules", "Using Guile Modules",
2645 and "Creating Guile Modules" in the manual for more information.
2647 ** Support for renaming bindings on module export
2649 Wherever Guile accepts a symbol as an argument to specify a binding to
2650 export, it now also accepts a pair of symbols, indicating that a binding
2651 should be renamed on export. See "Creating Guile Modules" in the manual
2652 for more information.
2654 ** New procedure: `module-export-all!'
2656 This procedure exports all current and future bindings from a module.
2657 Use as `(module-export-all! (current-module))'.
2659 ** New procedure `reload-module', and `,reload' REPL command
2661 See "Module System Reflection" and "Module Commands" in the manual, for
2664 ** `eval-case' has been deprecated, and replaced by `eval-when'.
2666 The semantics of `eval-when' are easier to understand. See "Eval When"
2667 in the manual, for more information.
2669 ** Guile is now more strict about prohibiting definitions in expression
2672 Although previous versions of Guile accepted it, the following
2673 expression is not valid, in R5RS or R6RS:
2675 (if test (define foo 'bar) (define foo 'baz))
2677 In this specific case, it would be better to do:
2679 (define foo (if test 'bar 'baz))
2681 It is possible to circumvent this restriction with e.g.
2682 `(module-define! (current-module) 'foo 'baz)'. Contact the list if you
2685 ** Support for `letrec*'
2687 Guile now supports `letrec*', a recursive lexical binding operator in
2688 which the identifiers are bound in order. See "Local Bindings" in the
2689 manual, for more details.
2691 ** Internal definitions now expand to `letrec*'
2693 Following the R6RS, internal definitions now expand to letrec* instead
2694 of letrec. The following program is invalid for R5RS, but valid for
2699 (define baz (+ bar 20))
2702 ;; R5RS and Guile <= 1.8:
2703 (foo) => Unbound variable: bar
2704 ;; R6RS and Guile >= 2.0:
2707 This change should not affect correct R5RS programs, or programs written
2708 in earlier Guile dialects.
2710 ** Macro expansion produces structures instead of s-expressions
2712 In the olden days, macroexpanding an s-expression would yield another
2713 s-expression. Though the lexical variables were renamed, expansions of
2714 core forms like `if' and `begin' were still non-hygienic, as they relied
2715 on the toplevel definitions of `if' et al being the conventional ones.
2717 The solution is to expand to structures instead of s-expressions. There
2718 is an `if' structure, a `begin' structure, a `toplevel-ref' structure,
2719 etc. The expander already did this for compilation, producing Tree-IL
2720 directly; it has been changed now to do so when expanding for the
2723 ** Defmacros must now produce valid Scheme expressions.
2725 It used to be that defmacros could unquote in Scheme values, as a way of
2726 supporting partial evaluation, and avoiding some hygiene issues. For
2729 (define (helper x) ...)
2730 (define-macro (foo bar)
2733 Assuming this macro is in the `(baz)' module, the direct translation of
2736 (define (helper x) ...)
2737 (define-macro (foo bar)
2738 `((@@ (baz) helper) ,bar))
2740 Of course, one could just use a hygienic macro instead:
2744 ((_ bar) (helper bar))))
2746 ** Guile's psyntax now supports docstrings and internal definitions.
2748 The following Scheme is not strictly legal:
2755 However its intent is fairly clear. Guile interprets "bar" to be the
2756 docstring of `foo', and the definition of `baz' is still in definition
2759 ** Support for settable identifier syntax
2761 Following the R6RS, "variable transformers" are settable
2762 identifier-syntax. See "Identifier macros" in the manual, for more
2765 ** syntax-case treats `_' as a placeholder
2767 Following R6RS, a `_' in a syntax-rules or syntax-case pattern matches
2768 anything, and binds no pattern variables. Unlike the R6RS, Guile also
2769 permits `_' to be in the literals list for a pattern.
2771 ** Macros need to be defined before their first use.
2773 It used to be that with lazy memoization, this might work:
2777 (define-macro (ref x) x)
2780 But now, the body of `foo' is interpreted to mean a call to the toplevel
2781 `ref' function, instead of a macro expansion. The solution is to define
2782 macros before code that uses them.
2784 ** Functions needed by macros at expand-time need to be present at
2787 For example, this code will work at the REPL:
2789 (define (double-helper x) (* x x))
2790 (define-macro (double-literal x) (double-helper x))
2791 (double-literal 2) => 4
2793 But it will not work when a file is compiled, because the definition of
2794 `double-helper' is not present at expand-time. The solution is to wrap
2795 the definition of `double-helper' in `eval-when':
2797 (eval-when (load compile eval)
2798 (define (double-helper x) (* x x)))
2799 (define-macro (double-literal x) (double-helper x))
2800 (double-literal 2) => 4
2802 See the documentation for eval-when for more information.
2804 ** `macroexpand' produces structures, not S-expressions.
2806 Given the need to maintain referential transparency, both lexically and
2807 modular, the result of expanding Scheme expressions is no longer itself
2808 an s-expression. If you want a human-readable approximation of the
2809 result of `macroexpand', call `tree-il->scheme' from `(language
2812 ** Removed function: `macroexpand-1'
2814 It is unclear how to implement `macroexpand-1' with syntax-case, though
2815 PLT Scheme does prove that it is possible.
2817 ** New reader macros: #' #` #, #,@
2819 These macros translate, respectively, to `syntax', `quasisyntax',
2820 `unsyntax', and `unsyntax-splicing'. See the R6RS for more information.
2821 These reader macros may be overridden by `read-hash-extend'.
2823 ** Incompatible change to #'
2825 Guile did have a #' hash-extension, by default, which just returned the
2826 subsequent datum: #'foo => foo. In the unlikely event that anyone
2827 actually used this, this behavior may be reinstated via the
2828 `read-hash-extend' mechanism.
2830 ** `unquote' and `unquote-splicing' accept multiple expressions
2832 As per the R6RS, these syntax operators can now accept any number of
2833 expressions to unquote.
2835 ** Scheme expresssions may be commented out with #;
2837 #; comments out an entire expression. See SRFI-62 or the R6RS for more
2840 ** Prompts: Delimited, composable continuations
2842 Guile now has prompts as part of its primitive language. See "Prompts"
2843 in the manual, for more information.
2845 Expressions entered in at the REPL, or from the command line, are
2846 surrounded by a prompt with the default prompt tag.
2848 ** `make-stack' with a tail-called procedural narrowing argument no longer
2849 works (with compiled procedures)
2851 It used to be the case that a captured stack could be narrowed to select
2852 calls only up to or from a certain procedure, even if that procedure
2853 already tail-called another procedure. This was because the debug
2854 information from the original procedure was kept on the stack.
2856 Now with the new compiler, the stack only contains active frames from
2857 the current continuation. A narrow to a procedure that is not in the
2858 stack will result in an empty stack. To fix this, narrow to a procedure
2859 that is active in the current continuation, or narrow to a specific
2860 number of stack frames.
2862 ** Backtraces through compiled procedures only show procedures that are
2863 active in the current continuation
2865 Similarly to the previous issue, backtraces in compiled code may be
2866 different from backtraces in interpreted code. There are no semantic
2867 differences, however. Please mail bug-guile@gnu.org if you see any
2868 deficiencies with Guile's backtraces.
2870 ** `positions' reader option enabled by default
2872 This change allows primitive-load without --auto-compile to also
2873 propagate source information through the expander, for better errors and
2874 to let macros know their source locations. The compiler was already
2875 turning it on anyway.
2877 ** New macro: `current-source-location'
2879 The macro returns the current source location (to be documented).
2881 ** syntax-rules and syntax-case macros now propagate source information
2882 through to the expanded code
2884 This should result in better backtraces.
2886 ** The currying behavior of `define' has been removed.
2888 Before, `(define ((f a) b) (* a b))' would translate to
2890 (define f (lambda (a) (lambda (b) (* a b))))
2892 Now a syntax error is signaled, as this syntax is not supported by
2893 default. Use the `(ice-9 curried-definitions)' module to get back the
2896 ** New procedure, `define!'
2898 `define!' is a procedure that takes two arguments, a symbol and a value,
2899 and binds the value to the symbol in the current module. It's useful to
2900 programmatically make definitions in the current module, and is slightly
2901 less verbose than `module-define!'.
2903 ** All modules have names now
2905 Before, you could have anonymous modules: modules without names. Now,
2906 because of hygiene and macros, all modules have names. If a module was
2907 created without a name, the first time `module-name' is called on it, a
2908 fresh name will be lazily generated for it.
2910 ** The module namespace is now separate from the value namespace
2912 It was a little-known implementation detail of Guile's module system
2913 that it was built on a single hierarchical namespace of values -- that
2914 if there was a module named `(foo bar)', then in the module named
2915 `(foo)' there was a binding from `bar' to the `(foo bar)' module.
2917 This was a neat trick, but presented a number of problems. One problem
2918 was that the bindings in a module were not apparent from the module
2919 itself; perhaps the `(foo)' module had a private binding for `bar', and
2920 then an external contributor defined `(foo bar)'. In the end there can
2921 be only one binding, so one of the two will see the wrong thing, and
2922 produce an obtuse error of unclear provenance.
2924 Also, the public interface of a module was also bound in the value
2925 namespace, as `%module-public-interface'. This was a hack from the early
2926 days of Guile's modules.
2928 Both of these warts have been fixed by the addition of fields in the
2929 `module' data type. Access to modules and their interfaces from the
2930 value namespace has been deprecated, and all accessors use the new
2931 record accessors appropriately.
2933 When Guile is built with support for deprecated code, as is the default,
2934 the value namespace is still searched for modules and public interfaces,
2935 and a deprecation warning is raised as appropriate.
2937 Finally, to support lazy loading of modules as one used to be able to do
2938 with module binder procedures, Guile now has submodule binders, called
2939 if a given submodule is not found. See boot-9.scm for more information.
2941 ** New procedures: module-ref-submodule, module-define-submodule,
2942 nested-ref-module, nested-define-module!, local-ref-module,
2945 These new accessors are like their bare variants, but operate on
2946 namespaces instead of values.
2948 ** The (app modules) module tree is officially deprecated
2950 It used to be that one could access a module named `(foo bar)' via
2951 `(nested-ref the-root-module '(app modules foo bar))'. The `(app
2952 modules)' bit was a never-used and never-documented abstraction, and has
2953 been deprecated. See the following mail for a full discussion:
2955 http://lists.gnu.org/archive/html/guile-devel/2010-04/msg00168.html
2957 The `%app' binding is also deprecated.
2959 ** `module-filename' field and accessor
2961 Modules now record the file in which they are defined. This field may be
2962 accessed with the new `module-filename' procedure.
2964 ** Modules load within a known environment
2966 It takes a few procedure calls to define a module, and those procedure
2967 calls need to be in scope. Now we ensure that the current module when
2968 loading a module is one that has the needed bindings, instead of relying
2971 ** `load' is a macro (!) that resolves paths relative to source file dir
2973 The familiar Schem `load' procedure is now a macro that captures the
2974 name of the source file being expanded, and dispatches to the new
2975 `load-in-vicinity'. Referencing `load' by bare name returns a closure
2976 that embeds the current source file name.
2978 This fix allows `load' of relative paths to be resolved with respect to
2979 the location of the file that calls `load'.
2981 ** Many syntax errors have different texts now
2983 Syntax errors still throw to the `syntax-error' key, but the arguments
2984 are often different now. Perhaps in the future, Guile will switch to
2985 using standard SRFI-35 conditions.
2987 ** Returning multiple values to compiled code will silently truncate the
2988 values to the expected number
2990 For example, the interpreter would raise an error evaluating the form,
2991 `(+ (values 1 2) (values 3 4))', because it would see the operands as
2992 being two compound "values" objects, to which `+' does not apply.
2994 The compiler, on the other hand, receives multiple values on the stack,
2995 not as a compound object. Given that it must check the number of values
2996 anyway, if too many values are provided for a continuation, it chooses
2997 to truncate those values, effectively evaluating `(+ 1 3)' instead.
2999 The idea is that the semantics that the compiler implements is more
3000 intuitive, and the use of the interpreter will fade out with time.
3001 This behavior is allowed both by the R5RS and the R6RS.
3003 ** Multiple values in compiled code are not represented by compound
3006 This change may manifest itself in the following situation:
3008 (let ((val (foo))) (do-something) val)
3010 In the interpreter, if `foo' returns multiple values, multiple values
3011 are produced from the `let' expression. In the compiler, those values
3012 are truncated to the first value, and that first value is returned. In
3013 the compiler, if `foo' returns no values, an error will be raised, while
3014 the interpreter would proceed.
3016 Both of these behaviors are allowed by R5RS and R6RS. The compiler's
3017 behavior is more correct, however. If you wish to preserve a potentially
3018 multiply-valued return, you will need to set up a multiple-value
3019 continuation, using `call-with-values'.
3021 ** Defmacros are now implemented in terms of syntax-case.
3023 The practical ramification of this is that the `defmacro?' predicate has
3024 been removed, along with `defmacro-transformer', `macro-table',
3025 `xformer-table', `assert-defmacro?!', `set-defmacro-transformer!' and
3026 `defmacro:transformer'. This is because defmacros are simply macros. If
3027 any of these procedures provided useful facilities to you, we encourage
3028 you to contact the Guile developers.
3030 ** Hygienic macros documented as the primary syntactic extension mechanism.
3032 The macro documentation was finally fleshed out with some documentation
3033 on `syntax-rules' and `syntax-case' macros, and other parts of the macro
3034 expansion process. See "Macros" in the manual, for details.
3036 ** psyntax is now the default expander
3038 Scheme code is now expanded by default by the psyntax hygienic macro
3039 expander. Expansion is performed completely before compilation or
3042 Notably, syntax errors will be signalled before interpretation begins.
3043 In the past, many syntax errors were only detected at runtime if the
3044 code in question was memoized.
3046 As part of its expansion, psyntax renames all lexically-bound
3047 identifiers. Original identifier names are preserved and given to the
3048 compiler, but the interpreter will see the renamed variables, e.g.,
3049 `x432' instead of `x'.
3051 Note that the psyntax that Guile uses is a fork, as Guile already had
3052 modules before incompatible modules were added to psyntax -- about 10
3053 years ago! Thus there are surely a number of bugs that have been fixed
3054 in psyntax since then. If you find one, please notify bug-guile@gnu.org.
3056 ** syntax-rules and syntax-case are available by default.
3058 There is no longer any need to import the `(ice-9 syncase)' module
3059 (which is now deprecated). The expander may be invoked directly via
3060 `macroexpand', though it is normally searched for via the current module
3063 Also, the helper routines for syntax-case are available in the default
3064 environment as well: `syntax->datum', `datum->syntax',
3065 `bound-identifier=?', `free-identifier=?', `generate-temporaries',
3066 `identifier?', and `syntax-violation'. See the R6RS for documentation.
3068 ** Tail patterns in syntax-case
3070 Guile has pulled in some more recent changes from the psyntax portable
3071 syntax expander, to implement support for "tail patterns". Such patterns
3072 are supported by syntax-rules and syntax-case. This allows a syntax-case
3073 match clause to have ellipses, then a pattern at the end. For example:
3076 (syntax-rules (else)
3077 ((_ val match-clause ... (else e e* ...))
3080 Note how there is MATCH-CLAUSE, which is ellipsized, then there is a
3081 tail pattern for the else clause. Thanks to Andreas Rottmann for the
3082 patch, and Kent Dybvig for the code.
3084 ** Lexical bindings introduced by hygienic macros may not be referenced
3085 by nonhygienic macros.
3087 If a lexical binding is introduced by a hygienic macro, it may not be
3088 referenced by a nonhygienic macro. For example, this works:
3091 (define-macro (bind-x val body)
3092 `(let ((x ,val)) ,body))
3093 (define-macro (ref x)
3095 (bind-x 10 (ref x)))
3100 (define-syntax bind-x
3102 ((_ val body) (let ((x val)) body))))
3103 (define-macro (ref x)
3105 (bind-x 10 (ref x)))
3107 It is not normal to run into this situation with existing code. However,
3108 if you have defmacros that expand to hygienic macros, it is possible to
3109 run into situations like this. For example, if you have a defmacro that
3110 generates a `while' expression, the `break' bound by the `while' may not
3111 be visible within other parts of your defmacro. The solution is to port
3112 from defmacros to syntax-rules or syntax-case.
3114 ** Macros may no longer be referenced as first-class values.
3116 In the past, you could evaluate e.g. `if', and get its macro value. Now,
3117 expanding this form raises a syntax error.
3119 Macros still /exist/ as first-class values, but they must be
3120 /referenced/ via the module system, e.g. `(module-ref (current-module)
3123 ** Macros may now have docstrings.
3125 `object-documentation' from `(ice-9 documentation)' may be used to
3126 retrieve the docstring, once you have a macro value -- but see the above
3127 note about first-class macros. Docstrings are associated with the syntax
3128 transformer procedures.
3130 ** `case-lambda' is now available in the default environment.
3132 The binding in the default environment is equivalent to the one from the
3133 `(srfi srfi-16)' module. Use the srfi-16 module explicitly if you wish
3134 to maintain compatibility with Guile 1.8 and earlier.
3136 ** Procedures may now have more than one arity.
3138 This can be the case, for example, in case-lambda procedures. The
3139 arities of compiled procedures may be accessed via procedures from the
3140 `(system vm program)' module; see "Compiled Procedures", "Optional
3141 Arguments", and "Case-lambda" in the manual.
3143 ** Deprecate arity access via (procedure-properties proc 'arity)
3145 Instead of accessing a procedure's arity as a property, use the new
3146 `procedure-minimum-arity' function, which gives the most permissive
3147 arity that the function has, in the same format as the old arity
3150 ** `lambda*' and `define*' are now available in the default environment
3152 As with `case-lambda', `(ice-9 optargs)' continues to be supported, for
3153 compatibility purposes. No semantic change has been made (we hope).
3154 Optional and keyword arguments now dispatch via special VM operations,
3155 without the need to cons rest arguments, making them very fast.
3157 ** New syntax: define-once
3159 `define-once' is like Lisp's `defvar': it creates a toplevel binding,
3160 but only if one does not exist already.
3162 ** New function, `truncated-print', with `format' support
3164 `(ice-9 pretty-print)' now exports `truncated-print', a printer that
3165 will ensure that the output stays within a certain width, truncating the
3166 output in what is hopefully an intelligent manner. See the manual for
3169 There is a new `format' specifier, `~@y', for doing a truncated
3170 print (as opposed to `~y', which does a pretty-print). See the `format'
3171 documentation for more details.
3173 ** Better pretty-printing
3175 Indentation recognizes more special forms, like `syntax-case', and read
3176 macros like `quote' are printed better.
3178 ** Passing a number as the destination of `format' is deprecated
3180 The `format' procedure in `(ice-9 format)' now emits a deprecation
3181 warning if a number is passed as its first argument.
3183 Also, it used to be that you could omit passing a port to `format', in
3184 some cases. This still works, but has been formally deprecated.
3186 ** SRFI-4 vectors reimplemented in terms of R6RS bytevectors
3188 Guile now implements SRFI-4 vectors using bytevectors. Often when you
3189 have a numeric vector, you end up wanting to write its bytes somewhere,
3190 or have access to the underlying bytes, or read in bytes from somewhere
3191 else. Bytevectors are very good at this sort of thing. But the SRFI-4
3192 APIs are nicer to use when doing number-crunching, because they are
3193 addressed by element and not by byte.
3195 So as a compromise, Guile allows all bytevector functions to operate on
3196 numeric vectors. They address the underlying bytes in the native
3197 endianness, as one would expect.
3199 Following the same reasoning, that it's just bytes underneath, Guile
3200 also allows uniform vectors of a given type to be accessed as if they
3201 were of any type. One can fill a u32vector, and access its elements with
3202 u8vector-ref. One can use f64vector-ref on bytevectors. It's all the
3205 In this way, uniform numeric vectors may be written to and read from
3206 input/output ports using the procedures that operate on bytevectors.
3208 Calls to SRFI-4 accessors (ref and set functions) from Scheme are now
3209 inlined to the VM instructions for bytevector access.
3211 See "SRFI-4" in the manual, for more information.
3213 ** Nonstandard SRFI-4 procedures now available from `(srfi srfi-4 gnu)'
3215 Guile's `(srfi srfi-4)' now only exports those srfi-4 procedures that
3216 are part of the standard. Complex uniform vectors and the
3217 `any->FOOvector' family are now available only from `(srfi srfi-4 gnu)'.
3219 Guile's default environment imports `(srfi srfi-4)', and probably should
3220 import `(srfi srfi-4 gnu)' as well.
3222 See "SRFI-4 Extensions" in the manual, for more information.
3224 ** New syntax: include-from-path.
3226 `include-from-path' is like `include', except it looks for its file in
3227 the load path. It can be used to compile other files into a file.
3229 ** New syntax: quasisyntax.
3231 `quasisyntax' is to `syntax' as `quasiquote' is to `quote'. See the R6RS
3232 documentation for more information. Thanks to Andre van Tonder for the
3235 ** `*unspecified*' is identifier syntax
3237 `*unspecified*' is no longer a variable, so it is optimized properly by
3238 the compiler, and is not `set!'-able.
3240 ** Changes and bugfixes in numerics code
3242 *** Added six new sets of fast quotient and remainder operators
3244 Added six new sets of fast quotient and remainder operator pairs with
3245 different semantics than the R5RS operators. They support not only
3246 integers, but all reals, including exact rationals and inexact
3247 floating point numbers.
3249 These procedures accept two real numbers N and D, where the divisor D
3250 must be non-zero. Each set of operators computes an integer quotient
3251 Q and a real remainder R such that N = Q*D + R and |R| < |D|. They
3252 differ only in how N/D is rounded to produce Q.
3254 `euclidean-quotient' returns the integer Q and `euclidean-remainder'
3255 returns the real R such that N = Q*D + R and 0 <= R < |D|. `euclidean/'
3256 returns both Q and R, and is more efficient than computing each
3257 separately. Note that when D > 0, `euclidean-quotient' returns
3258 floor(N/D), and when D < 0 it returns ceiling(N/D).
3260 `centered-quotient', `centered-remainder', and `centered/' are similar
3261 except that the range of remainders is -abs(D/2) <= R < abs(D/2), and
3262 `centered-quotient' rounds N/D to the nearest integer. Note that these
3263 operators are equivalent to the R6RS integer division operators `div',
3264 `mod', `div-and-mod', `div0', `mod0', and `div0-and-mod0'.
3266 `floor-quotient' and `floor-remainder' compute Q and R, respectively,
3267 where Q has been rounded toward negative infinity. `floor/' returns
3268 both Q and R, and is more efficient than computing each separately.
3269 Note that when applied to integers, `floor-remainder' is equivalent to
3270 the R5RS integer-only `modulo' operator. `ceiling-quotient',
3271 `ceiling-remainder', and `ceiling/' are similar except that Q is
3272 rounded toward positive infinity.
3274 For `truncate-quotient', `truncate-remainder', and `truncate/', Q is
3275 rounded toward zero. Note that when applied to integers,
3276 `truncate-quotient' and `truncate-remainder' are equivalent to the
3277 R5RS integer-only operators `quotient' and `remainder'.
3279 For `round-quotient', `round-remainder', and `round/', Q is rounded to
3280 the nearest integer, with ties going to the nearest even integer.
3282 *** Complex number changes
3284 Guile is now able to represent non-real complex numbers whose
3285 imaginary part is an _inexact_ zero (0.0 or -0.0), per R6RS.
3286 Previously, such numbers were immediately changed into inexact reals.
3288 (real? 0.0+0.0i) now returns #f, per R6RS, although (zero? 0.0+0.0i)
3289 still returns #t, per R6RS. (= 0 0.0+0.0i) and (= 0.0 0.0+0.0i) are
3290 #t, but the same comparisons using `eqv?' or `equal?' are #f.
3292 Like other non-real numbers, these complex numbers with inexact zero
3293 imaginary part will raise exceptions is passed to procedures requiring
3294 reals, such as `<', `>', `<=', `>=', `min', `max', `positive?',
3295 `negative?', `inf?', `nan?', `finite?', etc.
3297 **** `make-rectangular' changes
3299 scm_make_rectangular `make-rectangular' now returns a real number only
3300 if the imaginary part is an _exact_ 0. Previously, it would return a
3301 real number if the imaginary part was an inexact zero.
3303 scm_c_make_rectangular now always returns a non-real complex number,
3304 even if the imaginary part is zero. Previously, it would return a
3305 real number if the imaginary part was zero.
3307 **** `make-polar' changes
3309 scm_make_polar `make-polar' now returns a real number only if the
3310 angle or magnitude is an _exact_ 0. If the magnitude is an exact 0,
3311 it now returns an exact 0. Previously, it would return a real
3312 number if the imaginary part was an inexact zero.
3314 scm_c_make_polar now always returns a non-real complex number, even if
3315 the imaginary part is 0.0. Previously, it would return a real number
3316 if the imaginary part was 0.0.
3318 **** `imag-part' changes
3320 scm_imag_part `imag-part' now returns an exact 0 if applied to an
3321 inexact real number. Previously it returned an inexact zero in this
3324 *** `eqv?' and `equal?' now compare numbers equivalently
3326 scm_equal_p `equal?' now behaves equivalently to scm_eqv_p `eqv?' for
3327 numeric values, per R5RS. Previously, equal? worked differently,
3328 e.g. `(equal? 0.0 -0.0)' returned #t but `(eqv? 0.0 -0.0)' returned #f,
3329 and `(equal? +nan.0 +nan.0)' returned #f but `(eqv? +nan.0 +nan.0)'
3332 *** `(equal? +nan.0 +nan.0)' now returns #t
3334 Previously, `(equal? +nan.0 +nan.0)' returned #f, although
3335 `(let ((x +nan.0)) (equal? x x))' and `(eqv? +nan.0 +nan.0)'
3336 both returned #t. R5RS requires that `equal?' behave like
3337 `eqv?' when comparing numbers.
3339 *** Change in handling products `*' involving exact 0
3341 scm_product `*' now handles exact 0 differently. A product containing
3342 an exact 0 now returns an exact 0 if and only if the other arguments
3343 are all exact. An inexact zero is returned if and only if the other
3344 arguments are all finite but not all exact. If an infinite or NaN
3345 value is present, a NaN value is returned. Previously, any product
3346 containing an exact 0 yielded an exact 0, regardless of the other
3349 *** `expt' and `integer-expt' changes when the base is 0
3351 While `(expt 0 0)' is still 1, and `(expt 0 N)' for N > 0 is still
3352 zero, `(expt 0 N)' for N < 0 is now a NaN value, and likewise for
3353 integer-expt. This is more correct, and conforming to R6RS, but seems
3354 to be incompatible with R5RS, which would return 0 for all non-zero
3357 *** `expt' and `integer-expt' are more generic, less strict
3359 When raising to an exact non-negative integer exponent, `expt' and
3360 `integer-expt' are now able to exponentiate any object that can be
3361 multiplied using `*'. They can also raise an object to an exact
3362 negative integer power if its reciprocal can be taken using `/'.
3363 In order to allow this, the type of the first argument is no longer
3364 checked when raising to an exact integer power. If the exponent is 0
3365 or 1, the first parameter is not manipulated at all, and need not
3366 even support multiplication.
3368 *** Infinities are no longer integers, nor rationals
3370 scm_integer_p `integer?' and scm_rational_p `rational?' now return #f
3371 for infinities, per R6RS. Previously they returned #t for real
3372 infinities. The real infinities and NaNs are still considered real by
3373 scm_real `real?' however, per R6RS.
3375 *** NaNs are no longer rationals
3377 scm_rational_p `rational?' now returns #f for NaN values, per R6RS.
3378 Previously it returned #t for real NaN values. They are still
3379 considered real by scm_real `real?' however, per R6RS.
3381 *** `inf?' and `nan?' now throw exceptions for non-reals
3383 The domain of `inf?' and `nan?' is the real numbers. Guile now signals
3384 an error when a non-real number or non-number is passed to these
3385 procedures. (Note that NaNs _are_ considered numbers by scheme, despite
3388 *** `rationalize' bugfixes and changes
3390 Fixed bugs in scm_rationalize `rationalize'. Previously, it returned
3391 exact integers unmodified, although that was incorrect if the epsilon
3392 was at least 1 or inexact, e.g. (rationalize 4 1) should return 3 per
3393 R5RS and R6RS, but previously it returned 4. It also now handles
3394 cases involving infinities and NaNs properly, per R6RS.
3396 *** Trigonometric functions now return exact numbers in some cases
3398 scm_sin `sin', scm_cos `cos', scm_tan `tan', scm_asin `asin', scm_acos
3399 `acos', scm_atan `atan', scm_sinh `sinh', scm_cosh `cosh', scm_tanh
3400 `tanh', scm_sys_asinh `asinh', scm_sys_acosh `acosh', and
3401 scm_sys_atanh `atanh' now return exact results in some cases.
3403 *** New procedure: `finite?'
3405 Add scm_finite_p `finite?' from R6RS to guile core, which returns #t
3406 if and only if its argument is neither infinite nor a NaN. Note that
3407 this is not the same as (not (inf? x)) or (not (infinite? x)), since
3408 NaNs are neither finite nor infinite.
3410 *** Improved exactness handling for complex number parsing
3412 When parsing non-real complex numbers, exactness specifiers are now
3413 applied to each component, as is done in PLT Scheme. For complex
3414 numbers written in rectangular form, exactness specifiers are applied
3415 to the real and imaginary parts before calling scm_make_rectangular.
3416 For complex numbers written in polar form, exactness specifiers are
3417 applied to the magnitude and angle before calling scm_make_polar.
3419 Previously, exactness specifiers were applied to the number as a whole
3420 _after_ calling scm_make_rectangular or scm_make_polar.
3422 For example, (string->number "#i5.0+0i") now does the equivalent of:
3424 (make-rectangular (exact->inexact 5.0) (exact->inexact 0))
3426 which yields 5.0+0.0i. Previously it did the equivalent of:
3428 (exact->inexact (make-rectangular 5.0 0))
3432 ** Unicode characters
3434 Unicode characters may be entered in octal format via e.g. `#\454', or
3435 created via (integer->char 300). A hex external representation will
3436 probably be introduced at some point.
3440 Internally, strings are now represented either in the `latin-1'
3441 encoding, one byte per character, or in UTF-32, with four bytes per
3442 character. Strings manage their own allocation, switching if needed.
3444 Extended characters may be written in a literal string using the
3445 hexadecimal escapes `\xXX', `\uXXXX', or `\UXXXXXX', for 8-bit, 16-bit,
3446 or 24-bit codepoints, respectively, or entered directly in the native
3447 encoding of the port on which the string is read.
3451 One may now use U+03BB (GREEK SMALL LETTER LAMBDA) as an identifier.
3453 ** Support for non-ASCII source code files
3455 The default reader now handles source code files for some of the
3456 non-ASCII character encodings, such as UTF-8. A non-ASCII source file
3457 should have an encoding declaration near the top of the file. Also,
3458 there is a new function, `file-encoding', that scans a port for a coding
3459 declaration. See the section of the manual entitled, "Character Encoding
3462 The pre-1.9.3 reader handled 8-bit clean but otherwise unspecified source
3463 code. This use is now discouraged. Binary input and output is
3464 currently supported by opening ports in the ISO-8859-1 locale.
3466 ** Source files default to UTF-8.
3468 If source files do not specify their encoding via a `coding:' block,
3469 the default encoding is UTF-8, instead of being taken from the current
3472 ** Interactive Guile installs the current locale.
3474 Instead of leaving the user in the "C" locale, running the Guile REPL
3475 installs the current locale. [FIXME xref?]
3477 ** Support for locale transcoding when reading from and writing to ports
3479 Ports now have an associated character encoding, and port read and write
3480 operations do conversion to and from locales automatically. Ports also
3481 have an associated strategy for how to deal with locale conversion
3484 See the documentation in the manual for the four new support functions,
3485 `set-port-encoding!', `port-encoding', `set-port-conversion-strategy!',
3486 and `port-conversion-strategy'.
3488 ** String and SRFI-13 functions can operate on Unicode strings
3490 ** Unicode support for SRFI-14 character sets
3492 The default character sets are no longer locale dependent and contain
3493 characters from the whole Unicode range. There is a new predefined
3494 character set, `char-set:designated', which contains all assigned
3495 Unicode characters. There is a new debugging function, `%char-set-dump'.
3497 ** Character functions operate on Unicode characters
3499 `char-upcase' and `char-downcase' use default Unicode casing rules.
3500 Character comparisons such as `char<?' and `char-ci<?' now sort based on
3501 Unicode code points.
3503 ** Global variables `scm_charnames' and `scm_charnums' are removed
3505 These variables contained the names of control characters and were
3506 used when writing characters. While these were global, they were
3507 never intended to be public API. They have been replaced with private
3510 ** EBCDIC support is removed
3512 There was an EBCDIC compile flag that altered some of the character
3513 processing. It appeared that full EBCDIC support was never completed
3514 and was unmaintained.
3516 ** Compile-time warnings
3518 Guile can warn about potentially unbound free variables. Pass the
3519 -Wunbound-variable on the `guile-tools compile' command line, or add
3520 `#:warnings '(unbound-variable)' to your `compile' or `compile-file'
3521 invocation. Warnings are also enabled by default for expressions entered
3524 Guile can also warn when you pass the wrong number of arguments to a
3525 procedure, with -Warity-mismatch, or `arity-mismatch' in the
3526 `#:warnings' as above.
3528 Other warnings include `-Wunused-variable' and `-Wunused-toplevel', to
3529 warn about unused local or global (top-level) variables, and `-Wformat',
3530 to check for various errors related to the `format' procedure.
3532 ** A new `memoize-symbol' evaluator trap has been added.
3534 This trap can be used for efficiently implementing a Scheme code
3537 ** Duplicate bindings among used modules are resolved lazily.
3539 This slightly improves program startup times.
3541 ** New thread cancellation and thread cleanup API
3543 See `cancel-thread', `set-thread-cleanup!', and `thread-cleanup'.
3545 ** New threads are in `(guile-user)' by default, not `(guile)'
3547 It used to be that a new thread entering Guile would do so in the
3548 `(guile)' module, unless this was the first time Guile was initialized,
3549 in which case it was `(guile-user)'. This has been fixed to have all
3550 new threads unknown to Guile default to `(guile-user)'.
3552 ** New helpers: `print-exception', `set-exception-printer!'
3554 These functions implement an extensible exception printer. Guile
3555 registers printers for all of the exceptions it throws. Users may add
3556 their own printers. There is also `scm_print_exception', for use by C
3557 programs. Pleasantly, this allows SRFI-35 and R6RS exceptions to be
3558 printed appropriately.
3560 ** GOOPS dispatch in scheme
3562 As an implementation detail, GOOPS dispatch is no longer implemented by
3563 special evaluator bytecodes, but rather directly via a Scheme function
3564 associated with an applicable struct. There is some VM support for the
3565 underlying primitives, like `class-of'.
3567 This change will in the future allow users to customize generic function
3568 dispatch without incurring a performance penalty, and allow us to
3569 implement method combinations.
3571 ** Applicable struct support
3573 One may now make structs from Scheme that may be applied as procedures.
3574 To do so, make a struct whose vtable is `<applicable-struct-vtable>'.
3575 That struct will be the vtable of your applicable structs; instances of
3576 that new struct are assumed to have the procedure in their first slot.
3577 `<applicable-struct-vtable>' is like Common Lisp's
3578 `funcallable-standard-class'. Likewise there is
3579 `<applicable-struct-with-setter-vtable>', which looks for the setter in
3580 the second slot. This needs to be better documented.
3584 GOOPS had a number of concepts that were relevant to the days of Tcl,
3585 but not any more: operators and entities, mainly. These objects were
3586 never documented, and it is unlikely that they were ever used. Operators
3587 were a kind of generic specific to the Tcl support. Entities were
3588 replaced by applicable structs, mentioned above.
3590 ** New struct slot allocation: "hidden"
3592 A hidden slot is readable and writable, but will not be initialized by a
3593 call to make-struct. For example in your layout you would say "ph"
3594 instead of "pw". Hidden slots are useful for adding new slots to a
3595 vtable without breaking existing invocations to make-struct.
3597 ** eqv? not a generic
3599 One used to be able to extend `eqv?' as a primitive-generic, but no
3600 more. Because `eqv?' is in the expansion of `case' (via `memv'), which
3601 should be able to compile to static dispatch tables, it doesn't make
3602 sense to allow extensions that would subvert this optimization.
3604 ** `inet-ntop' and `inet-pton' are always available.
3606 Guile now use a portable implementation of `inet_pton'/`inet_ntop', so
3607 there is no more need to use `inet-aton'/`inet-ntoa'. The latter
3608 functions are deprecated.
3610 ** `getopt-long' parsing errors throw to `quit', not `misc-error'
3612 This change should inhibit backtraces on argument parsing errors.
3613 `getopt-long' has been modified to print out the error that it throws
3616 ** New primitive: `tmpfile'.
3618 See "File System" in the manual.
3620 ** Random generator state may be serialized to a datum
3622 `random-state->datum' will serialize a random state to a datum, which
3623 may be written out, read back in later, and revivified using
3624 `datum->random-state'. See "Random" in the manual, for more details.
3626 ** Fix random number generator on 64-bit platforms
3628 There was a nasty bug on 64-bit platforms in which asking for a random
3629 integer with a range between 2**32 and 2**64 caused a segfault. After
3630 many embarrassing iterations, this was fixed.
3632 ** Fast bit operations.
3634 The bit-twiddling operations `ash', `logand', `logior', and `logxor' now
3635 have dedicated bytecodes. Guile is not just for symbolic computation,
3636 it's for number crunching too.
3638 ** Faster SRFI-9 record access
3640 SRFI-9 records are now implemented directly on top of Guile's structs,
3641 and their accessors are defined in such a way that normal call-sites
3642 inline to special VM opcodes, while still allowing for the general case
3643 (e.g. passing a record accessor to `apply').
3645 ** R6RS block comment support
3647 Guile now supports R6RS nested block comments. The start of a comment is
3648 marked with `#|', and the end with `|#'.
3650 ** `guile-2' cond-expand feature
3652 To test if your code is running under Guile 2.0 (or its alpha releases),
3653 test for the `guile-2' cond-expand feature. Like this:
3655 (cond-expand (guile-2 (eval-when (compile)
3656 ;; This must be evaluated at compile time.
3657 (fluid-set! current-reader my-reader)))
3659 ;; Earlier versions of Guile do not have a
3660 ;; separate compilation phase.
3661 (fluid-set! current-reader my-reader)))
3663 ** New global variables: %load-compiled-path, %load-compiled-extensions
3665 These are analogous to %load-path and %load-extensions.
3667 ** New fluid: `%file-port-name-canonicalization'
3669 This fluid parameterizes the file names that are associated with file
3670 ports. If %file-port-name-canonicalization is 'absolute, then file names
3671 are canonicalized to be absolute paths. If it is 'relative, then the
3672 name is canonicalized, but any prefix corresponding to a member of
3673 `%load-path' is stripped off. Otherwise the names are passed through
3676 In addition, the `compile-file' and `compile-and-load' procedures bind
3677 %file-port-name-canonicalization to their `#:canonicalization' keyword
3678 argument, which defaults to 'relative. In this way, one might compile
3679 "../module/ice-9/boot-9.scm", but the path that gets residualized into
3680 the .go is "ice-9/boot-9.scm".
3682 ** New procedure, `make-promise'
3684 `(make-promise (lambda () foo))' is equivalent to `(delay foo)'.
3686 ** `defined?' may accept a module as its second argument
3688 Previously it only accepted internal structures from the evaluator.
3690 ** New entry into %guile-build-info: `ccachedir'
3692 ** Fix bug in `module-bound?'.
3694 `module-bound?' was returning true if a module did have a local
3695 variable, but one that was unbound, but another imported module bound
3696 the variable. This was an error, and was fixed.
3698 ** `(ice-9 syncase)' has been deprecated.
3700 As syntax-case is available by default, importing `(ice-9 syncase)' has
3701 no effect, and will trigger a deprecation warning.
3703 ** New readline history functions
3705 The (ice-9 readline) module now provides add-history, read-history,
3706 write-history and clear-history, which wrap the corresponding GNU
3707 History library functions.
3709 ** Removed deprecated uniform array procedures:
3710 dimensions->uniform-array, list->uniform-array, array-prototype
3712 Instead, use make-typed-array, list->typed-array, or array-type,
3715 ** Deprecate the old `scm-style-repl'
3717 The following bindings from boot-9 are now found in `(ice-9
3718 scm-style-repl)': `scm-style-repl', `error-catching-loop',
3719 `error-catching-repl', `bad-throw', `scm-repl-silent'
3720 `assert-repl-silence', `repl-print-unspecified',
3721 `assert-repl-print-unspecified', `scm-repl-verbose',
3722 `assert-repl-verbosity', `scm-repl-prompt', `set-repl-prompt!', `repl',
3723 `default-pre-unwind-handler', `handle-system-error',
3725 The following bindings have been deprecated, with no replacement:
3726 `pre-unwind-handler-dispatch'.
3728 The following bindings have been totally removed:
3729 `before-signal-stack'.
3731 Deprecated forwarding shims have been installed so that users that
3732 expect these bindings in the main namespace will still work, but receive
3733 a deprecation warning.
3735 ** `set-batch-mode?!' replaced by `ensure-batch-mode!'
3737 "Batch mode" is a flag used to tell a program that it is not running
3738 interactively. One usually turns it on after a fork. It may not be
3739 turned off. `ensure-batch-mode!' deprecates the old `set-batch-mode?!',
3740 because it is a better interface, as it can only turn on batch mode, not
3743 ** Deprecate `save-stack', `the-last-stack'
3745 It used to be that the way to debug programs in Guile was to capture the
3746 stack at the time of error, drop back to the REPL, then debug that
3747 stack. But this approach didn't compose, was tricky to get right in the
3748 presence of threads, and was not very powerful.
3750 So `save-stack', `stack-saved?', and `the-last-stack' have been moved to
3751 `(ice-9 save-stack)', with deprecated bindings left in the root module.
3753 ** `top-repl' has its own module
3755 The `top-repl' binding, called with Guile is run interactively, is now
3756 is its own module, `(ice-9 top-repl)'. A deprecated forwarding shim was
3757 left in the default environment.
3759 ** `display-error' takes a frame
3761 The `display-error' / `scm_display_error' helper now takes a frame as an
3762 argument instead of a stack. Stacks are still supported in deprecated
3763 builds. Additionally, `display-error' will again source location
3764 information for the error.
3766 ** No more `(ice-9 debug)'
3768 This module had some debugging helpers that are no longer applicable to
3769 the current debugging model. Importing this module will produce a
3770 deprecation warning. Users should contact bug-guile for support.
3772 ** Remove obsolete debug-options
3774 Removed `breakpoints', `trace', `procnames', `indent', `frames',
3775 `maxdepth', and `debug' debug-options.
3777 ** `backtrace' debug option on by default
3779 Given that Guile 2.0 can always give you a backtrace, backtraces are now
3782 ** `turn-on-debugging' deprecated
3784 ** Remove obsolete print-options
3786 The `source' and `closure-hook' print options are obsolete, and have
3789 ** Remove obsolete read-options
3791 The "elisp-strings" and "elisp-vectors" read options were unused and
3792 obsolete, so they have been removed.
3794 ** Remove eval-options and trap-options
3796 Eval-options and trap-options are obsolete with the new VM and
3799 ** Remove (ice-9 debugger) and (ice-9 debugging)
3801 See "Traps" and "Interactive Debugging" in the manual, for information
3802 on their replacements.
3804 ** Remove the GDS Emacs integration
3806 See "Using Guile in Emacs" in the manual, for info on how we think you
3807 should use Guile with Emacs.
3809 ** Deprecated: `lazy-catch'
3811 `lazy-catch' was a form that captured the stack at the point of a
3812 `throw', but the dynamic state at the point of the `catch'. It was a bit
3813 crazy. Please change to use `catch', possibly with a throw-handler, or
3814 `with-throw-handler'.
3816 ** Deprecated: primitive properties
3818 The `primitive-make-property', `primitive-property-set!',
3819 `primitive-property-ref', and `primitive-property-del!' procedures were
3820 crufty and only used to implement object properties, which has a new,
3821 threadsafe implementation. Use object properties or weak hash tables
3824 ** Deprecated `@bind' syntax
3826 `@bind' was part of an older implementation of the Emacs Lisp language,
3827 and is no longer used.
3829 ** Miscellaneous other deprecations
3831 `cuserid' has been deprecated, as it only returns 8 bytes of a user's
3832 login. Use `(passwd:name (getpwuid (geteuid)))' instead.
3834 Additionally, the procedures `apply-to-args', `has-suffix?', `scheme-file-suffix'
3835 `get-option', `for-next-option', `display-usage-report',
3836 `transform-usage-lambda', `collect', and `set-batch-mode?!' have all
3839 ** Add support for unbound fluids
3841 See `make-unbound-fluid', `fluid-unset!', and `fluid-bound?' in the
3844 ** Add `variable-unset!'
3846 See "Variables" in the manual, for more details.
3848 ** Last but not least, the `λ' macro can be used in lieu of `lambda'
3850 * Changes to the C interface
3852 ** Guile now uses libgc, the Boehm-Demers-Weiser garbage collector
3854 The semantics of `scm_gc_malloc ()' have been changed, in a
3855 backward-compatible way. A new allocation routine,
3856 `scm_gc_malloc_pointerless ()', was added.
3858 Libgc is a conservative GC, which we hope will make interaction with C
3859 code easier and less error-prone.
3861 ** New procedures: `scm_to_stringn', `scm_from_stringn'
3862 ** New procedures: scm_{to,from}_{utf8,latin1}_symbol{n,}
3863 ** New procedures: scm_{to,from}_{utf8,utf32,latin1}_string{n,}
3865 These new procedures convert to and from string representations in
3866 particular encodings.
3868 Users should continue to use locale encoding for user input, user
3869 output, or interacting with the C library.
3871 Use the Latin-1 functions for ASCII, and for literals in source code.
3873 Use UTF-8 functions for interaction with modern libraries which deal in
3874 UTF-8, and UTF-32 for interaction with utf32-using libraries.
3876 Otherwise, use scm_to_stringn or scm_from_stringn with a specific
3879 ** New type definitions for `scm_t_intptr' and friends.
3881 `SCM_T_UINTPTR_MAX', `SCM_T_INTPTR_MIN', `SCM_T_INTPTR_MAX',
3882 `SIZEOF_SCM_T_BITS', `scm_t_intptr' and `scm_t_uintptr' are now
3883 available to C. Have fun!
3885 ** The GH interface (deprecated in version 1.6, 2001) was removed.
3887 ** Internal `scm_i_' functions now have "hidden" linkage with GCC/ELF
3889 This makes these internal functions technically not callable from
3892 ** Functions for handling `scm_option' now no longer require an argument
3893 indicating length of the `scm_t_option' array.
3895 ** Procedures-with-setters are now implemented using applicable structs
3897 From a user's perspective this doesn't mean very much. But if, for some
3898 odd reason, you used the SCM_PROCEDURE_WITH_SETTER_P, SCM_PROCEDURE, or
3899 SCM_SETTER macros, know that they're deprecated now. Also, scm_tc7_pws
3902 ** Remove old evaluator closures
3904 There used to be ranges of typecodes allocated to interpreted data
3905 structures, but that it no longer the case, given that interpreted
3906 procedure are now just regular VM closures. As a result, there is a
3907 newly free tc3, and a number of removed macros. See the ChangeLog for
3910 ** Primitive procedures are now VM trampoline procedures
3912 It used to be that there were something like 12 different typecodes
3913 allocated to primitive procedures, each with its own calling convention.
3914 Now there is only one, the gsubr. This may affect user code if you were
3915 defining a procedure using scm_c_make_subr rather scm_c_make_gsubr. The
3916 solution is to switch to use scm_c_make_gsubr. This solution works well
3917 both with the old 1.8 and with the current 1.9 branch.
3919 Guile's old evaluator used to have special cases for applying "gsubrs",
3920 primitive procedures with specified numbers of required, optional, and
3921 rest arguments. Now, however, Guile represents gsubrs as normal VM
3922 procedures, with appropriate bytecode to parse out the correct number of
3923 arguments, including optional and rest arguments, and then with a
3924 special bytecode to apply the gsubr.
3926 This allows primitive procedures to appear on the VM stack, allowing
3927 them to be accurately counted in profiles. Also they now have more
3928 debugging information attached to them -- their number of arguments, for
3929 example. In addition, the VM can completely inline the application
3930 mechanics, allowing for faster primitive calls.
3932 However there are some changes on the C level. There is no more
3933 `scm_tc7_gsubr' or `scm_tcs_subrs' typecode for primitive procedures, as
3934 they are just VM procedures. Likewise the macros `SCM_GSUBR_TYPE',
3935 `SCM_GSUBR_MAKTYPE', `SCM_GSUBR_REQ', `SCM_GSUBR_OPT', and
3936 `SCM_GSUBR_REST' are gone, as are `SCM_SUBR_META_INFO', `SCM_SUBR_PROPS'
3937 `SCM_SET_SUBR_GENERIC_LOC', and `SCM_SUBR_ARITY_TO_TYPE'.
3939 Perhaps more significantly, `scm_c_make_subr',
3940 `scm_c_make_subr_with_generic', `scm_c_define_subr', and
3941 `scm_c_define_subr_with_generic'. They all operated on subr typecodes,
3942 and there are no more subr typecodes. Use the scm_c_make_gsubr family
3945 Normal users of gsubrs should not be affected, though, as the
3946 scm_c_make_gsubr family still is the correct way to create primitive
3949 ** Remove deprecated array C interfaces
3951 Removed the deprecated array functions `scm_i_arrayp',
3952 `scm_i_array_ndim', `scm_i_array_mem', `scm_i_array_v',
3953 `scm_i_array_base', `scm_i_array_dims', and the deprecated macros
3954 `SCM_ARRAYP', `SCM_ARRAY_NDIM', `SCM_ARRAY_CONTP', `SCM_ARRAY_MEM',
3955 `SCM_ARRAY_V', `SCM_ARRAY_BASE', and `SCM_ARRAY_DIMS'.
3957 ** Remove unused snarf macros
3959 `SCM_DEFINE1', `SCM_PRIMITIVE_GENERIC_1', `SCM_PROC1, and `SCM_GPROC1'
3960 are no more. Use SCM_DEFINE or SCM_PRIMITIVE_GENERIC instead.
3962 ** New functions: `scm_call_n', `scm_c_run_hookn'
3964 `scm_call_n' applies to apply a function to an array of arguments.
3965 `scm_c_run_hookn' runs a hook with an array of arguments.
3967 ** Some SMOB types changed to have static typecodes
3969 Fluids, dynamic states, and hash tables used to be SMOB objects, but now
3970 they have statically allocated tc7 typecodes.
3972 ** Preparations for changing SMOB representation
3974 If things go right, we'll be changing the SMOB representation soon. To
3975 that end, we did a lot of cleanups to calls to e.g. SCM_CELL_WORD_2(x) when
3976 the code meant SCM_SMOB_DATA_2(x); user code will need similar changes
3977 in the future. Code accessing SMOBs using SCM_CELL macros was never
3978 correct, but until now things still worked. Users should be aware of
3981 ** Changed invocation mechanics of applicable SMOBs
3983 Guile's old evaluator used to have special cases for applying SMOB
3984 objects. Now, with the VM, when Guile sees a SMOB, it looks up a VM
3985 trampoline procedure for it, and use the normal mechanics to apply the
3986 trampoline. This simplifies procedure application in the normal,
3989 The upshot is that the mechanics used to apply a SMOB are different from
3990 1.8. Descriptors no longer have `apply_0', `apply_1', `apply_2', and
3991 `apply_3' functions, and the macros SCM_SMOB_APPLY_0 and friends are now
3992 deprecated. Just use the scm_call_0 family of procedures.
3994 ** Removed support shlibs for SRFIs 1, 4, 13, 14, and 60
3996 Though these SRFI support libraries did expose API, they encoded a
3997 strange version string into their library names. That version was never
3998 programmatically exported, so there was no way people could use the
4001 This was a fortunate oversight, as it allows us to remove the need for
4002 extra, needless shared libraries --- the C support code for SRFIs 4, 13,
4003 and 14 was already in core --- and allow us to incrementally return the
4004 SRFI implementation to Scheme.
4006 ** New C function: scm_module_public_interface
4008 This procedure corresponds to Scheme's `module-public-interface'.
4010 ** Undeprecate `scm_the_root_module ()'
4012 It's useful to be able to get the root module from C without doing a
4015 ** Inline vector allocation
4017 Instead of having vectors point out into the heap for their data, their
4018 data is now allocated inline to the vector object itself. The same is
4019 true for bytevectors, by default, though there is an indirection
4020 available which should allow for making a bytevector from an existing
4023 ** New struct constructors that don't involve making lists
4025 `scm_c_make_struct' and `scm_c_make_structv' are new varargs and array
4026 constructors, respectively, for structs. You might find them useful.
4030 In Guile 1.8, there were debugging frames on the C stack. Now there is
4031 no more need to explicitly mark the stack in this way, because Guile has
4032 a VM stack that it knows how to walk, which simplifies the C API
4033 considerably. See the ChangeLog for details; the relevant interface is
4034 in libguile/stacks.h. The Scheme API has not been changed significantly.
4036 ** Removal of Guile's primitive object system.
4038 There were a number of pieces in `objects.[ch]' that tried to be a
4039 minimal object system, but were never documented, and were quickly
4040 obseleted by GOOPS' merge into Guile proper. So `scm_make_class_object',
4041 `scm_make_subclass_object', `scm_metaclass_standard', and like symbols
4042 from objects.h are no more. In the very unlikely case in which these
4043 were useful to you, we urge you to contact guile-devel.
4047 Actually the future is still in the state that it was, is, and ever
4048 shall be, Amen, except that `futures.c' and `futures.h' are no longer a
4049 part of it. These files were experimental, never compiled, and would be
4050 better implemented in Scheme anyway. In the future, that is.
4052 ** Deprecate trampolines
4054 There used to be C functions `scm_trampoline_0', `scm_trampoline_1', and
4055 so on. The point was to do some precomputation on the type of the
4056 procedure, then return a specialized "call" procedure. However this
4057 optimization wasn't actually an optimization, so it is now deprecated.
4058 Just use `scm_call_0', etc instead.
4060 ** Deprecated `scm_badargsp'
4062 This function is unused in Guile, but was part of its API.
4064 ** Better support for Lisp `nil'.
4066 The bit representation of `nil' has been tweaked so that it is now very
4067 efficient to check e.g. if a value is equal to Scheme's end-of-list or
4068 Lisp's nil. Additionally there are a heap of new, specific predicates
4069 like scm_is_null_or_nil.
4071 ** Better integration of Lisp `nil'.
4073 `scm_is_boolean', `scm_is_false', and `scm_is_null' all return true now
4074 for Lisp's `nil'. This shouldn't affect any Scheme code at this point,
4075 but when we start to integrate more with Emacs, it is possible that we
4076 break code that assumes that, for example, `(not x)' implies that `x' is
4077 `eq?' to `#f'. This is not a common assumption. Refactoring affected
4078 code to rely on properties instead of identities will improve code
4079 correctness. See "Nil" in the manual, for more details.
4081 ** Support for static allocation of strings, symbols, and subrs.
4083 Calls to snarfing CPP macros like SCM_DEFINE macro will now allocate
4084 much of their associated data as static variables, reducing Guile's
4087 ** `scm_stat' has an additional argument, `exception_on_error'
4088 ** `scm_primitive_load_path' has an additional argument `exception_on_not_found'
4090 ** `scm_set_port_seek' and `scm_set_port_truncate' use the `scm_t_off' type
4092 Previously they would use the `off_t' type, which is fragile since its
4093 definition depends on the application's value for `_FILE_OFFSET_BITS'.
4095 ** The `long_long' C type, deprecated in 1.8, has been removed
4097 ** Removed deprecated uniform array procedures: scm_make_uve,
4098 scm_array_prototype, scm_list_to_uniform_array,
4099 scm_dimensions_to_uniform_array, scm_make_ra, scm_shap2ra, scm_cvref,
4100 scm_ra_set_contp, scm_aind, scm_raprin1
4102 These functions have been deprecated since early 2005.
4104 * Changes to the distribution
4106 ** Guile's license is now LGPLv3+
4108 In other words the GNU Lesser General Public License, version 3 or
4109 later (at the discretion of each person that chooses to redistribute
4114 Guile's build is visually quieter, due to the use of Automake 1.11's
4115 AM_SILENT_RULES. Build as `make V=1' to see all of the output.
4117 ** GOOPS documentation folded into Guile reference manual
4119 GOOPS, Guile's object system, used to be documented in separate manuals.
4120 This content is now included in Guile's manual directly.
4122 ** `guile-config' will be deprecated in favor of `pkg-config'
4124 `guile-config' has been rewritten to get its information from
4125 `pkg-config', so this should be a transparent change. Note however that
4126 guile.m4 has yet to be modified to call pkg-config instead of
4129 ** Guile now provides `guile-2.0.pc' instead of `guile-1.8.pc'
4131 Programs that use `pkg-config' to find Guile or one of its Autoconf
4132 macros should now require `guile-2.0' instead of `guile-1.8'.
4134 ** New installation directory: $(pkglibdir)/1.9/ccache
4136 If $(libdir) is /usr/lib, for example, Guile will install its .go files
4137 to /usr/lib/guile/1.9/ccache. These files are architecture-specific.
4139 ** Parallel installability fixes
4141 Guile now installs its header files to a effective-version-specific
4142 directory, and includes the effective version (e.g. 2.0) in the library
4143 name (e.g. libguile-2.0.so).
4145 This change should be transparent to users, who should detect Guile via
4146 the guile.m4 macro, or the guile-2.0.pc pkg-config file. It will allow
4147 parallel installs for multiple versions of Guile development
4150 ** Dynamically loadable extensions may be placed in a Guile-specific path
4152 Before, Guile only searched the system library paths for extensions
4153 (e.g. /usr/lib), which meant that the names of Guile extensions had to
4154 be globally unique. Installing them to a Guile-specific extensions
4155 directory is cleaner. Use `pkg-config --variable=extensiondir
4156 guile-2.0' to get the location of the extensions directory.
4158 ** User Scheme code may be placed in a version-specific path
4160 Before, there was only one way to install user Scheme code to a
4161 version-specific Guile directory: install to Guile's own path,
4162 e.g. /usr/share/guile/2.0. The site directory,
4163 e.g. /usr/share/guile/site, was unversioned. This has been changed to
4164 add a version-specific site directory, e.g. /usr/share/guile/site/2.0,
4165 searched before the global site directory.
4167 ** New dependency: libgc
4169 See http://www.hpl.hp.com/personal/Hans_Boehm/gc/, for more information.
4171 ** New dependency: GNU libunistring
4173 See http://www.gnu.org/software/libunistring/, for more information. Our
4174 Unicode support uses routines from libunistring.
4176 ** New dependency: libffi
4178 See http://sourceware.org/libffi/, for more information.
4182 Changes in 1.8.8 (since 1.8.7)
4186 ** Fix possible buffer overruns when parsing numbers
4187 ** Avoid clash with system setjmp/longjmp on IA64
4188 ** Fix `wrong type arg' exceptions with IPv6 addresses
4191 Changes in 1.8.7 (since 1.8.6)
4193 * New modules (see the manual for details)
4195 ** `(srfi srfi-98)', an interface to access environment variables
4199 ** Fix compilation with `--disable-deprecated'
4200 ** Fix %fast-slot-ref/set!, to avoid possible segmentation fault
4201 ** Fix MinGW build problem caused by HAVE_STRUCT_TIMESPEC confusion
4202 ** Fix build problem when scm_t_timespec is different from struct timespec
4203 ** Fix build when compiled with -Wundef -Werror
4204 ** More build fixes for `alphaev56-dec-osf5.1b' (Tru64)
4205 ** Build fixes for `powerpc-ibm-aix5.3.0.0' (AIX 5.3)
4206 ** With GCC, always compile with `-mieee' on `alpha*' and `sh*'
4207 ** Better diagnose broken `(strftime "%z" ...)' in `time.test' (bug #24130)
4208 ** Fix parsing of SRFI-88/postfix keywords longer than 128 characters
4209 ** Fix reading of complex numbers where both parts are inexact decimals
4211 ** Allow @ macro to work with (ice-9 syncase)
4213 Previously, use of the @ macro in a module whose code is being
4214 transformed by (ice-9 syncase) would cause an "Invalid syntax" error.
4215 Now it works as you would expect (giving the value of the specified
4218 ** Have `scm_take_locale_symbol ()' return an interned symbol (bug #25865)
4221 Changes in 1.8.6 (since 1.8.5)
4223 * New features (see the manual for details)
4225 ** New convenience function `scm_c_symbol_length ()'
4227 ** Single stepping through code from Emacs
4229 When you use GDS to evaluate Scheme code from Emacs, you can now use
4230 `C-u' to indicate that you want to single step through that code. See
4231 `Evaluating Scheme Code' in the manual for more details.
4233 ** New "guile(1)" man page!
4235 * Changes to the distribution
4237 ** Automake's `AM_MAINTAINER_MODE' is no longer used
4239 Thus, the `--enable-maintainer-mode' configure option is no longer
4240 available: Guile is now always configured in "maintainer mode".
4242 ** `ChangeLog' files are no longer updated
4244 Instead, changes are detailed in the version control system's logs. See
4245 the top-level `ChangeLog' files for details.
4250 ** `symbol->string' now returns a read-only string, as per R5RS
4251 ** Fix incorrect handling of the FLAGS argument of `fold-matches'
4252 ** `guile-config link' now prints `-L$libdir' before `-lguile'
4253 ** Fix memory corruption involving GOOPS' `class-redefinition'
4254 ** Fix possible deadlock in `mutex-lock'
4255 ** Fix build issue on Tru64 and ia64-hp-hpux11.23 (`SCM_UNPACK' macro)
4256 ** Fix build issue on mips, mipsel, powerpc and ia64 (stack direction)
4257 ** Fix build issue on hppa2.0w-hp-hpux11.11 (`dirent64' and `readdir64_r')
4258 ** Fix build issue on i386-unknown-freebsd7.0 ("break strict-aliasing rules")
4259 ** Fix misleading output from `(help rationalize)'
4260 ** Fix build failure on Debian hppa architecture (bad stack growth detection)
4261 ** Fix `gcd' when called with a single, negative argument.
4262 ** Fix `Stack overflow' errors seen when building on some platforms
4263 ** Fix bug when `scm_with_guile ()' was called several times from the
4265 ** The handler of SRFI-34 `with-exception-handler' is now invoked in the
4266 dynamic environment of the call to `raise'
4267 ** Fix potential deadlock in `make-struct'
4268 ** Fix compilation problem with libltdl from Libtool 2.2.x
4269 ** Fix sloppy bound checking in `string-{ref,set!}' with the empty string
4272 Changes in 1.8.5 (since 1.8.4)
4274 * Infrastructure changes
4276 ** Guile repository switched from CVS to Git
4278 The new repository can be accessed using
4279 "git-clone git://git.sv.gnu.org/guile.git", or can be browsed on-line at
4280 http://git.sv.gnu.org/gitweb/?p=guile.git . See `README' for details.
4282 ** Add support for `pkg-config'
4284 See "Autoconf Support" in the manual for details.
4286 * New modules (see the manual for details)
4290 * New features (see the manual for details)
4292 ** New `postfix' read option, for SRFI-88 keyword syntax
4293 ** Some I/O primitives have been inlined, which improves I/O performance
4294 ** New object-based traps infrastructure
4296 This is a GOOPS-based infrastructure that builds on Guile's low-level
4297 evaluator trap calls and facilitates the development of debugging
4298 features like single-stepping, breakpoints, tracing and profiling.
4299 See the `Traps' node of the manual for details.
4301 ** New support for working on Guile code from within Emacs
4303 Guile now incorporates the `GDS' library (previously distributed
4304 separately) for working on Guile code from within Emacs. See the
4305 `Using Guile In Emacs' node of the manual for details.
4309 ** `scm_add_slot ()' no longer segfaults (fixes bug #22369)
4310 ** Fixed `(ice-9 match)' for patterns like `((_ ...) ...)'
4312 Previously, expressions like `(match '((foo) (bar)) (((_ ...) ...) #t))'
4313 would trigger an unbound variable error for `match:andmap'.
4315 ** `(oop goops describe)' now properly provides the `describe' feature
4316 ** Fixed `args-fold' from `(srfi srfi-37)'
4318 Previously, parsing short option names of argument-less options would
4319 lead to a stack overflow.
4321 ** `(srfi srfi-35)' is now visible through `cond-expand'
4322 ** Fixed type-checking for the second argument of `eval'
4323 ** Fixed type-checking for SRFI-1 `partition'
4324 ** Fixed `struct-ref' and `struct-set!' on "light structs"
4325 ** Honor struct field access rights in GOOPS
4326 ** Changed the storage strategy of source properties, which fixes a deadlock
4327 ** Allow compilation of Guile-using programs in C99 mode with GCC 4.3 and later
4328 ** Fixed build issue for GNU/Linux on IA64
4329 ** Fixed build issues on NetBSD 1.6
4330 ** Fixed build issue on Solaris 2.10 x86_64
4331 ** Fixed build issue with DEC/Compaq/HP's compiler
4332 ** Fixed `scm_from_complex_double' build issue on FreeBSD
4333 ** Fixed `alloca' build issue on FreeBSD 6
4334 ** Removed use of non-portable makefile constructs
4335 ** Fixed shadowing of libc's <random.h> on Tru64, which broke compilation
4336 ** Make sure all tests honor `$TMPDIR'
4339 Changes in 1.8.4 (since 1.8.3)
4343 ** CR (ASCII 0x0d) is (again) recognized as a token delimiter by the reader
4344 ** Fixed a segmentation fault which occurred when displaying the
4345 backtrace of a stack with a promise object (made by `delay') in it.
4346 ** Make `accept' leave guile mode while blocking
4347 ** `scm_c_read ()' and `scm_c_write ()' now type-check their port argument
4348 ** Fixed a build problem on AIX (use of func_data identifier)
4349 ** Fixed a segmentation fault which occurred when hashx-ref or hashx-set! was
4350 called with an associator proc that returns neither a pair nor #f.
4351 ** Secondary threads now always return a valid module for (current-module).
4352 ** Avoid MacOS build problems caused by incorrect combination of "64"
4353 system and library calls.
4354 ** `guile-snarf' now honors `$TMPDIR'
4355 ** `guile-config compile' now reports CPPFLAGS used at compile-time
4356 ** Fixed build with Sun Studio (Solaris 9)
4357 ** Fixed wrong-type-arg errors when creating zero length SRFI-4
4358 uniform vectors on AIX.
4359 ** Fixed a deadlock that occurs upon GC with multiple threads.
4360 ** Fixed compile problem with GCC on Solaris and AIX (use of _Complex_I)
4361 ** Fixed autotool-derived build problems on AIX 6.1.
4362 ** Fixed NetBSD/alpha support
4363 ** Fixed MacOS build problem caused by use of rl_get_keymap(_name)
4365 * New modules (see the manual for details)
4369 * Documentation fixes and improvements
4371 ** Removed premature breakpoint documentation
4373 The features described are not available in the series of 1.8.x
4374 releases, so the documentation was misleading and has been removed.
4376 ** More about Guile's default *random-state* variable
4378 ** GOOPS: more about how to use `next-method'
4380 * Changes to the distribution
4382 ** Corrected a few files that referred incorrectly to the old GPL + special exception licence
4384 In fact Guile since 1.8.0 has been licensed with the GNU Lesser
4385 General Public License, and the few incorrect files have now been
4386 fixed to agree with the rest of the Guile distribution.
4388 ** Removed unnecessary extra copies of COPYING*
4390 The distribution now contains a single COPYING.LESSER at its top level.
4393 Changes in 1.8.3 (since 1.8.2)
4395 * New modules (see the manual for details)
4402 ** The `(ice-9 slib)' module now works as expected
4403 ** Expressions like "(set! 'x #t)" no longer yield a crash
4404 ** Warnings about duplicate bindings now go to stderr
4405 ** A memory leak in `make-socket-address' was fixed
4406 ** Alignment issues (e.g., on SPARC) in network routines were fixed
4407 ** A threading issue that showed up at least on NetBSD was fixed
4408 ** Build problems on Solaris and IRIX fixed
4410 * Implementation improvements
4412 ** The reader is now faster, which reduces startup time
4413 ** Procedures returned by `record-accessor' and `record-modifier' are faster
4416 Changes in 1.8.2 (since 1.8.1):
4418 * New procedures (see the manual for details)
4420 ** set-program-arguments
4423 * Incompatible changes
4425 ** The body of a top-level `define' no longer sees the binding being created
4427 In a top-level `define', the binding being created is no longer visible
4428 from the `define' body. This breaks code like
4429 "(define foo (begin (set! foo 1) (+ foo 1)))", where `foo' is now
4430 unbound in the body. However, such code was not R5RS-compliant anyway,
4435 ** Fractions were not `equal?' if stored in unreduced form.
4436 (A subtle problem, since printing a value reduced it, making it work.)
4437 ** srfi-60 `copy-bit' failed on 64-bit systems
4438 ** "guile --use-srfi" option at the REPL can replace core functions
4439 (Programs run with that option were ok, but in the interactive REPL
4440 the core bindings got priority, preventing SRFI replacements or
4442 ** `regexp-exec' doesn't abort() on #\nul in the input or bad flags arg
4443 ** `kill' on mingw throws an error for a PID other than oneself
4444 ** Procedure names are attached to procedure-with-setters
4445 ** Array read syntax works with negative lower bound
4446 ** `array-in-bounds?' fix if an array has different lower bounds on each index
4447 ** `*' returns exact 0 for "(* inexact 0)"
4448 This follows what it always did for "(* 0 inexact)".
4449 ** SRFI-19: Value returned by `(current-time time-process)' was incorrect
4450 ** SRFI-19: `date->julian-day' did not account for timezone offset
4451 ** `ttyname' no longer crashes when passed a non-tty argument
4452 ** `inet-ntop' no longer crashes on SPARC when passed an `AF_INET' address
4453 ** Small memory leaks have been fixed in `make-fluid' and `add-history'
4454 ** GOOPS: Fixed a bug in `method-more-specific?'
4455 ** Build problems on Solaris fixed
4456 ** Build problems on HP-UX IA64 fixed
4457 ** Build problems on MinGW fixed
4460 Changes in 1.8.1 (since 1.8.0):
4462 * LFS functions are now used to access 64-bit files on 32-bit systems.
4464 * New procedures (see the manual for details)
4466 ** primitive-_exit - [Scheme] the-root-module
4467 ** scm_primitive__exit - [C]
4468 ** make-completion-function - [Scheme] (ice-9 readline)
4469 ** scm_c_locale_stringn_to_number - [C]
4470 ** scm_srfi1_append_reverse [C]
4471 ** scm_srfi1_append_reverse_x [C]
4479 ** Build problems have been fixed on MacOS, SunOS, and QNX.
4481 ** `strftime' fix sign of %z timezone offset.
4483 ** A one-dimensional array can now be 'equal?' to a vector.
4485 ** Structures, records, and SRFI-9 records can now be compared with `equal?'.
4487 ** SRFI-14 standard char sets are recomputed upon a successful `setlocale'.
4489 ** `record-accessor' and `record-modifier' now have strict type checks.
4491 Record accessor and modifier procedures now throw an error if the
4492 record type of the record they're given is not the type expected.
4493 (Previously accessors returned #f and modifiers silently did nothing).
4495 ** It is now OK to use both autoload and use-modules on a given module.
4497 ** `apply' checks the number of arguments more carefully on "0 or 1" funcs.
4499 Previously there was no checking on primatives like make-vector that
4500 accept "one or two" arguments. Now there is.
4502 ** The srfi-1 assoc function now calls its equality predicate properly.
4504 Previously srfi-1 assoc would call the equality predicate with the key
4505 last. According to the SRFI, the key should be first.
4507 ** A bug in n-par-for-each and n-for-each-par-map has been fixed.
4509 ** The array-set! procedure no longer segfaults when given a bit vector.
4511 ** Bugs in make-shared-array have been fixed.
4513 ** string<? and friends now follow char<? etc order on 8-bit chars.
4515 ** The format procedure now handles inf and nan values for ~f correctly.
4517 ** exact->inexact should no longer overflow when given certain large fractions.
4519 ** srfi-9 accessor and modifier procedures now have strict record type checks.
4521 This matches the srfi-9 specification.
4523 ** (ice-9 ftw) procedures won't ignore different files with same inode number.
4525 Previously the (ice-9 ftw) procedures would ignore any file that had
4526 the same inode number as a file they had already seen, even if that
4527 file was on a different device.
4530 Changes in 1.8.0 (changes since the 1.6.x series):
4532 * Changes to the distribution
4534 ** Guile is now licensed with the GNU Lesser General Public License.
4536 ** The manual is now licensed with the GNU Free Documentation License.
4538 ** Guile now requires GNU MP (http://swox.com/gmp).
4540 Guile now uses the GNU MP library for arbitrary precision arithmetic.
4542 ** Guile now has separate private and public configuration headers.
4544 That is, things like HAVE_STRING_H no longer leak from Guile's
4547 ** Guile now provides and uses an "effective" version number.
4549 Guile now provides scm_effective_version and effective-version
4550 functions which return the "effective" version number. This is just
4551 the normal full version string without the final micro-version number,
4552 so the current effective-version is "1.8". The effective version
4553 should remain unchanged during a stable series, and should be used for
4554 items like the versioned share directory name
4555 i.e. /usr/share/guile/1.8.
4557 Providing an unchanging version number during a stable release for
4558 things like the versioned share directory can be particularly
4559 important for Guile "add-on" packages, since it provides a directory
4560 that they can install to that won't be changed out from under them
4561 with each micro release during a stable series.
4563 ** Thread implementation has changed.
4565 When you configure "--with-threads=null", you will get the usual
4566 threading API (call-with-new-thread, make-mutex, etc), but you can't
4567 actually create new threads. Also, "--with-threads=no" is now
4568 equivalent to "--with-threads=null". This means that the thread API
4569 is always present, although you might not be able to create new
4572 When you configure "--with-threads=pthreads" or "--with-threads=yes",
4573 you will get threads that are implemented with the portable POSIX
4574 threads. These threads can run concurrently (unlike the previous
4575 "coop" thread implementation), but need to cooperate for things like
4578 The default is "pthreads", unless your platform doesn't have pthreads,
4579 in which case "null" threads are used.
4581 See the manual for details, nodes "Initialization", "Multi-Threading",
4582 "Blocking", and others.
4584 ** There is the new notion of 'discouraged' features.
4586 This is a milder form of deprecation.
4588 Things that are discouraged should not be used in new code, but it is
4589 OK to leave them in old code for now. When a discouraged feature is
4590 used, no warning message is printed like there is for 'deprecated'
4591 features. Also, things that are merely discouraged are nevertheless
4592 implemented efficiently, while deprecated features can be very slow.
4594 You can omit discouraged features from libguile by configuring it with
4595 the '--disable-discouraged' option.
4597 ** Deprecation warnings can be controlled at run-time.
4599 (debug-enable 'warn-deprecated) switches them on and (debug-disable
4600 'warn-deprecated) switches them off.
4602 ** Support for SRFI 61, extended cond syntax for multiple values has
4605 This SRFI is always available.
4607 ** Support for require-extension, SRFI-55, has been added.
4609 The SRFI-55 special form `require-extension' has been added. It is
4610 available at startup, and provides a portable way to load Scheme
4611 extensions. SRFI-55 only requires support for one type of extension,
4612 "srfi"; so a set of SRFIs may be loaded via (require-extension (srfi 1
4615 ** New module (srfi srfi-26) provides support for `cut' and `cute'.
4617 The (srfi srfi-26) module is an implementation of SRFI-26 which
4618 provides the `cut' and `cute' syntax. These may be used to specialize
4619 parameters without currying.
4621 ** New module (srfi srfi-31)
4623 This is an implementation of SRFI-31 which provides a special form
4624 `rec' for recursive evaluation.
4626 ** The modules (srfi srfi-13), (srfi srfi-14) and (srfi srfi-4) have
4627 been merged with the core, making their functionality always
4630 The modules are still available, tho, and you could use them together
4631 with a renaming import, for example.
4633 ** Guile no longer includes its own version of libltdl.
4635 The official version is good enough now.
4637 ** The --enable-htmldoc option has been removed from 'configure'.
4639 Support for translating the documentation into HTML is now always
4640 provided. Use 'make html'.
4642 ** New module (ice-9 serialize):
4644 (serialize FORM1 ...) and (parallelize FORM1 ...) are useful when you
4645 don't trust the thread safety of most of your program, but where you
4646 have some section(s) of code which you consider can run in parallel to
4647 other sections. See ice-9/serialize.scm for more information.
4649 ** The configure option '--disable-arrays' has been removed.
4651 Support for arrays and uniform numeric arrays is now always included
4654 * Changes to the stand-alone interpreter
4656 ** New command line option `-L'.
4658 This option adds a directory to the front of the load path.
4660 ** New command line option `--no-debug'.
4662 Specifying `--no-debug' on the command line will keep the debugging
4663 evaluator turned off, even for interactive sessions.
4665 ** User-init file ~/.guile is now loaded with the debugging evaluator.
4667 Previously, the normal evaluator would have been used. Using the
4668 debugging evaluator gives better error messages.
4670 ** The '-e' option now 'read's its argument.
4672 This is to allow the new '(@ MODULE-NAME VARIABLE-NAME)' construct to
4673 be used with '-e'. For example, you can now write a script like
4676 exec guile -e '(@ (demo) main)' -s "$0" "$@"
4679 (define-module (demo)
4683 (format #t "Demo: ~a~%" args))
4686 * Changes to Scheme functions and syntax
4688 ** Guardians have changed back to their original semantics
4690 Guardians now behave like described in the paper by Dybvig et al. In
4691 particular, they no longer make guarantees about the order in which
4692 they return objects, and they can no longer be greedy.
4694 They no longer drop cyclic data structures.
4696 The C function scm_make_guardian has been changed incompatibly and no
4697 longer takes the 'greedy_p' argument.
4699 ** New function hashx-remove!
4701 This function completes the set of 'hashx' functions.
4703 ** The concept of dynamic roots has been factored into continuation
4704 barriers and dynamic states.
4706 Each thread has a current dynamic state that carries the values of the
4707 fluids. You can create and copy dynamic states and use them as the
4708 second argument for 'eval'. See "Fluids and Dynamic States" in the
4711 To restrict the influence that captured continuations can have on the
4712 control flow, you can errect continuation barriers. See "Continuation
4713 Barriers" in the manual.
4715 The function call-with-dynamic-root now essentially temporarily
4716 installs a new dynamic state and errects a continuation barrier.
4718 ** The default load path no longer includes "." at the end.
4720 Automatically loading modules from the current directory should not
4721 happen by default. If you want to allow it in a more controlled
4722 manner, set the environment variable GUILE_LOAD_PATH or the Scheme
4723 variable %load-path.
4725 ** The uniform vector and array support has been overhauled.
4727 It now complies with SRFI-4 and the weird prototype based uniform
4728 array creation has been deprecated. See the manual for more details.
4730 Some non-compatible changes have been made:
4731 - characters can no longer be stored into byte arrays.
4732 - strings and bit vectors are no longer considered to be uniform numeric
4734 - array-rank throws an error for non-arrays instead of returning zero.
4735 - array-ref does no longer accept non-arrays when no indices are given.
4737 There is the new notion of 'generalized vectors' and corresponding
4738 procedures like 'generalized-vector-ref'. Generalized vectors include
4739 strings, bitvectors, ordinary vectors, and uniform numeric vectors.
4741 Arrays use generalized vectors as their storage, so that you still
4742 have arrays of characters, bits, etc. However, uniform-array-read!
4743 and uniform-array-write can no longer read/write strings and
4746 ** There is now support for copy-on-write substrings, mutation-sharing
4747 substrings and read-only strings.
4749 Three new procedures are related to this: substring/shared,
4750 substring/copy, and substring/read-only. See the manual for more
4753 ** Backtraces will now highlight the value that caused the error.
4755 By default, these values are enclosed in "{...}", such as in this
4764 <unnamed port>:1:1: In procedure car in expression (car (quote a)):
4765 <unnamed port>:1:1: Wrong type (expecting pair): a
4766 ABORT: (wrong-type-arg)
4768 The prefix and suffix used for highlighting can be set via the two new
4769 printer options 'highlight-prefix' and 'highlight-suffix'. For
4770 example, putting this into ~/.guile will output the bad value in bold
4771 on an ANSI terminal:
4773 (print-set! highlight-prefix "\x1b[1m")
4774 (print-set! highlight-suffix "\x1b[22m")
4777 ** 'gettext' support for internationalization has been added.
4779 See the manual for details.
4781 ** New syntax '@' and '@@':
4783 You can now directly refer to variables exported from a module by
4786 (@ MODULE-NAME VARIABLE-NAME)
4788 For example (@ (ice-9 pretty-print) pretty-print) will directly access
4789 the pretty-print variable exported from the (ice-9 pretty-print)
4790 module. You don't need to 'use' that module first. You can also use
4791 '@' as a target of 'set!', as in (set! (@ mod var) val).
4793 The related syntax (@@ MODULE-NAME VARIABLE-NAME) works just like '@',
4794 but it can also access variables that have not been exported. It is
4795 intended only for kluges and temporary fixes and for debugging, not
4798 ** Keyword syntax has been made more disciplined.
4800 Previously, the name of a keyword was read as a 'token' but printed as
4801 a symbol. Now, it is read as a general Scheme datum which must be a
4812 ERROR: In expression (a b c):
4816 ERROR: Unbound variable: foo
4821 ERROR: Wrong type (expecting symbol): 12
4825 ERROR: Wrong type (expecting symbol): (a b c)
4829 ** The printing of symbols that might look like keywords can be
4832 The new printer option 'quote-keywordish-symbols' controls how symbols
4833 are printed that have a colon as their first or last character. The
4834 default now is to only quote a symbol with #{...}# when the read
4835 option 'keywords' is not '#f'. Thus:
4837 guile> (define foo (string->symbol ":foo"))
4838 guile> (read-set! keywords #f)
4841 guile> (read-set! keywords 'prefix)
4844 guile> (print-set! quote-keywordish-symbols #f)
4848 ** 'while' now provides 'break' and 'continue'
4850 break and continue were previously bound in a while loop, but not
4851 documented, and continue didn't quite work properly. The undocumented
4852 parameter to break which gave a return value for the while has been
4855 ** 'call-with-current-continuation' is now also available under the name
4858 ** The module system now checks for duplicate bindings.
4860 The module system now can check for name conflicts among imported
4863 The behavior can be controlled by specifying one or more 'duplicates'
4864 handlers. For example, to make Guile return an error for every name
4867 (define-module (foo)
4872 The new default behavior of the module system when a name collision
4873 has been detected is to
4875 1. Give priority to bindings marked as a replacement.
4876 2. Issue a warning (different warning if overriding core binding).
4877 3. Give priority to the last encountered binding (this corresponds to
4880 If you want the old behavior back without replacements or warnings you
4883 (default-duplicate-binding-handler 'last)
4885 to your .guile init file.
4887 ** New define-module option: :replace
4889 :replace works as :export, but, in addition, marks the binding as a
4892 A typical example is `format' in (ice-9 format) which is a replacement
4893 for the core binding `format'.
4895 ** Adding prefixes to imported bindings in the module system
4897 There is now a new :use-module option :prefix. It can be used to add
4898 a prefix to all imported bindings.
4900 (define-module (foo)
4901 :use-module ((bar) :prefix bar:))
4903 will import all bindings exported from bar, but rename them by adding
4906 ** Conflicting generic functions can be automatically merged.
4908 When two imported bindings conflict and they are both generic
4909 functions, the two functions can now be merged automatically. This is
4910 activated with the 'duplicates' handler 'merge-generics'.
4912 ** New function: effective-version
4914 Returns the "effective" version number. This is just the normal full
4915 version string without the final micro-version number. See "Changes
4916 to the distribution" above.
4918 ** New threading functions: parallel, letpar, par-map, and friends
4920 These are convenient ways to run calculations in parallel in new
4921 threads. See "Parallel forms" in the manual for details.
4923 ** New function 'try-mutex'.
4925 This function will attempt to lock a mutex but will return immediately
4926 instead of blocking and indicate failure.
4928 ** Waiting on a condition variable can have a timeout.
4930 The function 'wait-condition-variable' now takes a third, optional
4931 argument that specifies the point in time where the waiting should be
4934 ** New function 'broadcast-condition-variable'.
4936 ** New functions 'all-threads' and 'current-thread'.
4938 ** Signals and system asyncs work better with threads.
4940 The function 'sigaction' now takes a fourth, optional, argument that
4941 specifies the thread that the handler should run in. When the
4942 argument is omitted, the handler will run in the thread that called
4945 Likewise, 'system-async-mark' takes a second, optional, argument that
4946 specifies the thread that the async should run in. When it is
4947 omitted, the async will run in the thread that called
4948 'system-async-mark'.
4950 C code can use the new functions scm_sigaction_for_thread and
4951 scm_system_async_mark_for_thread to pass the new thread argument.
4953 When a thread blocks on a mutex, a condition variable or is waiting
4954 for IO to be possible, it will still execute system asyncs. This can
4955 be used to interrupt such a thread by making it execute a 'throw', for
4958 ** The function 'system-async' is deprecated.
4960 You can now pass any zero-argument procedure to 'system-async-mark'.
4961 The function 'system-async' will just return its argument unchanged
4964 ** New functions 'call-with-blocked-asyncs' and
4965 'call-with-unblocked-asyncs'
4967 The expression (call-with-blocked-asyncs PROC) will call PROC and will
4968 block execution of system asyncs for the current thread by one level
4969 while PROC runs. Likewise, call-with-unblocked-asyncs will call a
4970 procedure and will unblock the execution of system asyncs by one
4971 level for the current thread.
4973 Only system asyncs are affected by these functions.
4975 ** The functions 'mask-signals' and 'unmask-signals' are deprecated.
4977 Use 'call-with-blocked-asyncs' or 'call-with-unblocked-asyncs'
4978 instead. Those functions are easier to use correctly and can be
4981 ** New function 'unsetenv'.
4983 ** New macro 'define-syntax-public'.
4985 It works like 'define-syntax' and also exports the defined macro (but
4988 ** There is support for Infinity and NaNs.
4990 Following PLT Scheme, Guile can now work with infinite numbers, and
4993 There is new syntax for numbers: "+inf.0" (infinity), "-inf.0"
4994 (negative infinity), "+nan.0" (not-a-number), and "-nan.0" (same as
4995 "+nan.0"). These numbers are inexact and have no exact counterpart.
4997 Dividing by an inexact zero returns +inf.0 or -inf.0, depending on the
4998 sign of the dividend. The infinities are integers, and they answer #t
4999 for both 'even?' and 'odd?'. The +nan.0 value is not an integer and is
5000 not '=' to itself, but '+nan.0' is 'eqv?' to itself.
5011 ERROR: Numerical overflow
5013 Two new predicates 'inf?' and 'nan?' can be used to test for the
5016 ** Inexact zero can have a sign.
5018 Guile can now distinguish between plus and minus inexact zero, if your
5019 platform supports this, too. The two zeros are equal according to
5020 '=', but not according to 'eqv?'. For example
5031 ** Guile now has exact rationals.
5033 Guile can now represent fractions such as 1/3 exactly. Computing with
5034 them is also done exactly, of course:
5039 ** 'floor', 'ceiling', 'round' and 'truncate' now return exact numbers
5040 for exact arguments.
5042 For example: (floor 2) now returns an exact 2 where in the past it
5043 returned an inexact 2.0. Likewise, (floor 5/4) returns an exact 1.
5045 ** inexact->exact no longer returns only integers.
5047 Without exact rationals, the closest exact number was always an
5048 integer, but now inexact->exact returns the fraction that is exactly
5049 equal to a floating point number. For example:
5051 (inexact->exact 1.234)
5052 => 694680242521899/562949953421312
5054 When you want the old behavior, use 'round' explicitly:
5056 (inexact->exact (round 1.234))
5059 ** New function 'rationalize'.
5061 This function finds a simple fraction that is close to a given real
5062 number. For example (and compare with inexact->exact above):
5064 (rationalize (inexact->exact 1.234) 1/2000)
5067 Note that, as required by R5RS, rationalize returns only then an exact
5068 result when both its arguments are exact.
5070 ** 'odd?' and 'even?' work also for inexact integers.
5072 Previously, (odd? 1.0) would signal an error since only exact integers
5073 were recognized as integers. Now (odd? 1.0) returns #t, (odd? 2.0)
5074 returns #f and (odd? 1.5) signals an error.
5076 ** Guile now has uninterned symbols.
5078 The new function 'make-symbol' will return an uninterned symbol. This
5079 is a symbol that is unique and is guaranteed to remain unique.
5080 However, uninterned symbols can not yet be read back in.
5082 Use the new function 'symbol-interned?' to check whether a symbol is
5085 ** pretty-print has more options.
5087 The function pretty-print from the (ice-9 pretty-print) module can now
5088 also be invoked with keyword arguments that control things like
5089 maximum output width. See the manual for details.
5091 ** Variables have no longer a special behavior for `equal?'.
5093 Previously, comparing two variables with `equal?' would recursivly
5094 compare their values. This is no longer done. Variables are now only
5095 `equal?' if they are `eq?'.
5097 ** `(begin)' is now valid.
5099 You can now use an empty `begin' form. It will yield #<unspecified>
5100 when evaluated and simply be ignored in a definition context.
5102 ** Deprecated: procedure->macro
5104 Change your code to use 'define-macro' or r5rs macros. Also, be aware
5105 that macro expansion will not be done during evaluation, but prior to
5108 ** Soft ports now allow a `char-ready?' procedure
5110 The vector argument to `make-soft-port' can now have a length of
5111 either 5 or 6. (Previously the length had to be 5.) The optional 6th
5112 element is interpreted as an `input-waiting' thunk -- i.e. a thunk
5113 that returns the number of characters that can be read immediately
5114 without the soft port blocking.
5116 ** Deprecated: undefine
5118 There is no replacement for undefine.
5120 ** The functions make-keyword-from-dash-symbol and keyword-dash-symbol
5121 have been discouraged.
5123 They are relics from a time where a keyword like #:foo was used
5124 directly as a Tcl option "-foo" and thus keywords were internally
5125 stored as a symbol with a starting dash. We now store a symbol
5128 Use symbol->keyword and keyword->symbol instead.
5130 ** The `cheap' debug option is now obsolete
5132 Evaluator trap calls are now unconditionally "cheap" - in other words,
5133 they pass a debug object to the trap handler rather than a full
5134 continuation. The trap handler code can capture a full continuation
5135 by using `call-with-current-continuation' in the usual way, if it so
5138 The `cheap' option is retained for now so as not to break existing
5139 code which gets or sets it, but setting it now has no effect. It will
5140 be removed in the next major Guile release.
5142 ** Evaluator trap calls now support `tweaking'
5144 `Tweaking' means that the trap handler code can modify the Scheme
5145 expression that is about to be evaluated (in the case of an
5146 enter-frame trap) or the value that is being returned (in the case of
5147 an exit-frame trap). The trap handler code indicates that it wants to
5148 do this by returning a pair whose car is the symbol 'instead and whose
5149 cdr is the modified expression or return value.
5151 * Changes to the C interface
5153 ** The functions scm_hash_fn_remove_x and scm_hashx_remove_x no longer
5154 take a 'delete' function argument.
5156 This argument makes no sense since the delete function is used to
5157 remove a pair from an alist, and this must not be configurable.
5159 This is an incompatible change.
5161 ** The GH interface is now subject to the deprecation mechanism
5163 The GH interface has been deprecated for quite some time but now it is
5164 actually removed from Guile when it is configured with
5165 --disable-deprecated.
5167 See the manual "Transitioning away from GH" for more information.
5169 ** A new family of functions for converting between C values and
5170 Scheme values has been added.
5172 These functions follow a common naming scheme and are designed to be
5173 easier to use, thread-safe and more future-proof than the older
5176 - int scm_is_* (...)
5178 These are predicates that return a C boolean: 1 or 0. Instead of
5179 SCM_NFALSEP, you can now use scm_is_true, for example.
5181 - <type> scm_to_<type> (SCM val, ...)
5183 These are functions that convert a Scheme value into an appropriate
5184 C value. For example, you can use scm_to_int to safely convert from
5187 - SCM scm_from_<type> (<type> val, ...)
5189 These functions convert from a C type to a SCM value; for example,
5190 scm_from_int for ints.
5192 There is a huge number of these functions, for numbers, strings,
5193 symbols, vectors, etc. They are documented in the reference manual in
5194 the API section together with the types that they apply to.
5196 ** New functions for dealing with complex numbers in C have been added.
5198 The new functions are scm_c_make_rectangular, scm_c_make_polar,
5199 scm_c_real_part, scm_c_imag_part, scm_c_magnitude and scm_c_angle.
5200 They work like scm_make_rectangular etc but take or return doubles
5203 ** The function scm_make_complex has been discouraged.
5205 Use scm_c_make_rectangular instead.
5207 ** The INUM macros have been deprecated.
5209 A lot of code uses these macros to do general integer conversions,
5210 although the macros only work correctly with fixnums. Use the
5211 following alternatives.
5213 SCM_INUMP -> scm_is_integer or similar
5214 SCM_NINUMP -> !scm_is_integer or similar
5215 SCM_MAKINUM -> scm_from_int or similar
5216 SCM_INUM -> scm_to_int or similar
5218 SCM_VALIDATE_INUM_* -> Do not use these; scm_to_int, etc. will
5219 do the validating for you.
5221 ** The scm_num2<type> and scm_<type>2num functions and scm_make_real
5222 have been discouraged.
5224 Use the newer scm_to_<type> and scm_from_<type> functions instead for
5225 new code. The functions have been discouraged since they don't fit
5228 ** The 'boolean' macros SCM_FALSEP etc have been discouraged.
5230 They have strange names, especially SCM_NFALSEP, and SCM_BOOLP
5231 evaluates its argument twice. Use scm_is_true, etc. instead for new
5234 ** The macro SCM_EQ_P has been discouraged.
5236 Use scm_is_eq for new code, which fits better into the naming
5239 ** The macros SCM_CONSP, SCM_NCONSP, SCM_NULLP, and SCM_NNULLP have
5242 Use the function scm_is_pair or scm_is_null instead.
5244 ** The functions scm_round and scm_truncate have been deprecated and
5245 are now available as scm_c_round and scm_c_truncate, respectively.
5247 These functions occupy the names that scm_round_number and
5248 scm_truncate_number should have.
5250 ** The functions scm_c_string2str, scm_c_substring2str, and
5251 scm_c_symbol2str have been deprecated.
5253 Use scm_to_locale_stringbuf or similar instead, maybe together with
5256 ** New functions scm_c_make_string, scm_c_string_length,
5257 scm_c_string_ref, scm_c_string_set_x, scm_c_substring,
5258 scm_c_substring_shared, scm_c_substring_copy.
5260 These are like scm_make_string, scm_length, etc. but are slightly
5261 easier to use from C.
5263 ** The macros SCM_STRINGP, SCM_STRING_CHARS, SCM_STRING_LENGTH,
5264 SCM_SYMBOL_CHARS, and SCM_SYMBOL_LENGTH have been deprecated.
5266 They export too many assumptions about the implementation of strings
5267 and symbols that are no longer true in the presence of
5268 mutation-sharing substrings and when Guile switches to some form of
5271 When working with strings, it is often best to use the normal string
5272 functions provided by Guile, such as scm_c_string_ref,
5273 scm_c_string_set_x, scm_string_append, etc. Be sure to look in the
5274 manual since many more such functions are now provided than
5277 When you want to convert a SCM string to a C string, use the
5278 scm_to_locale_string function or similar instead. For symbols, use
5279 scm_symbol_to_string and then work with that string. Because of the
5280 new string representation, scm_symbol_to_string does not need to copy
5281 and is thus quite efficient.
5283 ** Some string, symbol and keyword functions have been discouraged.
5285 They don't fit into the uniform naming scheme and are not explicit
5286 about the character encoding.
5288 Replace according to the following table:
5290 scm_allocate_string -> scm_c_make_string
5291 scm_take_str -> scm_take_locale_stringn
5292 scm_take0str -> scm_take_locale_string
5293 scm_mem2string -> scm_from_locale_stringn
5294 scm_str2string -> scm_from_locale_string
5295 scm_makfrom0str -> scm_from_locale_string
5296 scm_mem2symbol -> scm_from_locale_symboln
5297 scm_mem2uninterned_symbol -> scm_from_locale_stringn + scm_make_symbol
5298 scm_str2symbol -> scm_from_locale_symbol
5300 SCM_SYMBOL_HASH -> scm_hashq
5301 SCM_SYMBOL_INTERNED_P -> scm_symbol_interned_p
5303 scm_c_make_keyword -> scm_from_locale_keyword
5305 ** The functions scm_keyword_to_symbol and sym_symbol_to_keyword are
5306 now also available to C code.
5308 ** SCM_KEYWORDP and SCM_KEYWORDSYM have been deprecated.
5310 Use scm_is_keyword and scm_keyword_to_symbol instead, but note that
5311 the latter returns the true name of the keyword, not the 'dash name',
5312 as SCM_KEYWORDSYM used to do.
5314 ** A new way to access arrays in a thread-safe and efficient way has
5317 See the manual, node "Accessing Arrays From C".
5319 ** The old uniform vector and bitvector implementations have been
5320 unceremoniously removed.
5322 This implementation exposed the details of the tagging system of
5323 Guile. Use the new C API explained in the manual in node "Uniform
5324 Numeric Vectors" and "Bit Vectors", respectively.
5326 The following macros are gone: SCM_UVECTOR_BASE, SCM_SET_UVECTOR_BASE,
5327 SCM_UVECTOR_MAXLENGTH, SCM_UVECTOR_LENGTH, SCM_MAKE_UVECTOR_TAG,
5328 SCM_SET_UVECTOR_LENGTH, SCM_BITVECTOR_P, SCM_BITVECTOR_BASE,
5329 SCM_SET_BITVECTOR_BASE, SCM_BITVECTOR_MAX_LENGTH,
5330 SCM_BITVECTOR_LENGTH, SCM_MAKE_BITVECTOR_TAG,
5331 SCM_SET_BITVECTOR_LENGTH, SCM_BITVEC_REF, SCM_BITVEC_SET,
5334 ** The macros dealing with vectors have been deprecated.
5336 Use the new functions scm_is_vector, scm_vector_elements,
5337 scm_vector_writable_elements, etc, or scm_is_simple_vector,
5338 SCM_SIMPLE_VECTOR_REF, SCM_SIMPLE_VECTOR_SET, etc instead. See the
5339 manual for more details.
5341 Deprecated are SCM_VECTORP, SCM_VELTS, SCM_VECTOR_MAX_LENGTH,
5342 SCM_VECTOR_LENGTH, SCM_VECTOR_REF, SCM_VECTOR_SET, SCM_WRITABLE_VELTS.
5344 The following macros have been removed: SCM_VECTOR_BASE,
5345 SCM_SET_VECTOR_BASE, SCM_MAKE_VECTOR_TAG, SCM_SET_VECTOR_LENGTH,
5346 SCM_VELTS_AS_STACKITEMS, SCM_SETVELTS, SCM_GC_WRITABLE_VELTS.
5348 ** Some C functions and macros related to arrays have been deprecated.
5350 Migrate according to the following table:
5352 scm_make_uve -> scm_make_typed_array, scm_make_u8vector etc.
5353 scm_make_ra -> scm_make_array
5354 scm_shap2ra -> scm_make_array
5355 scm_cvref -> scm_c_generalized_vector_ref
5356 scm_ra_set_contp -> do not use
5357 scm_aind -> scm_array_handle_pos
5358 scm_raprin1 -> scm_display or scm_write
5360 SCM_ARRAYP -> scm_is_array
5361 SCM_ARRAY_NDIM -> scm_c_array_rank
5362 SCM_ARRAY_DIMS -> scm_array_handle_dims
5363 SCM_ARRAY_CONTP -> do not use
5364 SCM_ARRAY_MEM -> do not use
5365 SCM_ARRAY_V -> scm_array_handle_elements or similar
5366 SCM_ARRAY_BASE -> do not use
5368 ** SCM_CELL_WORD_LOC has been deprecated.
5370 Use the new macro SCM_CELL_OBJECT_LOC instead, which returns a pointer
5371 to a SCM, as opposed to a pointer to a scm_t_bits.
5373 This was done to allow the correct use of pointers into the Scheme
5374 heap. Previously, the heap words were of type scm_t_bits and local
5375 variables and function arguments were of type SCM, making it
5376 non-standards-conformant to have a pointer that can point to both.
5378 ** New macros SCM_SMOB_DATA_2, SCM_SMOB_DATA_3, etc.
5380 These macros should be used instead of SCM_CELL_WORD_2/3 to access the
5381 second and third words of double smobs. Likewise for
5382 SCM_SET_SMOB_DATA_2 and SCM_SET_SMOB_DATA_3.
5384 Also, there is SCM_SMOB_FLAGS and SCM_SET_SMOB_FLAGS that should be
5385 used to get and set the 16 exra bits in the zeroth word of a smob.
5387 And finally, there is SCM_SMOB_OBJECT and SCM_SMOB_SET_OBJECT for
5388 accesing the first immediate word of a smob as a SCM value, and there
5389 is SCM_SMOB_OBJECT_LOC for getting a pointer to the first immediate
5390 smob word. Like wise for SCM_SMOB_OBJECT_2, etc.
5392 ** New way to deal with non-local exits and re-entries.
5394 There is a new set of functions that essentially do what
5395 scm_internal_dynamic_wind does, but in a way that is more convenient
5396 for C code in some situations. Here is a quick example of how to
5397 prevent a potential memory leak:
5404 scm_dynwind_begin (0);
5406 mem = scm_malloc (100);
5407 scm_dynwind_unwind_handler (free, mem, SCM_F_WIND_EXPLICITLY);
5409 /* MEM would leak if BAR throws an error.
5410 SCM_DYNWIND_UNWIND_HANDLER frees it nevertheless.
5417 /* Because of SCM_F_WIND_EXPLICITLY, MEM will be freed by
5418 SCM_DYNWIND_END as well.
5422 For full documentation, see the node "Dynamic Wind" in the manual.
5424 ** New function scm_dynwind_free
5426 This function calls 'free' on a given pointer when a dynwind context
5427 is left. Thus the call to scm_dynwind_unwind_handler above could be
5428 replaced with simply scm_dynwind_free (mem).
5430 ** New functions scm_c_call_with_blocked_asyncs and
5431 scm_c_call_with_unblocked_asyncs
5433 Like scm_call_with_blocked_asyncs etc. but for C functions.
5435 ** New functions scm_dynwind_block_asyncs and scm_dynwind_unblock_asyncs
5437 In addition to scm_c_call_with_blocked_asyncs you can now also use
5438 scm_dynwind_block_asyncs in a 'dynwind context' (see above). Likewise for
5439 scm_c_call_with_unblocked_asyncs and scm_dynwind_unblock_asyncs.
5441 ** The macros SCM_DEFER_INTS, SCM_ALLOW_INTS, SCM_REDEFER_INTS,
5442 SCM_REALLOW_INTS have been deprecated.
5444 They do no longer fulfill their original role of blocking signal
5445 delivery. Depending on what you want to achieve, replace a pair of
5446 SCM_DEFER_INTS and SCM_ALLOW_INTS with a dynwind context that locks a
5447 mutex, blocks asyncs, or both. See node "Critical Sections" in the
5450 ** The value 'scm_mask_ints' is no longer writable.
5452 Previously, you could set scm_mask_ints directly. This is no longer
5453 possible. Use scm_c_call_with_blocked_asyncs and
5454 scm_c_call_with_unblocked_asyncs instead.
5456 ** New way to temporarily set the current input, output or error ports
5458 C code can now use scm_dynwind_current_<foo>_port in a 'dynwind
5459 context' (see above). <foo> is one of "input", "output" or "error".
5461 ** New way to temporarily set fluids
5463 C code can now use scm_dynwind_fluid in a 'dynwind context' (see
5464 above) to temporarily set the value of a fluid.
5466 ** New types scm_t_intmax and scm_t_uintmax.
5468 On platforms that have them, these types are identical to intmax_t and
5469 uintmax_t, respectively. On other platforms, they are identical to
5470 the largest integer types that Guile knows about.
5472 ** The functions scm_unmemocopy and scm_unmemoize have been removed.
5474 You should not have used them.
5476 ** Many public #defines with generic names have been made private.
5478 #defines with generic names like HAVE_FOO or SIZEOF_FOO have been made
5479 private or renamed with a more suitable public name.
5481 ** The macro SCM_TYP16S has been deprecated.
5483 This macro is not intended for public use.
5485 ** The macro SCM_SLOPPY_INEXACTP has been deprecated.
5487 Use scm_is_true (scm_inexact_p (...)) instead.
5489 ** The macro SCM_SLOPPY_REALP has been deprecated.
5491 Use scm_is_real instead.
5493 ** The macro SCM_SLOPPY_COMPLEXP has been deprecated.
5495 Use scm_is_complex instead.
5497 ** Some preprocessor defines have been deprecated.
5499 These defines indicated whether a certain feature was present in Guile
5500 or not. Going forward, assume that the features are always present.
5502 The macros are: USE_THREADS, GUILE_ISELECT, READER_EXTENSIONS,
5503 DEBUG_EXTENSIONS, DYNAMIC_LINKING.
5505 The following macros have been removed completely: MEMOIZE_LOCALS,
5506 SCM_RECKLESS, SCM_CAUTIOUS.
5508 ** The preprocessor define STACK_DIRECTION has been deprecated.
5510 There should be no need to know about the stack direction for ordinary
5513 ** New function: scm_effective_version
5515 Returns the "effective" version number. This is just the normal full
5516 version string without the final micro-version number. See "Changes
5517 to the distribution" above.
5519 ** The function scm_call_with_new_thread has a new prototype.
5521 Instead of taking a list with the thunk and handler, these two
5522 arguments are now passed directly:
5524 SCM scm_call_with_new_thread (SCM thunk, SCM handler);
5526 This is an incompatible change.
5528 ** New snarfer macro SCM_DEFINE_PUBLIC.
5530 This is like SCM_DEFINE, but also calls scm_c_export for the defined
5531 function in the init section.
5533 ** The snarfer macro SCM_SNARF_INIT is now officially supported.
5535 ** Garbage collector rewrite.
5537 The garbage collector is cleaned up a lot, and now uses lazy
5538 sweeping. This is reflected in the output of (gc-stats); since cells
5539 are being freed when they are allocated, the cells-allocated field
5540 stays roughly constant.
5542 For malloc related triggers, the behavior is changed. It uses the same
5543 heuristic as the cell-triggered collections. It may be tuned with the
5544 environment variables GUILE_MIN_YIELD_MALLOC. This is the percentage
5545 for minimum yield of malloc related triggers. The default is 40.
5546 GUILE_INIT_MALLOC_LIMIT sets the initial trigger for doing a GC. The
5549 Debugging operations for the freelist have been deprecated, along with
5550 the C variables that control garbage collection. The environment
5551 variables GUILE_MAX_SEGMENT_SIZE, GUILE_INIT_SEGMENT_SIZE_2,
5552 GUILE_INIT_SEGMENT_SIZE_1, and GUILE_MIN_YIELD_2 should be used.
5554 For understanding the memory usage of a GUILE program, the routine
5555 gc-live-object-stats returns an alist containing the number of live
5556 objects for every type.
5559 ** The function scm_definedp has been renamed to scm_defined_p
5561 The name scm_definedp is deprecated.
5563 ** The struct scm_cell type has been renamed to scm_t_cell
5565 This is in accordance to Guile's naming scheme for types. Note that
5566 the name scm_cell is now used for a function that allocates and
5567 initializes a new cell (see below).
5569 ** New functions for memory management
5571 A new set of functions for memory management has been added since the
5572 old way (scm_must_malloc, scm_must_free, etc) was error prone and
5573 indeed, Guile itself contained some long standing bugs that could
5574 cause aborts in long running programs.
5576 The new functions are more symmetrical and do not need cooperation
5577 from smob free routines, among other improvements.
5579 The new functions are scm_malloc, scm_realloc, scm_calloc, scm_strdup,
5580 scm_strndup, scm_gc_malloc, scm_gc_calloc, scm_gc_realloc,
5581 scm_gc_free, scm_gc_register_collectable_memory, and
5582 scm_gc_unregister_collectable_memory. Refer to the manual for more
5583 details and for upgrading instructions.
5585 The old functions for memory management have been deprecated. They
5586 are: scm_must_malloc, scm_must_realloc, scm_must_free,
5587 scm_must_strdup, scm_must_strndup, scm_done_malloc, scm_done_free.
5589 ** Declarations of exported features are marked with SCM_API.
5591 Every declaration of a feature that belongs to the exported Guile API
5592 has been marked by adding the macro "SCM_API" to the start of the
5593 declaration. This macro can expand into different things, the most
5594 common of which is just "extern" for Unix platforms. On Win32, it can
5595 be used to control which symbols are exported from a DLL.
5597 If you `#define SCM_IMPORT' before including <libguile.h>, SCM_API
5598 will expand into "__declspec (dllimport) extern", which is needed for
5599 linking to the Guile DLL in Windows.
5601 There are also SCM_RL_IMPORT, SCM_SRFI1314_IMPORT, and
5602 SCM_SRFI4_IMPORT, for the corresponding libraries.
5604 ** SCM_NEWCELL and SCM_NEWCELL2 have been deprecated.
5606 Use the new functions scm_cell and scm_double_cell instead. The old
5607 macros had problems because with them allocation and initialization
5608 was separated and the GC could sometimes observe half initialized
5609 cells. Only careful coding by the user of SCM_NEWCELL and
5610 SCM_NEWCELL2 could make this safe and efficient.
5612 ** CHECK_ENTRY, CHECK_APPLY and CHECK_EXIT have been deprecated.
5614 Use the variables scm_check_entry_p, scm_check_apply_p and scm_check_exit_p
5617 ** SRCBRKP has been deprecated.
5619 Use scm_c_source_property_breakpoint_p instead.
5621 ** Deprecated: scm_makmacro
5623 Change your code to use either scm_makmmacro or to define macros in
5624 Scheme, using 'define-macro'.
5626 ** New function scm_c_port_for_each.
5628 This function is like scm_port_for_each but takes a pointer to a C
5629 function as the callback instead of a SCM value.
5631 ** The names scm_internal_select, scm_thread_sleep, and
5632 scm_thread_usleep have been discouraged.
5634 Use scm_std_select, scm_std_sleep, scm_std_usleep instead.
5636 ** The GC can no longer be blocked.
5638 The global flags scm_gc_heap_lock and scm_block_gc have been removed.
5639 The GC can now run (partially) concurrently with other code and thus
5640 blocking it is not well defined.
5642 ** Many definitions have been removed that were previously deprecated.
5644 scm_lisp_nil, scm_lisp_t, s_nil_ify, scm_m_nil_ify, s_t_ify,
5645 scm_m_t_ify, s_0_cond, scm_m_0_cond, s_0_ify, scm_m_0_ify, s_1_ify,
5646 scm_m_1_ify, scm_debug_newcell, scm_debug_newcell2,
5647 scm_tc16_allocated, SCM_SET_SYMBOL_HASH, SCM_IM_NIL_IFY, SCM_IM_T_IFY,
5648 SCM_IM_0_COND, SCM_IM_0_IFY, SCM_IM_1_IFY, SCM_GC_SET_ALLOCATED,
5649 scm_debug_newcell, scm_debug_newcell2, SCM_HUP_SIGNAL, SCM_INT_SIGNAL,
5650 SCM_FPE_SIGNAL, SCM_BUS_SIGNAL, SCM_SEGV_SIGNAL, SCM_ALRM_SIGNAL,
5651 SCM_GC_SIGNAL, SCM_TICK_SIGNAL, SCM_SIG_ORD, SCM_ORD_SIG,
5652 SCM_NUM_SIGS, scm_top_level_lookup_closure_var,
5653 *top-level-lookup-closure*, scm_system_transformer, scm_eval_3,
5654 scm_eval2, root_module_lookup_closure, SCM_SLOPPY_STRINGP,
5655 SCM_RWSTRINGP, scm_read_only_string_p, scm_make_shared_substring,
5656 scm_tc7_substring, sym_huh, SCM_VARVCELL, SCM_UDVARIABLEP,
5657 SCM_DEFVARIABLEP, scm_mkbig, scm_big2inum, scm_adjbig, scm_normbig,
5658 scm_copybig, scm_2ulong2big, scm_dbl2big, scm_big2dbl, SCM_FIXNUM_BIT,
5659 SCM_SETCHARS, SCM_SLOPPY_SUBSTRP, SCM_SUBSTR_STR, SCM_SUBSTR_OFFSET,
5660 SCM_LENGTH_MAX, SCM_SETLENGTH, SCM_ROSTRINGP, SCM_ROLENGTH,
5661 SCM_ROCHARS, SCM_ROUCHARS, SCM_SUBSTRP, SCM_COERCE_SUBSTR,
5662 scm_sym2vcell, scm_intern, scm_intern0, scm_sysintern, scm_sysintern0,
5663 scm_sysintern0_no_module_lookup, scm_init_symbols_deprecated,
5664 scm_vector_set_length_x, scm_contregs, scm_debug_info,
5665 scm_debug_frame, SCM_DSIDEVAL, SCM_CONST_LONG, SCM_VCELL,
5666 SCM_GLOBAL_VCELL, SCM_VCELL_INIT, SCM_GLOBAL_VCELL_INIT,
5667 SCM_HUGE_LENGTH, SCM_VALIDATE_STRINGORSUBSTR, SCM_VALIDATE_ROSTRING,
5668 SCM_VALIDATE_ROSTRING_COPY, SCM_VALIDATE_NULLORROSTRING_COPY,
5669 SCM_VALIDATE_RWSTRING, DIGITS, scm_small_istr2int, scm_istr2int,
5670 scm_istr2flo, scm_istring2number, scm_istr2int, scm_istr2flo,
5671 scm_istring2number, scm_vtable_index_vcell, scm_si_vcell, SCM_ECONSP,
5672 SCM_NECONSP, SCM_GLOC_VAR, SCM_GLOC_VAL, SCM_GLOC_SET_VAL,
5673 SCM_GLOC_VAL_LOC, scm_make_gloc, scm_gloc_p, scm_tc16_variable,
5674 SCM_CHARS, SCM_LENGTH, SCM_SET_STRING_CHARS, SCM_SET_STRING_LENGTH.
5676 * Changes to bundled modules
5680 Using the (ice-9 debug) module no longer automatically switches Guile
5681 to use the debugging evaluator. If you want to switch to the
5682 debugging evaluator (which is needed for backtrace information if you
5683 hit an error), please add an explicit "(debug-enable 'debug)" to your
5684 code just after the code to use (ice-9 debug).
5687 Changes since Guile 1.4:
5689 * Changes to the distribution
5691 ** A top-level TODO file is included.
5693 ** Guile now uses a versioning scheme similar to that of the Linux kernel.
5695 Guile now always uses three numbers to represent the version,
5696 i.e. "1.6.5". The first number, 1, is the major version number, the
5697 second number, 6, is the minor version number, and the third number,
5698 5, is the micro version number. Changes in major version number
5699 indicate major changes in Guile.
5701 Minor version numbers that are even denote stable releases, and odd
5702 minor version numbers denote development versions (which may be
5703 unstable). The micro version number indicates a minor sub-revision of
5704 a given MAJOR.MINOR release.
5706 In keeping with the new scheme, (minor-version) and scm_minor_version
5707 no longer return everything but the major version number. They now
5708 just return the minor version number. Two new functions
5709 (micro-version) and scm_micro_version have been added to report the
5710 micro version number.
5712 In addition, ./GUILE-VERSION now defines GUILE_MICRO_VERSION.
5714 ** New preprocessor definitions are available for checking versions.
5716 version.h now #defines SCM_MAJOR_VERSION, SCM_MINOR_VERSION, and
5717 SCM_MICRO_VERSION to the appropriate integer values.
5719 ** Guile now actively warns about deprecated features.
5721 The new configure option `--enable-deprecated=LEVEL' and the
5722 environment variable GUILE_WARN_DEPRECATED control this mechanism.
5723 See INSTALL and README for more information.
5725 ** Guile is much more likely to work on 64-bit architectures.
5727 Guile now compiles and passes "make check" with only two UNRESOLVED GC
5728 cases on Alpha and ia64 based machines now. Thanks to John Goerzen
5729 for the use of a test machine, and thanks to Stefan Jahn for ia64
5732 ** New functions: setitimer and getitimer.
5734 These implement a fairly direct interface to the libc functions of the
5737 ** The #. reader extension is now disabled by default.
5739 For safety reasons, #. evaluation is disabled by default. To
5740 re-enable it, set the fluid read-eval? to #t. For example:
5742 (fluid-set! read-eval? #t)
5744 but make sure you realize the potential security risks involved. With
5745 read-eval? enabled, reading a data file from an untrusted source can
5748 ** New SRFI modules have been added:
5750 SRFI-0 `cond-expand' is now supported in Guile, without requiring
5753 (srfi srfi-1) is a library containing many useful pair- and list-processing
5756 (srfi srfi-2) exports and-let*.
5758 (srfi srfi-4) implements homogeneous numeric vector datatypes.
5760 (srfi srfi-6) is a dummy module for now, since guile already provides
5761 all of the srfi-6 procedures by default: open-input-string,
5762 open-output-string, get-output-string.
5764 (srfi srfi-8) exports receive.
5766 (srfi srfi-9) exports define-record-type.
5768 (srfi srfi-10) exports define-reader-ctor and implements the reader
5771 (srfi srfi-11) exports let-values and let*-values.
5773 (srfi srfi-13) implements the SRFI String Library.
5775 (srfi srfi-14) implements the SRFI Character-Set Library.
5777 (srfi srfi-17) implements setter and getter-with-setter and redefines
5778 some accessor procedures as procedures with getters. (such as car,
5779 cdr, vector-ref etc.)
5781 (srfi srfi-19) implements the SRFI Time/Date Library.
5783 ** New scripts / "executable modules"
5785 Subdirectory "scripts" contains Scheme modules that are packaged to
5786 also be executable as scripts. At this time, these scripts are available:
5795 See README there for more info.
5797 These scripts can be invoked from the shell with the new program
5798 "guile-tools", which keeps track of installation directory for you.
5801 $ guile-tools display-commentary srfi/*.scm
5803 guile-tools is copied to the standard $bindir on "make install".
5805 ** New module (ice-9 stack-catch):
5807 stack-catch is like catch, but saves the current state of the stack in
5808 the fluid the-last-stack. This fluid can be useful when using the
5809 debugger and when re-throwing an error.
5811 ** The module (ice-9 and-let*) has been renamed to (ice-9 and-let-star)
5813 This has been done to prevent problems on lesser operating systems
5814 that can't tolerate `*'s in file names. The exported macro continues
5815 to be named `and-let*', of course.
5817 On systems that support it, there is also a compatibility module named
5818 (ice-9 and-let*). It will go away in the next release.
5820 ** New modules (oop goops) etc.:
5823 (oop goops describe)
5825 (oop goops active-slot)
5826 (oop goops composite-slot)
5828 The Guile Object Oriented Programming System (GOOPS) has been
5829 integrated into Guile. For further information, consult the GOOPS
5830 manual and tutorial in the `doc' directory.
5832 ** New module (ice-9 rdelim).
5834 This exports the following procedures which were previously defined
5835 in the default environment:
5837 read-line read-line! read-delimited read-delimited! %read-delimited!
5838 %read-line write-line
5840 For backwards compatibility the definitions are still imported into the
5841 default environment in this version of Guile. However you should add:
5843 (use-modules (ice-9 rdelim))
5845 to any program which uses the definitions, since this may change in
5848 Alternatively, if guile-scsh is installed, the (scsh rdelim) module
5849 can be used for similar functionality.
5851 ** New module (ice-9 rw)
5853 This is a subset of the (scsh rw) module from guile-scsh. Currently
5854 it defines two procedures:
5856 *** New function: read-string!/partial str [port_or_fdes [start [end]]]
5858 Read characters from a port or file descriptor into a string STR.
5859 A port must have an underlying file descriptor -- a so-called
5860 fport. This procedure is scsh-compatible and can efficiently read
5863 *** New function: write-string/partial str [port_or_fdes [start [end]]]
5865 Write characters from a string STR to a port or file descriptor.
5866 A port must have an underlying file descriptor -- a so-called
5867 fport. This procedure is mostly compatible and can efficiently
5868 write large strings.
5870 ** New module (ice-9 match)
5872 This module includes Andrew K. Wright's pattern matcher. See
5873 ice-9/match.scm for brief description or
5875 http://www.star-lab.com/wright/code.html
5877 for complete documentation.
5879 ** New module (ice-9 buffered-input)
5881 This module provides procedures to construct an input port from an
5882 underlying source of input that reads and returns its input in chunks.
5883 The underlying input source is a Scheme procedure, specified by the
5884 caller, which the port invokes whenever it needs more input.
5886 This is useful when building an input port whose back end is Readline
5887 or a UI element such as the GtkEntry widget.
5891 The reference and tutorial documentation that was previously
5892 distributed separately, as `guile-doc', is now included in the core
5893 Guile distribution. The documentation consists of the following
5896 - The Guile Tutorial (guile-tut.texi) contains a tutorial introduction
5899 - The Guile Reference Manual (guile.texi) contains (or is intended to
5900 contain) reference documentation on all aspects of Guile.
5902 - The GOOPS Manual (goops.texi) contains both tutorial-style and
5903 reference documentation for using GOOPS, Guile's Object Oriented
5906 - The Revised^5 Report on the Algorithmic Language Scheme
5909 See the README file in the `doc' directory for more details.
5911 ** There are a couple of examples in the examples/ directory now.
5913 * Changes to the stand-alone interpreter
5915 ** New command line option `--use-srfi'
5917 Using this option, SRFI modules can be loaded on startup and be
5918 available right from the beginning. This makes programming portable
5919 Scheme programs easier.
5921 The option `--use-srfi' expects a comma-separated list of numbers,
5922 each representing a SRFI number to be loaded into the interpreter
5923 before starting evaluating a script file or the REPL. Additionally,
5924 the feature identifier for the loaded SRFIs is recognized by
5925 `cond-expand' when using this option.
5928 $ guile --use-srfi=8,13
5929 guile> (receive (x z) (values 1 2) (+ 1 2))
5931 guile> (string-pad "bla" 20)
5934 ** Guile now always starts up in the `(guile-user)' module.
5936 Previously, scripts executed via the `-s' option would run in the
5937 `(guile)' module and the repl would run in the `(guile-user)' module.
5938 Now every user action takes place in the `(guile-user)' module by
5941 * Changes to Scheme functions and syntax
5943 ** Character classifiers work for non-ASCII characters.
5945 The predicates `char-alphabetic?', `char-numeric?',
5946 `char-whitespace?', `char-lower?', `char-upper?' and `char-is-both?'
5947 no longer check whether their arguments are ASCII characters.
5948 Previously, a character would only be considered alphabetic when it
5949 was also ASCII, for example.
5951 ** Previously deprecated Scheme functions have been removed:
5953 tag - no replacement.
5954 fseek - replaced by seek.
5955 list* - replaced by cons*.
5957 ** It's now possible to create modules with controlled environments
5961 (use-modules (ice-9 safe))
5962 (define m (make-safe-module))
5963 ;;; m will now be a module containing only a safe subset of R5RS
5964 (eval '(+ 1 2) m) --> 3
5965 (eval 'load m) --> ERROR: Unbound variable: load
5967 ** Evaluation of "()", the empty list, is now an error.
5969 Previously, the expression "()" evaluated to the empty list. This has
5970 been changed to signal a "missing expression" error. The correct way
5971 to write the empty list as a literal constant is to use quote: "'()".
5973 ** New concept of `Guile Extensions'.
5975 A Guile Extension is just a ordinary shared library that can be linked
5976 at run-time. We found it advantageous to give this simple concept a
5977 dedicated name to distinguish the issues related to shared libraries
5978 from the issues related to the module system.
5980 *** New function: load-extension
5982 Executing (load-extension lib init) is mostly equivalent to
5984 (dynamic-call init (dynamic-link lib))
5986 except when scm_register_extension has been called previously.
5987 Whenever appropriate, you should use `load-extension' instead of
5988 dynamic-link and dynamic-call.
5990 *** New C function: scm_c_register_extension
5992 This function registers a initialization function for use by
5993 `load-extension'. Use it when you don't want specific extensions to
5994 be loaded as shared libraries (for example on platforms that don't
5995 support dynamic linking).
5997 ** Auto-loading of compiled-code modules is deprecated.
5999 Guile used to be able to automatically find and link a shared
6000 library to satisfy requests for a module. For example, the module
6001 `(foo bar)' could be implemented by placing a shared library named
6002 "foo/libbar.so" (or with a different extension) in a directory on the
6005 This has been found to be too tricky, and is no longer supported. The
6006 shared libraries are now called "extensions". You should now write a
6007 small Scheme file that calls `load-extension' to load the shared
6008 library and initialize it explicitly.
6010 The shared libraries themselves should be installed in the usual
6011 places for shared libraries, with names like "libguile-foo-bar".
6013 For example, place this into a file "foo/bar.scm"
6015 (define-module (foo bar))
6017 (load-extension "libguile-foo-bar" "foobar_init")
6019 ** Backward incompatible change: eval EXP ENVIRONMENT-SPECIFIER
6021 `eval' is now R5RS, that is it takes two arguments.
6022 The second argument is an environment specifier, i.e. either
6024 (scheme-report-environment 5)
6025 (null-environment 5)
6026 (interaction-environment)
6032 ** The module system has been made more disciplined.
6034 The function `eval' will save and restore the current module around
6035 the evaluation of the specified expression. While this expression is
6036 evaluated, `(current-module)' will now return the right module, which
6037 is the module specified as the second argument to `eval'.
6039 A consequence of this change is that `eval' is not particularly
6040 useful when you want allow the evaluated code to change what module is
6041 designated as the current module and have this change persist from one
6042 call to `eval' to the next. The read-eval-print-loop is an example
6043 where `eval' is now inadequate. To compensate, there is a new
6044 function `primitive-eval' that does not take a module specifier and
6045 that does not save/restore the current module. You should use this
6046 function together with `set-current-module', `current-module', etc
6047 when you want to have more control over the state that is carried from
6048 one eval to the next.
6050 Additionally, it has been made sure that forms that are evaluated at
6051 the top level are always evaluated with respect to the current module.
6052 Previously, subforms of top-level forms such as `begin', `case',
6053 etc. did not respect changes to the current module although these
6054 subforms are at the top-level as well.
6056 To prevent strange behavior, the forms `define-module',
6057 `use-modules', `use-syntax', and `export' have been restricted to only
6058 work on the top level. The forms `define-public' and
6059 `defmacro-public' only export the new binding on the top level. They
6060 behave just like `define' and `defmacro', respectively, when they are
6061 used in a lexical environment.
6063 Also, `export' will no longer silently re-export bindings imported
6064 from a used module. It will emit a `deprecation' warning and will
6065 cease to perform any re-export in the next version. If you actually
6066 want to re-export bindings, use the new `re-export' in place of
6067 `export'. The new `re-export' will not make copies of variables when
6068 rexporting them, as `export' did wrongly.
6070 ** Module system now allows selection and renaming of imported bindings
6072 Previously, when using `use-modules' or the `#:use-module' clause in
6073 the `define-module' form, all the bindings (association of symbols to
6074 values) for imported modules were added to the "current module" on an
6075 as-is basis. This has been changed to allow finer control through two
6076 new facilities: selection and renaming.
6078 You can now select which of the imported module's bindings are to be
6079 visible in the current module by using the `:select' clause. This
6080 clause also can be used to rename individual bindings. For example:
6082 ;; import all bindings no questions asked
6083 (use-modules (ice-9 common-list))
6085 ;; import four bindings, renaming two of them;
6086 ;; the current module sees: every some zonk-y zonk-n
6087 (use-modules ((ice-9 common-list)
6089 (remove-if . zonk-y)
6090 (remove-if-not . zonk-n))))
6092 You can also programmatically rename all selected bindings using the
6093 `:renamer' clause, which specifies a proc that takes a symbol and
6094 returns another symbol. Because it is common practice to use a prefix,
6095 we now provide the convenience procedure `symbol-prefix-proc'. For
6098 ;; import four bindings, renaming two of them specifically,
6099 ;; and all four w/ prefix "CL:";
6100 ;; the current module sees: CL:every CL:some CL:zonk-y CL:zonk-n
6101 (use-modules ((ice-9 common-list)
6103 (remove-if . zonk-y)
6104 (remove-if-not . zonk-n))
6105 :renamer (symbol-prefix-proc 'CL:)))
6107 ;; import four bindings, renaming two of them specifically,
6108 ;; and all four by upcasing.
6109 ;; the current module sees: EVERY SOME ZONK-Y ZONK-N
6110 (define (upcase-symbol sym)
6111 (string->symbol (string-upcase (symbol->string sym))))
6113 (use-modules ((ice-9 common-list)
6115 (remove-if . zonk-y)
6116 (remove-if-not . zonk-n))
6117 :renamer upcase-symbol))
6119 Note that programmatic renaming is done *after* individual renaming.
6120 Also, the above examples show `use-modules', but the same facilities are
6121 available for the `#:use-module' clause of `define-module'.
6123 See manual for more info.
6125 ** The semantics of guardians have changed.
6127 The changes are for the most part compatible. An important criterion
6128 was to keep the typical usage of guardians as simple as before, but to
6129 make the semantics safer and (as a result) more useful.
6131 *** All objects returned from guardians are now properly alive.
6133 It is now guaranteed that any object referenced by an object returned
6134 from a guardian is alive. It's now impossible for a guardian to
6135 return a "contained" object before its "containing" object.
6137 One incompatible (but probably not very important) change resulting
6138 from this is that it is no longer possible to guard objects that
6139 indirectly reference themselves (i.e. are parts of cycles). If you do
6140 so accidentally, you'll get a warning.
6142 *** There are now two types of guardians: greedy and sharing.
6144 If you call (make-guardian #t) or just (make-guardian), you'll get a
6145 greedy guardian, and for (make-guardian #f) a sharing guardian.
6147 Greedy guardians are the default because they are more "defensive".
6148 You can only greedily guard an object once. If you guard an object
6149 more than once, once in a greedy guardian and the rest of times in
6150 sharing guardians, then it is guaranteed that the object won't be
6151 returned from sharing guardians as long as it is greedily guarded
6154 Guardians returned by calls to `make-guardian' can now take one more
6155 optional parameter, which says whether to throw an error in case an
6156 attempt is made to greedily guard an object that is already greedily
6157 guarded. The default is true, i.e. throw an error. If the parameter
6158 is false, the guardian invocation returns #t if guarding was
6159 successful and #f if it wasn't.
6161 Also, since greedy guarding is, in effect, a side-effecting operation
6162 on objects, a new function is introduced: `destroy-guardian!'.
6163 Invoking this function on a guardian renders it unoperative and, if
6164 the guardian is greedy, clears the "greedily guarded" property of the
6165 objects that were guarded by it, thus undoing the side effect.
6167 Note that all this hair is hardly very important, since guardian
6168 objects are usually permanent.
6170 ** Continuations created by call-with-current-continuation now accept
6171 any number of arguments, as required by R5RS.
6173 ** New function `issue-deprecation-warning'
6175 This function is used to display the deprecation messages that are
6176 controlled by GUILE_WARN_DEPRECATION as explained in the README.
6179 (issue-deprecation-warning "`id' is deprecated. Use `identity' instead.")
6183 ;; `id' is deprecated. Use `identity' instead.
6188 ** New syntax `begin-deprecated'
6190 When deprecated features are included (as determined by the configure
6191 option --enable-deprecated), `begin-deprecated' is identical to
6192 `begin'. When deprecated features are excluded, it always evaluates
6193 to `#f', ignoring the body forms.
6195 ** New function `make-object-property'
6197 This function returns a new `procedure with setter' P that can be used
6198 to attach a property to objects. When calling P as
6202 where `obj' is any kind of object, it attaches `val' to `obj' in such
6203 a way that it can be retrieved by calling P as
6207 This function will replace procedure properties, symbol properties and
6208 source properties eventually.
6210 ** Module (ice-9 optargs) now uses keywords instead of `#&'.
6212 Instead of #&optional, #&key, etc you should now use #:optional,
6213 #:key, etc. Since #:optional is a keyword, you can write it as just
6214 :optional when (read-set! keywords 'prefix) is active.
6216 The old reader syntax `#&' is still supported, but deprecated. It
6217 will be removed in the next release.
6219 ** New define-module option: pure
6221 Tells the module system not to include any bindings from the root
6226 (define-module (totally-empty-module)
6229 ** New define-module option: export NAME1 ...
6231 Export names NAME1 ...
6233 This option is required if you want to be able to export bindings from
6234 a module which doesn't import one of `define-public' or `export'.
6238 (define-module (foo)
6240 :use-module (ice-9 r5rs)
6243 ;;; Note that we're pure R5RS below this point!
6248 ** New function: object->string OBJ
6250 Return a Scheme string obtained by printing a given object.
6252 ** New function: port? X
6254 Returns a boolean indicating whether X is a port. Equivalent to
6255 `(or (input-port? X) (output-port? X))'.
6257 ** New function: file-port?
6259 Determines whether a given object is a port that is related to a file.
6261 ** New function: port-for-each proc
6263 Apply PROC to each port in the Guile port table in turn. The return
6264 value is unspecified. More specifically, PROC is applied exactly once
6265 to every port that exists in the system at the time PORT-FOR-EACH is
6266 invoked. Changes to the port table while PORT-FOR-EACH is running
6267 have no effect as far as PORT-FOR-EACH is concerned.
6269 ** New function: dup2 oldfd newfd
6271 A simple wrapper for the `dup2' system call. Copies the file
6272 descriptor OLDFD to descriptor number NEWFD, replacing the
6273 previous meaning of NEWFD. Both OLDFD and NEWFD must be integers.
6274 Unlike for dup->fdes or primitive-move->fdes, no attempt is made
6275 to move away ports which are using NEWFD. The return value is
6278 ** New function: close-fdes fd
6280 A simple wrapper for the `close' system call. Close file
6281 descriptor FD, which must be an integer. Unlike close (*note
6282 close: Ports and File Descriptors.), the file descriptor will be
6283 closed even if a port is using it. The return value is
6286 ** New function: crypt password salt
6288 Encrypts `password' using the standard unix password encryption
6291 ** New function: chroot path
6293 Change the root directory of the running process to `path'.
6295 ** New functions: getlogin, cuserid
6297 Return the login name or the user name of the current effective user
6300 ** New functions: getpriority which who, setpriority which who prio
6302 Get or set the priority of the running process.
6304 ** New function: getpass prompt
6306 Read a password from the terminal, first displaying `prompt' and
6309 ** New function: flock file operation
6311 Set/remove an advisory shared or exclusive lock on `file'.
6313 ** New functions: sethostname name, gethostname
6315 Set or get the hostname of the machine the current process is running
6318 ** New function: mkstemp! tmpl
6320 mkstemp creates a new unique file in the file system and returns a
6321 new buffered port open for reading and writing to the file. TMPL
6322 is a string specifying where the file should be created: it must
6323 end with `XXXXXX' and will be changed in place to return the name
6324 of the temporary file.
6326 ** New function: open-input-string string
6328 Return an input string port which delivers the characters from
6329 `string'. This procedure, together with `open-output-string' and
6330 `get-output-string' implements SRFI-6.
6332 ** New function: open-output-string
6334 Return an output string port which collects all data written to it.
6335 The data can then be retrieved by `get-output-string'.
6337 ** New function: get-output-string
6339 Return the contents of an output string port.
6341 ** New function: identity
6343 Return the argument.
6345 ** socket, connect, accept etc., now have support for IPv6. IPv6 addresses
6346 are represented in Scheme as integers with normal host byte ordering.
6348 ** New function: inet-pton family address
6350 Convert a printable string network address into an integer. Note that
6351 unlike the C version of this function, the result is an integer with
6352 normal host byte ordering. FAMILY can be `AF_INET' or `AF_INET6'.
6355 (inet-pton AF_INET "127.0.0.1") => 2130706433
6356 (inet-pton AF_INET6 "::1") => 1
6358 ** New function: inet-ntop family address
6360 Convert an integer network address into a printable string. Note that
6361 unlike the C version of this function, the input is an integer with
6362 normal host byte ordering. FAMILY can be `AF_INET' or `AF_INET6'.
6365 (inet-ntop AF_INET 2130706433) => "127.0.0.1"
6366 (inet-ntop AF_INET6 (- (expt 2 128) 1)) =>
6367 ffff:ffff:ffff:ffff:ffff:ffff:ffff:ffff
6371 Use `identity' instead.
6377 ** Deprecated: return-it
6381 ** Deprecated: string-character-length
6383 Use `string-length' instead.
6385 ** Deprecated: flags
6387 Use `logior' instead.
6389 ** Deprecated: close-all-ports-except.
6391 This was intended for closing ports in a child process after a fork,
6392 but it has the undesirable side effect of flushing buffers.
6393 port-for-each is more flexible.
6395 ** The (ice-9 popen) module now attempts to set up file descriptors in
6396 the child process from the current Scheme ports, instead of using the
6397 current values of file descriptors 0, 1, and 2 in the parent process.
6399 ** Removed function: builtin-weak-bindings
6401 There is no such concept as a weak binding any more.
6403 ** Removed constants: bignum-radix, scm-line-incrementors
6405 ** define-method: New syntax mandatory.
6407 The new method syntax is now mandatory:
6409 (define-method (NAME ARG-SPEC ...) BODY ...)
6410 (define-method (NAME ARG-SPEC ... . REST-ARG) BODY ...)
6412 ARG-SPEC ::= ARG-NAME | (ARG-NAME TYPE)
6413 REST-ARG ::= ARG-NAME
6415 If you have old code using the old syntax, import
6416 (oop goops old-define-method) before (oop goops) as in:
6418 (use-modules (oop goops old-define-method) (oop goops))
6420 ** Deprecated function: builtin-variable
6421 Removed function: builtin-bindings
6423 There is no longer a distinction between builtin or other variables.
6424 Use module system operations for all variables.
6426 ** Lazy-catch handlers are no longer allowed to return.
6428 That is, a call to `throw', `error', etc is now guaranteed to not
6431 ** Bugfixes for (ice-9 getopt-long)
6433 This module is now tested using test-suite/tests/getopt-long.test.
6434 The following bugs have been fixed:
6436 *** Parsing for options that are specified to have `optional' args now checks
6437 if the next element is an option instead of unconditionally taking it as the
6440 *** An error is now thrown for `--opt=val' when the option description
6441 does not specify `(value #t)' or `(value optional)'. This condition used to
6442 be accepted w/o error, contrary to the documentation.
6444 *** The error message for unrecognized options is now more informative.
6445 It used to be "not a record", an artifact of the implementation.
6447 *** The error message for `--opt' terminating the arg list (no value), when
6448 `(value #t)' is specified, is now more informative. It used to be "not enough
6451 *** "Clumped" single-char args now preserve trailing string, use it as arg.
6452 The expansion used to be like so:
6454 ("-abc5d" "--xyz") => ("-a" "-b" "-c" "--xyz")
6456 Note that the "5d" is dropped. Now it is like so:
6458 ("-abc5d" "--xyz") => ("-a" "-b" "-c" "5d" "--xyz")
6460 This enables single-char options to have adjoining arguments as long as their
6461 constituent characters are not potential single-char options.
6463 ** (ice-9 session) procedure `arity' now works with (ice-9 optargs) `lambda*'
6465 The `lambda*' and derivative forms in (ice-9 optargs) now set a procedure
6466 property `arglist', which can be retrieved by `arity'. The result is that
6467 `arity' can give more detailed information than before:
6471 guile> (use-modules (ice-9 optargs))
6472 guile> (define* (foo #:optional a b c) a)
6474 0 or more arguments in `lambda*:G0'.
6479 3 optional arguments: `a', `b' and `c'.
6480 guile> (define* (bar a b #:key c d #:allow-other-keys) a)
6482 2 required arguments: `a' and `b', 2 keyword arguments: `c'
6483 and `d', other keywords allowed.
6484 guile> (define* (baz a b #:optional c #:rest r) a)
6486 2 required arguments: `a' and `b', 1 optional argument: `c',
6489 * Changes to the C interface
6491 ** Types have been renamed from scm_*_t to scm_t_*.
6493 This has been done for POSIX sake. It reserves identifiers ending
6494 with "_t". What a concept.
6496 The old names are still available with status `deprecated'.
6498 ** scm_t_bits (former scm_bits_t) is now a unsigned type.
6500 ** Deprecated features have been removed.
6504 SCM_INPORTP, SCM_OUTPORTP SCM_ICHRP, SCM_ICHR, SCM_MAKICHR
6505 SCM_SETJMPBUF SCM_NSTRINGP SCM_NRWSTRINGP SCM_NVECTORP SCM_DOUBLE_CELLP
6507 *** C Functions removed
6509 scm_sysmissing scm_tag scm_tc16_flo scm_tc_flo
6510 scm_fseek - replaced by scm_seek.
6511 gc-thunk - replaced by after-gc-hook.
6512 gh_int2scmb - replaced by gh_bool2scm.
6513 scm_tc_dblr - replaced by scm_tc16_real.
6514 scm_tc_dblc - replaced by scm_tc16_complex.
6515 scm_list_star - replaced by scm_cons_star.
6517 ** Deprecated: scm_makfromstr
6519 Use scm_mem2string instead.
6521 ** Deprecated: scm_make_shared_substring
6523 Explicit shared substrings will disappear from Guile.
6525 Instead, "normal" strings will be implemented using sharing
6526 internally, combined with a copy-on-write strategy.
6528 ** Deprecated: scm_read_only_string_p
6530 The concept of read-only strings will disappear in next release of
6533 ** Deprecated: scm_sloppy_memq, scm_sloppy_memv, scm_sloppy_member
6535 Instead, use scm_c_memq or scm_memq, scm_memv, scm_member.
6537 ** New functions: scm_call_0, scm_call_1, scm_call_2, scm_call_3
6539 Call a procedure with the indicated number of arguments. See "Fly
6540 Evaluation" in the manual.
6542 ** New functions: scm_apply_0, scm_apply_1, scm_apply_2, scm_apply_3
6544 Call a procedure with the indicated number of arguments and a list of
6545 further arguments. See "Fly Evaluation" in the manual.
6547 ** New functions: scm_list_1, scm_list_2, scm_list_3, scm_list_4, scm_list_5
6549 Create a list of the given number of elements. See "List
6550 Constructors" in the manual.
6552 ** Renamed function: scm_listify has been replaced by scm_list_n.
6554 ** Deprecated macros: SCM_LIST0, SCM_LIST1, SCM_LIST2, SCM_LIST3, SCM_LIST4,
6555 SCM_LIST5, SCM_LIST6, SCM_LIST7, SCM_LIST8, SCM_LIST9.
6557 Use functions scm_list_N instead.
6559 ** New function: scm_c_read (SCM port, void *buffer, scm_sizet size)
6561 Used by an application to read arbitrary number of bytes from a port.
6562 Same semantics as libc read, except that scm_c_read only returns less
6563 than SIZE bytes if at end-of-file.
6565 Warning: Doesn't update port line and column counts!
6567 ** New function: scm_c_write (SCM port, const void *ptr, scm_sizet size)
6569 Used by an application to write arbitrary number of bytes to an SCM
6570 port. Similar semantics as libc write. However, unlike libc
6571 write, scm_c_write writes the requested number of bytes and has no
6574 Warning: Doesn't update port line and column counts!
6576 ** New function: scm_init_guile ()
6578 In contrast to scm_boot_guile, scm_init_guile will return normally
6579 after initializing Guile. It is not available on all systems, tho.
6581 ** New functions: scm_str2symbol, scm_mem2symbol
6583 The function scm_str2symbol takes a const char* pointing to a zero-terminated
6584 field of characters and creates a scheme symbol object from that C string.
6585 The function scm_mem2symbol takes a const char* and a number of characters and
6586 creates a symbol from the characters in that memory area.
6588 ** New functions: scm_primitive_make_property
6589 scm_primitive_property_ref
6590 scm_primitive_property_set_x
6591 scm_primitive_property_del_x
6593 These functions implement a new way to deal with object properties.
6594 See libguile/properties.c for their documentation.
6596 ** New function: scm_done_free (long size)
6598 This function is the inverse of scm_done_malloc. Use it to report the
6599 amount of smob memory you free. The previous method, which involved
6600 calling scm_done_malloc with negative argument, was somewhat
6601 unintuitive (and is still available, of course).
6603 ** New function: scm_c_memq (SCM obj, SCM list)
6605 This function provides a fast C level alternative for scm_memq for the case
6606 that the list parameter is known to be a proper list. The function is a
6607 replacement for scm_sloppy_memq, but is stricter in its requirements on its
6608 list input parameter, since for anything else but a proper list the function's
6609 behaviour is undefined - it may even crash or loop endlessly. Further, for
6610 the case that the object is not found in the list, scm_c_memq returns #f which
6611 is similar to scm_memq, but different from scm_sloppy_memq's behaviour.
6613 ** New functions: scm_remember_upto_here_1, scm_remember_upto_here_2,
6614 scm_remember_upto_here
6616 These functions replace the function scm_remember.
6618 ** Deprecated function: scm_remember
6620 Use one of the new functions scm_remember_upto_here_1,
6621 scm_remember_upto_here_2 or scm_remember_upto_here instead.
6623 ** New function: scm_allocate_string
6625 This function replaces the function scm_makstr.
6627 ** Deprecated function: scm_makstr
6629 Use the new function scm_allocate_string instead.
6631 ** New global variable scm_gc_running_p introduced.
6633 Use this variable to find out if garbage collection is being executed. Up to
6634 now applications have used scm_gc_heap_lock to test if garbage collection was
6635 running, which also works because of the fact that up to know only the garbage
6636 collector has set this variable. But, this is an implementation detail that
6637 may change. Further, scm_gc_heap_lock is not set throughout gc, thus the use
6638 of this variable is (and has been) not fully safe anyway.
6640 ** New macros: SCM_BITVECTOR_MAX_LENGTH, SCM_UVECTOR_MAX_LENGTH
6642 Use these instead of SCM_LENGTH_MAX.
6644 ** New macros: SCM_CONTINUATION_LENGTH, SCM_CCLO_LENGTH, SCM_STACK_LENGTH,
6645 SCM_STRING_LENGTH, SCM_SYMBOL_LENGTH, SCM_UVECTOR_LENGTH,
6646 SCM_BITVECTOR_LENGTH, SCM_VECTOR_LENGTH.
6648 Use these instead of SCM_LENGTH.
6650 ** New macros: SCM_SET_CONTINUATION_LENGTH, SCM_SET_STRING_LENGTH,
6651 SCM_SET_SYMBOL_LENGTH, SCM_SET_VECTOR_LENGTH, SCM_SET_UVECTOR_LENGTH,
6652 SCM_SET_BITVECTOR_LENGTH
6654 Use these instead of SCM_SETLENGTH
6656 ** New macros: SCM_STRING_CHARS, SCM_SYMBOL_CHARS, SCM_CCLO_BASE,
6657 SCM_VECTOR_BASE, SCM_UVECTOR_BASE, SCM_BITVECTOR_BASE, SCM_COMPLEX_MEM,
6660 Use these instead of SCM_CHARS, SCM_UCHARS, SCM_ROCHARS, SCM_ROUCHARS or
6663 ** New macros: SCM_SET_BIGNUM_BASE, SCM_SET_STRING_CHARS,
6664 SCM_SET_SYMBOL_CHARS, SCM_SET_UVECTOR_BASE, SCM_SET_BITVECTOR_BASE,
6667 Use these instead of SCM_SETCHARS.
6669 ** New macro: SCM_BITVECTOR_P
6671 ** New macro: SCM_STRING_COERCE_0TERMINATION_X
6673 Use instead of SCM_COERCE_SUBSTR.
6675 ** New macros: SCM_DIR_OPEN_P, SCM_DIR_FLAG_OPEN
6677 For directory objects, use these instead of SCM_OPDIRP and SCM_OPN.
6679 ** Deprecated macros: SCM_OUTOFRANGE, SCM_NALLOC, SCM_HUP_SIGNAL,
6680 SCM_INT_SIGNAL, SCM_FPE_SIGNAL, SCM_BUS_SIGNAL, SCM_SEGV_SIGNAL,
6681 SCM_ALRM_SIGNAL, SCM_GC_SIGNAL, SCM_TICK_SIGNAL, SCM_SIG_ORD,
6682 SCM_ORD_SIG, SCM_NUM_SIGS, SCM_SYMBOL_SLOTS, SCM_SLOTS, SCM_SLOPPY_STRINGP,
6683 SCM_VALIDATE_STRINGORSUBSTR, SCM_FREEP, SCM_NFREEP, SCM_CHARS, SCM_UCHARS,
6684 SCM_VALIDATE_ROSTRING, SCM_VALIDATE_ROSTRING_COPY,
6685 SCM_VALIDATE_NULLORROSTRING_COPY, SCM_ROLENGTH, SCM_LENGTH, SCM_HUGE_LENGTH,
6686 SCM_SUBSTRP, SCM_SUBSTR_STR, SCM_SUBSTR_OFFSET, SCM_COERCE_SUBSTR,
6687 SCM_ROSTRINGP, SCM_RWSTRINGP, SCM_VALIDATE_RWSTRING, SCM_ROCHARS,
6688 SCM_ROUCHARS, SCM_SETLENGTH, SCM_SETCHARS, SCM_LENGTH_MAX, SCM_GC8MARKP,
6689 SCM_SETGC8MARK, SCM_CLRGC8MARK, SCM_GCTYP16, SCM_GCCDR, SCM_SUBR_DOC,
6690 SCM_OPDIRP, SCM_VALIDATE_OPDIR, SCM_WTA, RETURN_SCM_WTA, SCM_CONST_LONG,
6691 SCM_WNA, SCM_FUNC_NAME, SCM_VALIDATE_NUMBER_COPY,
6692 SCM_VALIDATE_NUMBER_DEF_COPY, SCM_SLOPPY_CONSP, SCM_SLOPPY_NCONSP,
6693 SCM_SETAND_CDR, SCM_SETOR_CDR, SCM_SETAND_CAR, SCM_SETOR_CAR
6695 Use SCM_ASSERT_RANGE or SCM_VALIDATE_XXX_RANGE instead of SCM_OUTOFRANGE.
6696 Use scm_memory_error instead of SCM_NALLOC.
6697 Use SCM_STRINGP instead of SCM_SLOPPY_STRINGP.
6698 Use SCM_VALIDATE_STRING instead of SCM_VALIDATE_STRINGORSUBSTR.
6699 Use SCM_FREE_CELL_P instead of SCM_FREEP/SCM_NFREEP
6700 Use a type specific accessor macro instead of SCM_CHARS/SCM_UCHARS.
6701 Use a type specific accessor instead of SCM(_|_RO|_HUGE_)LENGTH.
6702 Use SCM_VALIDATE_(SYMBOL|STRING) instead of SCM_VALIDATE_ROSTRING.
6703 Use SCM_STRING_COERCE_0TERMINATION_X instead of SCM_COERCE_SUBSTR.
6704 Use SCM_STRINGP or SCM_SYMBOLP instead of SCM_ROSTRINGP.
6705 Use SCM_STRINGP instead of SCM_RWSTRINGP.
6706 Use SCM_VALIDATE_STRING instead of SCM_VALIDATE_RWSTRING.
6707 Use SCM_STRING_CHARS instead of SCM_ROCHARS.
6708 Use SCM_STRING_UCHARS instead of SCM_ROUCHARS.
6709 Use a type specific setter macro instead of SCM_SETLENGTH.
6710 Use a type specific setter macro instead of SCM_SETCHARS.
6711 Use a type specific length macro instead of SCM_LENGTH_MAX.
6712 Use SCM_GCMARKP instead of SCM_GC8MARKP.
6713 Use SCM_SETGCMARK instead of SCM_SETGC8MARK.
6714 Use SCM_CLRGCMARK instead of SCM_CLRGC8MARK.
6715 Use SCM_TYP16 instead of SCM_GCTYP16.
6716 Use SCM_CDR instead of SCM_GCCDR.
6717 Use SCM_DIR_OPEN_P instead of SCM_OPDIRP.
6718 Use SCM_MISC_ERROR or SCM_WRONG_TYPE_ARG instead of SCM_WTA.
6719 Use SCM_MISC_ERROR or SCM_WRONG_TYPE_ARG instead of RETURN_SCM_WTA.
6720 Use SCM_VCELL_INIT instead of SCM_CONST_LONG.
6721 Use SCM_WRONG_NUM_ARGS instead of SCM_WNA.
6722 Use SCM_CONSP instead of SCM_SLOPPY_CONSP.
6723 Use !SCM_CONSP instead of SCM_SLOPPY_NCONSP.
6725 ** Removed function: scm_struct_init
6727 ** Removed variable: scm_symhash_dim
6729 ** Renamed function: scm_make_cont has been replaced by
6730 scm_make_continuation, which has a different interface.
6732 ** Deprecated function: scm_call_catching_errors
6734 Use scm_catch or scm_lazy_catch from throw.[ch] instead.
6736 ** Deprecated function: scm_strhash
6738 Use scm_string_hash instead.
6740 ** Deprecated function: scm_vector_set_length_x
6742 Instead, create a fresh vector of the desired size and copy the contents.
6744 ** scm_gensym has changed prototype
6746 scm_gensym now only takes one argument.
6748 ** Deprecated type tags: scm_tc7_ssymbol, scm_tc7_msymbol, scm_tcs_symbols,
6751 There is now only a single symbol type scm_tc7_symbol.
6752 The tag scm_tc7_lvector was not used anyway.
6754 ** Deprecated function: scm_make_smob_type_mfpe, scm_set_smob_mfpe.
6756 Use scm_make_smob_type and scm_set_smob_XXX instead.
6758 ** New function scm_set_smob_apply.
6760 This can be used to set an apply function to a smob type.
6762 ** Deprecated function: scm_strprint_obj
6764 Use scm_object_to_string instead.
6766 ** Deprecated function: scm_wta
6768 Use scm_wrong_type_arg, or another appropriate error signalling function
6771 ** Explicit support for obarrays has been deprecated.
6773 Use `scm_str2symbol' and the generic hashtable functions instead.
6775 ** The concept of `vcells' has been deprecated.
6777 The data type `variable' is now used exclusively. `Vcells' have been
6778 a low-level concept so you are likely not affected by this change.
6780 *** Deprecated functions: scm_sym2vcell, scm_sysintern,
6781 scm_sysintern0, scm_symbol_value0, scm_intern, scm_intern0.
6783 Use scm_c_define or scm_c_lookup instead, as appropriate.
6785 *** New functions: scm_c_module_lookup, scm_c_lookup,
6786 scm_c_module_define, scm_c_define, scm_module_lookup, scm_lookup,
6787 scm_module_define, scm_define.
6789 These functions work with variables instead of with vcells.
6791 ** New functions for creating and defining `subr's and `gsubr's.
6793 The new functions more clearly distinguish between creating a subr (or
6794 gsubr) object and adding it to the current module.
6796 These new functions are available: scm_c_make_subr, scm_c_define_subr,
6797 scm_c_make_subr_with_generic, scm_c_define_subr_with_generic,
6798 scm_c_make_gsubr, scm_c_define_gsubr, scm_c_make_gsubr_with_generic,
6799 scm_c_define_gsubr_with_generic.
6801 ** Deprecated functions: scm_make_subr, scm_make_subr_opt,
6802 scm_make_subr_with_generic, scm_make_gsubr,
6803 scm_make_gsubr_with_generic.
6805 Use the new ones from above instead.
6807 ** C interface to the module system has changed.
6809 While we suggest that you avoid as many explicit module system
6810 operations from C as possible for the time being, the C interface has
6811 been made more similar to the high-level Scheme module system.
6813 *** New functions: scm_c_define_module, scm_c_use_module,
6814 scm_c_export, scm_c_resolve_module.
6816 They mostly work like their Scheme namesakes. scm_c_define_module
6817 takes a function that is called a context where the new module is
6820 *** Deprecated functions: scm_the_root_module, scm_make_module,
6821 scm_ensure_user_module, scm_load_scheme_module.
6823 Use the new functions instead.
6825 ** Renamed function: scm_internal_with_fluids becomes
6828 scm_internal_with_fluids is available as a deprecated function.
6830 ** New function: scm_c_with_fluid.
6832 Just like scm_c_with_fluids, but takes one fluid and one value instead
6835 ** Deprecated typedefs: long_long, ulong_long.
6837 They are of questionable utility and they pollute the global
6840 ** Deprecated typedef: scm_sizet
6842 It is of questionable utility now that Guile requires ANSI C, and is
6845 ** Deprecated typedefs: scm_port_rw_active, scm_port,
6846 scm_ptob_descriptor, scm_debug_info, scm_debug_frame, scm_fport,
6847 scm_option, scm_rstate, scm_rng, scm_array, scm_array_dim.
6849 Made more compliant with the naming policy by adding a _t at the end.
6851 ** Deprecated functions: scm_mkbig, scm_big2num, scm_adjbig,
6852 scm_normbig, scm_copybig, scm_2ulong2big, scm_dbl2big, scm_big2dbl
6854 With the exception of the mysterious scm_2ulong2big, they are still
6855 available under new names (scm_i_mkbig etc). These functions are not
6856 intended to be used in user code. You should avoid dealing with
6857 bignums directly, and should deal with numbers in general (which can
6860 ** Change in behavior: scm_num2long, scm_num2ulong
6862 The scm_num2[u]long functions don't any longer accept an inexact
6863 argument. This change in behavior is motivated by concordance with
6864 R5RS: It is more common that a primitive doesn't want to accept an
6865 inexact for an exact.
6867 ** New functions: scm_short2num, scm_ushort2num, scm_int2num,
6868 scm_uint2num, scm_size2num, scm_ptrdiff2num, scm_num2short,
6869 scm_num2ushort, scm_num2int, scm_num2uint, scm_num2ptrdiff,
6872 These are conversion functions between the various ANSI C integral
6873 types and Scheme numbers. NOTE: The scm_num2xxx functions don't
6874 accept an inexact argument.
6876 ** New functions: scm_float2num, scm_double2num,
6877 scm_num2float, scm_num2double.
6879 These are conversion functions between the two ANSI C float types and
6882 ** New number validation macros:
6883 SCM_NUM2{SIZE,PTRDIFF,SHORT,USHORT,INT,UINT}[_DEF]
6887 ** New functions: scm_gc_protect_object, scm_gc_unprotect_object
6889 These are just nicer-named old scm_protect_object and
6890 scm_unprotect_object.
6892 ** Deprecated functions: scm_protect_object, scm_unprotect_object
6894 ** New functions: scm_gc_[un]register_root, scm_gc_[un]register_roots
6896 These functions can be used to register pointers to locations that
6899 ** Deprecated function: scm_create_hook.
6901 Its sins are: misleading name, non-modularity and lack of general
6905 Changes since Guile 1.3.4:
6907 * Changes to the distribution
6909 ** Trees from nightly snapshots and CVS now require you to run autogen.sh.
6911 We've changed the way we handle generated files in the Guile source
6912 repository. As a result, the procedure for building trees obtained
6913 from the nightly FTP snapshots or via CVS has changed:
6914 - You must have appropriate versions of autoconf, automake, and
6915 libtool installed on your system. See README for info on how to
6916 obtain these programs.
6917 - Before configuring the tree, you must first run the script
6918 `autogen.sh' at the top of the source tree.
6920 The Guile repository used to contain not only source files, written by
6921 humans, but also some generated files, like configure scripts and
6922 Makefile.in files. Even though the contents of these files could be
6923 derived mechanically from other files present, we thought it would
6924 make the tree easier to build if we checked them into CVS.
6926 However, this approach means that minor differences between
6927 developer's installed tools and habits affected the whole team.
6928 So we have removed the generated files from the repository, and
6929 added the autogen.sh script, which will reconstruct them
6933 ** configure now has experimental options to remove support for certain
6936 --disable-arrays omit array and uniform array support
6937 --disable-posix omit posix interfaces
6938 --disable-networking omit networking interfaces
6939 --disable-regex omit regular expression interfaces
6941 These are likely to become separate modules some day.
6943 ** New configure option --enable-debug-freelist
6945 This enables a debugging version of SCM_NEWCELL(), and also registers
6946 an extra primitive, the setter `gc-set-debug-check-freelist!'.
6948 Configure with the --enable-debug-freelist option to enable
6949 the gc-set-debug-check-freelist! primitive, and then use:
6951 (gc-set-debug-check-freelist! #t) # turn on checking of the freelist
6952 (gc-set-debug-check-freelist! #f) # turn off checking
6954 Checking of the freelist forces a traversal of the freelist and
6955 a garbage collection before each allocation of a cell. This can
6956 slow down the interpreter dramatically, so the setter should be used to
6957 turn on this extra processing only when necessary.
6959 ** New configure option --enable-debug-malloc
6961 Include code for debugging of calls to scm_must_malloc/realloc/free.
6965 1. objects freed by scm_must_free has been mallocated by scm_must_malloc
6966 2. objects reallocated by scm_must_realloc has been allocated by
6968 3. reallocated objects are reallocated with the same what string
6970 But, most importantly, it records the number of allocated objects of
6971 each kind. This is useful when searching for memory leaks.
6973 A Guile compiled with this option provides the primitive
6974 `malloc-stats' which returns an alist with pairs of kind and the
6975 number of objects of that kind.
6977 ** All includes are now referenced relative to the root directory
6979 Since some users have had problems with mixups between Guile and
6980 system headers, we have decided to always refer to Guile headers via
6981 their parent directories. This essentially creates a "private name
6982 space" for Guile headers. This means that the compiler only is given
6983 -I options for the root build and root source directory.
6985 ** Header files kw.h and genio.h have been removed.
6987 ** The module (ice-9 getopt-gnu-style) has been removed.
6989 ** New module (ice-9 documentation)
6991 Implements the interface to documentation strings associated with
6994 ** New module (ice-9 time)
6996 Provides a macro `time', which displays execution time of a given form.
6998 ** New module (ice-9 history)
7000 Loading this module enables value history in the repl.
7002 * Changes to the stand-alone interpreter
7004 ** New command line option --debug
7006 Start Guile with debugging evaluator and backtraces enabled.
7008 This is useful when debugging your .guile init file or scripts.
7010 ** New help facility
7012 Usage: (help NAME) gives documentation about objects named NAME (a symbol)
7013 (help REGEXP) ditto for objects with names matching REGEXP (a string)
7014 (help 'NAME) gives documentation for NAME, even if it is not an object
7015 (help ,EXPR) gives documentation for object returned by EXPR
7016 (help (my module)) gives module commentary for `(my module)'
7017 (help) gives this text
7019 `help' searches among bindings exported from loaded modules, while
7020 `apropos' searches among bindings visible from the "current" module.
7022 Examples: (help help)
7024 (help "output-string")
7026 ** `help' and `apropos' now prints full module names
7028 ** Dynamic linking now uses libltdl from the libtool package.
7030 The old system dependent code for doing dynamic linking has been
7031 replaced with calls to the libltdl functions which do all the hairy
7034 The major improvement is that you can now directly pass libtool
7035 library names like "libfoo.la" to `dynamic-link' and `dynamic-link'
7036 will be able to do the best shared library job you can get, via
7039 The way dynamic libraries are found has changed and is not really
7040 portable across platforms, probably. It is therefore recommended to
7041 use absolute filenames when possible.
7043 If you pass a filename without an extension to `dynamic-link', it will
7044 try a few appropriate ones. Thus, the most platform ignorant way is
7045 to specify a name like "libfoo", without any directories and
7048 ** Guile COOP threads are now compatible with LinuxThreads
7050 Previously, COOP threading wasn't possible in applications linked with
7051 Linux POSIX threads due to their use of the stack pointer to find the
7052 thread context. This has now been fixed with a workaround which uses
7053 the pthreads to allocate the stack.
7055 ** New primitives: `pkgdata-dir', `site-dir', `library-dir'
7057 ** Positions of erring expression in scripts
7059 With version 1.3.4, the location of the erring expression in Guile
7060 scipts is no longer automatically reported. (This should have been
7061 documented before the 1.3.4 release.)
7063 You can get this information by enabling recording of positions of
7064 source expressions and running the debugging evaluator. Put this at
7065 the top of your script (or in your "site" file):
7067 (read-enable 'positions)
7068 (debug-enable 'debug)
7070 ** Backtraces in scripts
7072 It is now possible to get backtraces in scripts.
7076 (debug-enable 'debug 'backtrace)
7078 at the top of the script.
7080 (The first options enables the debugging evaluator.
7081 The second enables backtraces.)
7083 ** Part of module system symbol lookup now implemented in C
7085 The eval closure of most modules is now implemented in C. Since this
7086 was one of the bottlenecks for loading speed, Guile now loads code
7087 substantially faster than before.
7089 ** Attempting to get the value of an unbound variable now produces
7090 an exception with a key of 'unbound-variable instead of 'misc-error.
7092 ** The initial default output port is now unbuffered if it's using a
7093 tty device. Previously in this situation it was line-buffered.
7095 ** New hook: after-gc-hook
7097 after-gc-hook takes over the role of gc-thunk. This hook is run at
7098 the first SCM_TICK after a GC. (Thus, the code is run at the same
7099 point during evaluation as signal handlers.)
7101 Note that this hook should be used only for diagnostic and debugging
7102 purposes. It is not certain that it will continue to be well-defined
7103 when this hook is run in the future.
7105 C programmers: Note the new C level hooks scm_before_gc_c_hook,
7106 scm_before_sweep_c_hook, scm_after_gc_c_hook.
7108 ** Improvements to garbage collector
7110 Guile 1.4 has a new policy for triggering heap allocation and
7111 determining the sizes of heap segments. It fixes a number of problems
7114 1. The new policy can handle two separate pools of cells
7115 (2-word/4-word) better. (The old policy would run wild, allocating
7116 more and more memory for certain programs.)
7118 2. The old code would sometimes allocate far too much heap so that the
7119 Guile process became gigantic. The new code avoids this.
7121 3. The old code would sometimes allocate too little so that few cells
7122 were freed at GC so that, in turn, too much time was spent in GC.
7124 4. The old code would often trigger heap allocation several times in a
7125 row. (The new scheme predicts how large the segments needs to be
7126 in order not to need further allocation.)
7128 All in all, the new GC policy will make larger applications more
7131 The new GC scheme also is prepared for POSIX threading. Threads can
7132 allocate private pools of cells ("clusters") with just a single
7133 function call. Allocation of single cells from such a cluster can
7134 then proceed without any need of inter-thread synchronization.
7136 ** New environment variables controlling GC parameters
7138 GUILE_MAX_SEGMENT_SIZE Maximal segment size
7141 Allocation of 2-word cell heaps:
7143 GUILE_INIT_SEGMENT_SIZE_1 Size of initial heap segment in bytes
7146 GUILE_MIN_YIELD_1 Minimum number of freed cells at each
7147 GC in percent of total heap size
7150 Allocation of 4-word cell heaps
7151 (used for real numbers and misc other objects):
7153 GUILE_INIT_SEGMENT_SIZE_2, GUILE_MIN_YIELD_2
7155 (See entry "Way for application to customize GC parameters" under
7156 section "Changes to the scm_ interface" below.)
7158 ** Guile now implements reals using 4-word cells
7160 This speeds up computation with reals. (They were earlier allocated
7161 with `malloc'.) There is still some room for optimizations, however.
7163 ** Some further steps toward POSIX thread support have been taken
7165 *** Guile's critical sections (SCM_DEFER/ALLOW_INTS)
7166 don't have much effect any longer, and many of them will be removed in
7170 are only handled at the top of the evaluator loop, immediately after
7171 I/O, and in scm_equalp.
7173 *** The GC can allocate thread private pools of pairs.
7175 * Changes to Scheme functions and syntax
7177 ** close-input-port and close-output-port are now R5RS
7179 These procedures have been turned into primitives and have R5RS behaviour.
7181 ** New procedure: simple-format PORT MESSAGE ARG1 ...
7183 (ice-9 boot) makes `format' an alias for `simple-format' until possibly
7184 extended by the more sophisticated version in (ice-9 format)
7186 (simple-format port message . args)
7187 Write MESSAGE to DESTINATION, defaulting to `current-output-port'.
7188 MESSAGE can contain ~A (was %s) and ~S (was %S) escapes. When printed,
7189 the escapes are replaced with corresponding members of ARGS:
7190 ~A formats using `display' and ~S formats using `write'.
7191 If DESTINATION is #t, then use the `current-output-port',
7192 if DESTINATION is #f, then return a string containing the formatted text.
7193 Does not add a trailing newline."
7195 ** string-ref: the second argument is no longer optional.
7197 ** string, list->string: no longer accept strings in their arguments,
7198 only characters, for compatibility with R5RS.
7200 ** New procedure: port-closed? PORT
7201 Returns #t if PORT is closed or #f if it is open.
7203 ** Deprecated: list*
7205 The list* functionality is now provided by cons* (SRFI-1 compliant)
7207 ** New procedure: cons* ARG1 ARG2 ... ARGn
7209 Like `list', but the last arg provides the tail of the constructed list,
7210 returning (cons ARG1 (cons ARG2 (cons ... ARGn))).
7212 Requires at least one argument. If given one argument, that argument
7213 is returned as result.
7215 This function is called `list*' in some other Schemes and in Common LISP.
7217 ** Removed deprecated: serial-map, serial-array-copy!, serial-array-map!
7219 ** New procedure: object-documentation OBJECT
7221 Returns the documentation string associated with OBJECT. The
7222 procedure uses a caching mechanism so that subsequent lookups are
7225 Exported by (ice-9 documentation).
7227 ** module-name now returns full names of modules
7229 Previously, only the last part of the name was returned (`session' for
7230 `(ice-9 session)'). Ex: `(ice-9 session)'.
7232 * Changes to the gh_ interface
7234 ** Deprecated: gh_int2scmb
7236 Use gh_bool2scm instead.
7238 * Changes to the scm_ interface
7240 ** Guile primitives now carry docstrings!
7242 Thanks to Greg Badros!
7244 ** Guile primitives are defined in a new way: SCM_DEFINE/SCM_DEFINE1/SCM_PROC
7246 Now Guile primitives are defined using the SCM_DEFINE/SCM_DEFINE1/SCM_PROC
7247 macros and must contain a docstring that is extracted into foo.doc using a new
7248 guile-doc-snarf script (that uses guile-doc-snarf.awk).
7250 However, a major overhaul of these macros is scheduled for the next release of
7253 ** Guile primitives use a new technique for validation of arguments
7255 SCM_VALIDATE_* macros are defined to ease the redundancy and improve
7256 the readability of argument checking.
7258 ** All (nearly?) K&R prototypes for functions replaced with ANSI C equivalents.
7260 ** New macros: SCM_PACK, SCM_UNPACK
7262 Compose/decompose an SCM value.
7264 The SCM type is now treated as an abstract data type and may be defined as a
7265 long, a void* or as a struct, depending on the architecture and compile time
7266 options. This makes it easier to find several types of bugs, for example when
7267 SCM values are treated as integers without conversion. Values of the SCM type
7268 should be treated as "atomic" values. These macros are used when
7269 composing/decomposing an SCM value, either because you want to access
7270 individual bits, or because you want to treat it as an integer value.
7272 E.g., in order to set bit 7 in an SCM value x, use the expression
7274 SCM_PACK (SCM_UNPACK (x) | 0x80)
7276 ** The name property of hooks is deprecated.
7277 Thus, the use of SCM_HOOK_NAME and scm_make_hook_with_name is deprecated.
7279 You can emulate this feature by using object properties.
7281 ** Deprecated macros: SCM_INPORTP, SCM_OUTPORTP, SCM_CRDY, SCM_ICHRP,
7282 SCM_ICHR, SCM_MAKICHR, SCM_SETJMPBUF, SCM_NSTRINGP, SCM_NRWSTRINGP,
7285 These macros will be removed in a future release of Guile.
7287 ** The following types, functions and macros from numbers.h are deprecated:
7288 scm_dblproc, SCM_UNEGFIXABLE, SCM_FLOBUFLEN, SCM_INEXP, SCM_CPLXP, SCM_REAL,
7289 SCM_IMAG, SCM_REALPART, scm_makdbl, SCM_SINGP, SCM_NUM2DBL, SCM_NO_BIGDIG
7291 ** Port internals: the rw_random variable in the scm_port structure
7292 must be set to non-zero in any random access port. In recent Guile
7293 releases it was only set for bidirectional random-access ports.
7295 ** Port internals: the seek ptob procedure is now responsible for
7296 resetting the buffers if required. The change was made so that in the
7297 special case of reading the current position (i.e., seek p 0 SEEK_CUR)
7298 the fport and strport ptobs can avoid resetting the buffers,
7299 in particular to avoid discarding unread chars. An existing port
7300 type can be fixed by adding something like the following to the
7301 beginning of the ptob seek procedure:
7303 if (pt->rw_active == SCM_PORT_READ)
7304 scm_end_input (object);
7305 else if (pt->rw_active == SCM_PORT_WRITE)
7306 ptob->flush (object);
7308 although to actually avoid resetting the buffers and discard unread
7309 chars requires further hacking that depends on the characteristics
7312 ** Deprecated functions: scm_fseek, scm_tag
7314 These functions are no longer used and will be removed in a future version.
7316 ** The scm_sysmissing procedure is no longer used in libguile.
7317 Unless it turns out to be unexpectedly useful to somebody, it will be
7318 removed in a future version.
7320 ** The format of error message strings has changed
7322 The two C procedures: scm_display_error and scm_error, as well as the
7323 primitive `scm-error', now use scm_simple_format to do their work.
7324 This means that the message strings of all code must be updated to use
7325 ~A where %s was used before, and ~S where %S was used before.
7327 During the period when there still are a lot of old Guiles out there,
7328 you might want to support both old and new versions of Guile.
7330 There are basically two methods to achieve this. Both methods use
7333 AC_CHECK_FUNCS(scm_simple_format)
7335 in your configure.in.
7337 Method 1: Use the string concatenation features of ANSI C's
7342 #ifdef HAVE_SCM_SIMPLE_FORMAT
7348 Then represent each of your error messages using a preprocessor macro:
7350 #define E_SPIDER_ERROR "There's a spider in your " ## FMT_S ## "!!!"
7354 (define fmt-s (if (defined? 'simple-format) "~S" "%S"))
7355 (define make-message string-append)
7357 (define e-spider-error (make-message "There's a spider in your " fmt-s "!!!"))
7359 Method 2: Use the oldfmt function found in doc/oldfmt.c.
7363 scm_misc_error ("picnic", scm_c_oldfmt0 ("There's a spider in your ~S!!!"),
7368 (scm-error 'misc-error "picnic" (oldfmt "There's a spider in your ~S!!!")
7372 ** Deprecated: coop_mutex_init, coop_condition_variable_init
7374 Don't use the functions coop_mutex_init and
7375 coop_condition_variable_init. They will change.
7377 Use scm_mutex_init and scm_cond_init instead.
7379 ** New function: int scm_cond_timedwait (scm_cond_t *COND, scm_mutex_t *MUTEX, const struct timespec *ABSTIME)
7380 `scm_cond_timedwait' atomically unlocks MUTEX and waits on
7381 COND, as `scm_cond_wait' does, but it also bounds the duration
7382 of the wait. If COND has not been signaled before time ABSTIME,
7383 the mutex MUTEX is re-acquired and `scm_cond_timedwait'
7384 returns the error code `ETIMEDOUT'.
7386 The ABSTIME parameter specifies an absolute time, with the same
7387 origin as `time' and `gettimeofday': an ABSTIME of 0 corresponds
7388 to 00:00:00 GMT, January 1, 1970.
7390 ** New function: scm_cond_broadcast (scm_cond_t *COND)
7391 `scm_cond_broadcast' restarts all the threads that are waiting
7392 on the condition variable COND. Nothing happens if no threads are
7395 ** New function: scm_key_create (scm_key_t *KEY, void (*destr_function) (void *))
7396 `scm_key_create' allocates a new TSD key. The key is stored in
7397 the location pointed to by KEY. There is no limit on the number
7398 of keys allocated at a given time. The value initially associated
7399 with the returned key is `NULL' in all currently executing threads.
7401 The DESTR_FUNCTION argument, if not `NULL', specifies a destructor
7402 function associated with the key. When a thread terminates,
7403 DESTR_FUNCTION is called on the value associated with the key in
7404 that thread. The DESTR_FUNCTION is not called if a key is deleted
7405 with `scm_key_delete' or a value is changed with
7406 `scm_setspecific'. The order in which destructor functions are
7407 called at thread termination time is unspecified.
7409 Destructors are not yet implemented.
7411 ** New function: scm_setspecific (scm_key_t KEY, const void *POINTER)
7412 `scm_setspecific' changes the value associated with KEY in the
7413 calling thread, storing the given POINTER instead.
7415 ** New function: scm_getspecific (scm_key_t KEY)
7416 `scm_getspecific' returns the value currently associated with
7417 KEY in the calling thread.
7419 ** New function: scm_key_delete (scm_key_t KEY)
7420 `scm_key_delete' deallocates a TSD key. It does not check
7421 whether non-`NULL' values are associated with that key in the
7422 currently executing threads, nor call the destructor function
7423 associated with the key.
7425 ** New function: scm_c_hook_init (scm_c_hook_t *HOOK, void *HOOK_DATA, scm_c_hook_type_t TYPE)
7427 Initialize a C level hook HOOK with associated HOOK_DATA and type
7428 TYPE. (See scm_c_hook_run ().)
7430 ** New function: scm_c_hook_add (scm_c_hook_t *HOOK, scm_c_hook_function_t FUNC, void *FUNC_DATA, int APPENDP)
7432 Add hook function FUNC with associated FUNC_DATA to HOOK. If APPENDP
7433 is true, add it last, otherwise first. The same FUNC can be added
7434 multiple times if FUNC_DATA differ and vice versa.
7436 ** New function: scm_c_hook_remove (scm_c_hook_t *HOOK, scm_c_hook_function_t FUNC, void *FUNC_DATA)
7438 Remove hook function FUNC with associated FUNC_DATA from HOOK. A
7439 function is only removed if both FUNC and FUNC_DATA matches.
7441 ** New function: void *scm_c_hook_run (scm_c_hook_t *HOOK, void *DATA)
7443 Run hook HOOK passing DATA to the hook functions.
7445 If TYPE is SCM_C_HOOK_NORMAL, all hook functions are run. The value
7446 returned is undefined.
7448 If TYPE is SCM_C_HOOK_OR, hook functions are run until a function
7449 returns a non-NULL value. This value is returned as the result of
7450 scm_c_hook_run. If all functions return NULL, NULL is returned.
7452 If TYPE is SCM_C_HOOK_AND, hook functions are run until a function
7453 returns a NULL value, and NULL is returned. If all functions returns
7454 a non-NULL value, the last value is returned.
7456 ** New C level GC hooks
7458 Five new C level hooks has been added to the garbage collector.
7460 scm_before_gc_c_hook
7463 are run before locking and after unlocking the heap. The system is
7464 thus in a mode where evaluation can take place. (Except that
7465 scm_before_gc_c_hook must not allocate new cells.)
7467 scm_before_mark_c_hook
7468 scm_before_sweep_c_hook
7469 scm_after_sweep_c_hook
7471 are run when the heap is locked. These are intended for extension of
7472 the GC in a modular fashion. Examples are the weaks and guardians
7475 ** Way for application to customize GC parameters
7477 The application can set up other default values for the GC heap
7478 allocation parameters
7480 GUILE_INIT_HEAP_SIZE_1, GUILE_MIN_YIELD_1,
7481 GUILE_INIT_HEAP_SIZE_2, GUILE_MIN_YIELD_2,
7482 GUILE_MAX_SEGMENT_SIZE,
7486 scm_default_init_heap_size_1, scm_default_min_yield_1,
7487 scm_default_init_heap_size_2, scm_default_min_yield_2,
7488 scm_default_max_segment_size
7490 respectively before callong scm_boot_guile.
7492 (See entry "New environment variables ..." in section
7493 "Changes to the stand-alone interpreter" above.)
7495 ** scm_protect_object/scm_unprotect_object now nest
7497 This means that you can call scm_protect_object multiple times on an
7498 object and count on the object being protected until
7499 scm_unprotect_object has been call the same number of times.
7501 The functions also have better time complexity.
7503 Still, it is usually possible to structure the application in a way
7504 that you don't need to use these functions. For example, if you use a
7505 protected standard Guile list to keep track of live objects rather
7506 than some custom data type, objects will die a natural death when they
7507 are no longer needed.
7509 ** Deprecated type tags: scm_tc16_flo, scm_tc_flo, scm_tc_dblr, scm_tc_dblc
7511 Guile does not provide the float representation for inexact real numbers any
7512 more. Now, only doubles are used to represent inexact real numbers. Further,
7513 the tag names scm_tc_dblr and scm_tc_dblc have been changed to scm_tc16_real
7514 and scm_tc16_complex, respectively.
7516 ** Removed deprecated type scm_smobfuns
7518 ** Removed deprecated function scm_newsmob
7520 ** Warning: scm_make_smob_type_mfpe might become deprecated in a future release
7522 There is an ongoing discussion among the developers whether to
7523 deprecate `scm_make_smob_type_mfpe' or not. Please use the current
7524 standard interface (scm_make_smob_type, scm_set_smob_XXX) in new code
7525 until this issue has been settled.
7527 ** Removed deprecated type tag scm_tc16_kw
7529 ** Added type tag scm_tc16_keyword
7531 (This was introduced already in release 1.3.4 but was not documented
7534 ** gdb_print now prints "*** Guile not initialized ***" until Guile initialized
7536 * Changes to system call interfaces:
7538 ** The "select" procedure now tests port buffers for the ability to
7539 provide input or accept output. Previously only the underlying file
7540 descriptors were checked.
7542 ** New variable PIPE_BUF: the maximum number of bytes that can be
7543 atomically written to a pipe.
7545 ** If a facility is not available on the system when Guile is
7546 compiled, the corresponding primitive procedure will not be defined.
7547 Previously it would have been defined but would throw a system-error
7548 exception if called. Exception handlers which catch this case may
7549 need minor modification: an error will be thrown with key
7550 'unbound-variable instead of 'system-error. Alternatively it's
7551 now possible to use `defined?' to check whether the facility is
7554 ** Procedures which depend on the timezone should now give the correct
7555 result on systems which cache the TZ environment variable, even if TZ
7556 is changed without calling tzset.
7558 * Changes to the networking interfaces:
7560 ** New functions: htons, ntohs, htonl, ntohl: for converting short and
7561 long integers between network and host format. For now, it's not
7562 particularly convenient to do this kind of thing, but consider:
7564 (define write-network-long
7565 (lambda (value port)
7566 (let ((v (make-uniform-vector 1 1 0)))
7567 (uniform-vector-set! v 0 (htonl value))
7568 (uniform-vector-write v port))))
7570 (define read-network-long
7572 (let ((v (make-uniform-vector 1 1 0)))
7573 (uniform-vector-read! v port)
7574 (ntohl (uniform-vector-ref v 0)))))
7576 ** If inet-aton fails, it now throws an error with key 'misc-error
7577 instead of 'system-error, since errno is not relevant.
7579 ** Certain gethostbyname/gethostbyaddr failures now throw errors with
7580 specific keys instead of 'system-error. The latter is inappropriate
7581 since errno will not have been set. The keys are:
7582 'host-not-found, 'try-again, 'no-recovery and 'no-data.
7584 ** sethostent, setnetent, setprotoent, setservent: now take an
7585 optional argument STAYOPEN, which specifies whether the database
7586 remains open after a database entry is accessed randomly (e.g., using
7587 gethostbyname for the hosts database.) The default is #f. Previously
7591 Changes since Guile 1.3.2:
7593 * Changes to the stand-alone interpreter
7597 An initial version of the Guile debugger written by Chris Hanson has
7598 been added. The debugger is still under development but is included
7599 in the distribution anyway since it is already quite useful.
7605 after an error to enter the debugger. Type `help' inside the debugger
7606 for a description of available commands.
7608 If you prefer to have stack frames numbered and printed in
7609 anti-chronological order and prefer up in the stack to be down on the
7610 screen as is the case in gdb, you can put
7612 (debug-enable 'backwards)
7614 in your .guile startup file. (However, this means that Guile can't
7615 use indentation to indicate stack level.)
7617 The debugger is autoloaded into Guile at the first use.
7619 ** Further enhancements to backtraces
7621 There is a new debug option `width' which controls the maximum width
7622 on the screen of printed stack frames. Fancy printing parameters
7623 ("level" and "length" as in Common LISP) are adaptively adjusted for
7624 each stack frame to give maximum information while still fitting
7625 within the bounds. If the stack frame can't be made to fit by
7626 adjusting parameters, it is simply cut off at the end. This is marked
7629 ** Some modules are now only loaded when the repl is started
7631 The modules (ice-9 debug), (ice-9 session), (ice-9 threads) and (ice-9
7632 regex) are now loaded into (guile-user) only if the repl has been
7633 started. The effect is that the startup time for scripts has been
7634 reduced to 30% of what it was previously.
7636 Correctly written scripts load the modules they require at the top of
7637 the file and should not be affected by this change.
7639 ** Hooks are now represented as smobs
7641 * Changes to Scheme functions and syntax
7643 ** Readline support has changed again.
7645 The old (readline-activator) module is gone. Use (ice-9 readline)
7646 instead, which now contains all readline functionality. So the code
7647 to activate readline is now
7649 (use-modules (ice-9 readline))
7652 This should work at any time, including from the guile prompt.
7654 To avoid confusion about the terms of Guile's license, please only
7655 enable readline for your personal use; please don't make it the
7656 default for others. Here is why we make this rather odd-sounding
7659 Guile is normally licensed under a weakened form of the GNU General
7660 Public License, which allows you to link code with Guile without
7661 placing that code under the GPL. This exception is important to some
7664 However, since readline is distributed under the GNU General Public
7665 License, when you link Guile with readline, either statically or
7666 dynamically, you effectively change Guile's license to the strict GPL.
7667 Whenever you link any strictly GPL'd code into Guile, uses of Guile
7668 which are normally permitted become forbidden. This is a rather
7669 non-obvious consequence of the licensing terms.
7671 So, to make sure things remain clear, please let people choose for
7672 themselves whether to link GPL'd libraries like readline with Guile.
7674 ** regexp-substitute/global has changed slightly, but incompatibly.
7676 If you include a function in the item list, the string of the match
7677 object it receives is the same string passed to
7678 regexp-substitute/global, not some suffix of that string.
7679 Correspondingly, the match's positions are relative to the entire
7680 string, not the suffix.
7682 If the regexp can match the empty string, the way matches are chosen
7683 from the string has changed. regexp-substitute/global recognizes the
7684 same set of matches that list-matches does; see below.
7686 ** New function: list-matches REGEXP STRING [FLAGS]
7688 Return a list of match objects, one for every non-overlapping, maximal
7689 match of REGEXP in STRING. The matches appear in left-to-right order.
7690 list-matches only reports matches of the empty string if there are no
7691 other matches which begin on, end at, or include the empty match's
7694 If present, FLAGS is passed as the FLAGS argument to regexp-exec.
7696 ** New function: fold-matches REGEXP STRING INIT PROC [FLAGS]
7698 For each match of REGEXP in STRING, apply PROC to the match object,
7699 and the last value PROC returned, or INIT for the first call. Return
7700 the last value returned by PROC. We apply PROC to the matches as they
7701 appear from left to right.
7703 This function recognizes matches according to the same criteria as
7706 Thus, you could define list-matches like this:
7708 (define (list-matches regexp string . flags)
7709 (reverse! (apply fold-matches regexp string '() cons flags)))
7711 If present, FLAGS is passed as the FLAGS argument to regexp-exec.
7715 *** New function: hook? OBJ
7717 Return #t if OBJ is a hook, otherwise #f.
7719 *** New function: make-hook-with-name NAME [ARITY]
7721 Return a hook with name NAME and arity ARITY. The default value for
7722 ARITY is 0. The only effect of NAME is that it will appear when the
7723 hook object is printed to ease debugging.
7725 *** New function: hook-empty? HOOK
7727 Return #t if HOOK doesn't contain any procedures, otherwise #f.
7729 *** New function: hook->list HOOK
7731 Return a list of the procedures that are called when run-hook is
7734 ** `map' signals an error if its argument lists are not all the same length.
7736 This is the behavior required by R5RS, so this change is really a bug
7737 fix. But it seems to affect a lot of people's code, so we're
7738 mentioning it here anyway.
7740 ** Print-state handling has been made more transparent
7742 Under certain circumstances, ports are represented as a port with an
7743 associated print state. Earlier, this pair was represented as a pair
7744 (see "Some magic has been added to the printer" below). It is now
7745 indistinguishable (almost; see `get-print-state') from a port on the
7748 *** New function: port-with-print-state OUTPUT-PORT PRINT-STATE
7750 Return a new port with the associated print state PRINT-STATE.
7752 *** New function: get-print-state OUTPUT-PORT
7754 Return the print state associated with this port if it exists,
7755 otherwise return #f.
7757 *** New function: directory-stream? OBJECT
7759 Returns true iff OBJECT is a directory stream --- the sort of object
7760 returned by `opendir'.
7762 ** New function: using-readline?
7764 Return #t if readline is in use in the current repl.
7766 ** structs will be removed in 1.4
7768 Structs will be replaced in Guile 1.4. We will merge GOOPS into Guile
7769 and use GOOPS objects as the fundamental record type.
7771 * Changes to the scm_ interface
7773 ** structs will be removed in 1.4
7775 The entire current struct interface (struct.c, struct.h) will be
7776 replaced in Guile 1.4. We will merge GOOPS into libguile and use
7777 GOOPS objects as the fundamental record type.
7779 ** The internal representation of subr's has changed
7781 Instead of giving a hint to the subr name, the CAR field of the subr
7782 now contains an index to a subr entry in scm_subr_table.
7784 *** New variable: scm_subr_table
7786 An array of subr entries. A subr entry contains the name, properties
7787 and documentation associated with the subr. The properties and
7788 documentation slots are not yet used.
7790 ** A new scheme for "forwarding" calls to a builtin to a generic function
7792 It is now possible to extend the functionality of some Guile
7793 primitives by letting them defer a call to a GOOPS generic function on
7794 argument mismatch. This means that there is no loss of efficiency in
7799 (use-modules (oop goops)) ; Must be GOOPS version 0.2.
7800 (define-method + ((x <string>) (y <string>))
7801 (string-append x y))
7803 + will still be as efficient as usual in numerical calculations, but
7804 can also be used for concatenating strings.
7806 Who will be the first one to extend Guile's numerical tower to
7807 rationals? :) [OK, there a few other things to fix before this can
7808 be made in a clean way.]
7810 *** New snarf macros for defining primitives: SCM_GPROC, SCM_GPROC1
7812 New macro: SCM_GPROC (CNAME, SNAME, REQ, OPT, VAR, CFUNC, GENERIC)
7814 New macro: SCM_GPROC1 (CNAME, SNAME, TYPE, CFUNC, GENERIC)
7816 These do the same job as SCM_PROC and SCM_PROC1, but they also define
7817 a variable GENERIC which can be used by the dispatch macros below.
7819 [This is experimental code which may change soon.]
7821 *** New macros for forwarding control to a generic on arg type error
7823 New macro: SCM_WTA_DISPATCH_1 (GENERIC, ARG1, POS, SUBR)
7825 New macro: SCM_WTA_DISPATCH_2 (GENERIC, ARG1, ARG2, POS, SUBR)
7827 These correspond to the scm_wta function call, and have the same
7828 behaviour until the user has called the GOOPS primitive
7829 `enable-primitive-generic!'. After that, these macros will apply the
7830 generic function GENERIC to the argument(s) instead of calling
7833 [This is experimental code which may change soon.]
7835 *** New macros for argument testing with generic dispatch
7837 New macro: SCM_GASSERT1 (COND, GENERIC, ARG1, POS, SUBR)
7839 New macro: SCM_GASSERT2 (COND, GENERIC, ARG1, ARG2, POS, SUBR)
7841 These correspond to the SCM_ASSERT macro, but will defer control to
7842 GENERIC on error after `enable-primitive-generic!' has been called.
7844 [This is experimental code which may change soon.]
7846 ** New function: SCM scm_eval_body (SCM body, SCM env)
7848 Evaluates the body of a special form.
7850 ** The internal representation of struct's has changed
7852 Previously, four slots were allocated for the procedure(s) of entities
7853 and operators. The motivation for this representation had to do with
7854 the structure of the evaluator, the wish to support tail-recursive
7855 generic functions, and efficiency. Since the generic function
7856 dispatch mechanism has changed, there is no longer a need for such an
7857 expensive representation, and the representation has been simplified.
7859 This should not make any difference for most users.
7861 ** GOOPS support has been cleaned up.
7863 Some code has been moved from eval.c to objects.c and code in both of
7864 these compilation units has been cleaned up and better structured.
7866 *** New functions for applying generic functions
7868 New function: SCM scm_apply_generic (GENERIC, ARGS)
7869 New function: SCM scm_call_generic_0 (GENERIC)
7870 New function: SCM scm_call_generic_1 (GENERIC, ARG1)
7871 New function: SCM scm_call_generic_2 (GENERIC, ARG1, ARG2)
7872 New function: SCM scm_call_generic_3 (GENERIC, ARG1, ARG2, ARG3)
7874 ** Deprecated function: scm_make_named_hook
7876 It is now replaced by:
7878 ** New function: SCM scm_create_hook (const char *name, int arity)
7880 Creates a hook in the same way as make-hook above but also
7881 binds a variable named NAME to it.
7883 This is the typical way of creating a hook from C code.
7885 Currently, the variable is created in the "current" module.
7886 This might change when we get the new module system.
7888 [The behaviour is identical to scm_make_named_hook.]
7892 Changes since Guile 1.3:
7894 * Changes to mailing lists
7896 ** Some of the Guile mailing lists have moved to sourceware.cygnus.com.
7898 See the README file to find current addresses for all the Guile
7901 * Changes to the distribution
7903 ** Readline support is no longer included with Guile by default.
7905 Based on the different license terms of Guile and Readline, we
7906 concluded that Guile should not *by default* cause the linking of
7907 Readline into an application program. Readline support is now offered
7908 as a separate module, which is linked into an application only when
7909 you explicitly specify it.
7911 Although Guile is GNU software, its distribution terms add a special
7912 exception to the usual GNU General Public License (GPL). Guile's
7913 license includes a clause that allows you to link Guile with non-free
7914 programs. We add this exception so as not to put Guile at a
7915 disadvantage vis-a-vis other extensibility packages that support other
7918 In contrast, the GNU Readline library is distributed under the GNU
7919 General Public License pure and simple. This means that you may not
7920 link Readline, even dynamically, into an application unless it is
7921 distributed under a free software license that is compatible the GPL.
7923 Because of this difference in distribution terms, an application that
7924 can use Guile may not be able to use Readline. Now users will be
7925 explicitly offered two independent decisions about the use of these
7928 You can activate the readline support by issuing
7930 (use-modules (readline-activator))
7933 from your ".guile" file, for example.
7935 * Changes to the stand-alone interpreter
7937 ** All builtins now print as primitives.
7938 Previously builtin procedures not belonging to the fundamental subr
7939 types printed as #<compiled closure #<primitive-procedure gsubr-apply>>.
7940 Now, they print as #<primitive-procedure NAME>.
7942 ** Backtraces slightly more intelligible.
7943 gsubr-apply and macro transformer application frames no longer appear
7946 * Changes to Scheme functions and syntax
7948 ** Guile now correctly handles internal defines by rewriting them into
7949 their equivalent letrec. Previously, internal defines would
7950 incrementally add to the innermost environment, without checking
7951 whether the restrictions specified in RnRS were met. This lead to the
7952 correct behaviour when these restriction actually were met, but didn't
7953 catch all illegal uses. Such an illegal use could lead to crashes of
7954 the Guile interpreter or other unwanted results. An example of
7955 incorrect internal defines that made Guile behave erratically:
7967 The problem with this example is that the definition of `c' uses the
7968 value of `b' directly. This confuses the meoization machine of Guile
7969 so that the second call of `b' (this time in a larger environment that
7970 also contains bindings for `c' and `d') refers to the binding of `c'
7971 instead of `a'. You could also make Guile crash with a variation on
7976 (define (b flag) (if flag a 1))
7977 (define c (1+ (b flag)))
7985 From now on, Guile will issue an `Unbound variable: b' error message
7990 A hook contains a list of functions which should be called on
7991 particular occasions in an existing program. Hooks are used for
7994 A window manager might have a hook before-window-map-hook. The window
7995 manager uses the function run-hooks to call all functions stored in
7996 before-window-map-hook each time a window is mapped. The user can
7997 store functions in the hook using add-hook!.
7999 In Guile, hooks are first class objects.
8001 *** New function: make-hook [N_ARGS]
8003 Return a hook for hook functions which can take N_ARGS arguments.
8004 The default value for N_ARGS is 0.
8006 (See also scm_make_named_hook below.)
8008 *** New function: add-hook! HOOK PROC [APPEND_P]
8010 Put PROC at the beginning of the list of functions stored in HOOK.
8011 If APPEND_P is supplied, and non-false, put PROC at the end instead.
8013 PROC must be able to take the number of arguments specified when the
8016 If PROC already exists in HOOK, then remove it first.
8018 *** New function: remove-hook! HOOK PROC
8020 Remove PROC from the list of functions in HOOK.
8022 *** New function: reset-hook! HOOK
8024 Clear the list of hook functions stored in HOOK.
8026 *** New function: run-hook HOOK ARG1 ...
8028 Run all hook functions stored in HOOK with arguments ARG1 ... .
8029 The number of arguments supplied must correspond to the number given
8030 when the hook was created.
8032 ** The function `dynamic-link' now takes optional keyword arguments.
8033 The only keyword argument that is currently defined is `:global
8034 BOOL'. With it, you can control whether the shared library will be
8035 linked in global mode or not. In global mode, the symbols from the
8036 linked library can be used to resolve references from other
8037 dynamically linked libraries. In non-global mode, the linked
8038 library is essentially invisible and can only be accessed via
8039 `dynamic-func', etc. The default is now to link in global mode.
8040 Previously, the default has been non-global mode.
8042 The `#:global' keyword is only effective on platforms that support
8043 the dlopen family of functions.
8045 ** New function `provided?'
8047 - Function: provided? FEATURE
8048 Return true iff FEATURE is supported by this installation of
8049 Guile. FEATURE must be a symbol naming a feature; the global
8050 variable `*features*' is a list of available features.
8052 ** Changes to the module (ice-9 expect):
8054 *** The expect-strings macro now matches `$' in a regular expression
8055 only at a line-break or end-of-file by default. Previously it would
8056 match the end of the string accumulated so far. The old behaviour
8057 can be obtained by setting the variable `expect-strings-exec-flags'
8060 *** The expect-strings macro now uses a variable `expect-strings-exec-flags'
8061 for the regexp-exec flags. If `regexp/noteol' is included, then `$'
8062 in a regular expression will still match before a line-break or
8063 end-of-file. The default is `regexp/noteol'.
8065 *** The expect-strings macro now uses a variable
8066 `expect-strings-compile-flags' for the flags to be supplied to
8067 `make-regexp'. The default is `regexp/newline', which was previously
8070 *** The expect macro now supplies two arguments to a match procedure:
8071 the current accumulated string and a flag to indicate whether
8072 end-of-file has been reached. Previously only the string was supplied.
8073 If end-of-file is reached, the match procedure will be called an
8074 additional time with the same accumulated string as the previous call
8075 but with the flag set.
8077 ** New module (ice-9 format), implementing the Common Lisp `format' function.
8079 This code, and the documentation for it that appears here, was
8080 borrowed from SLIB, with minor adaptations for Guile.
8082 - Function: format DESTINATION FORMAT-STRING . ARGUMENTS
8083 An almost complete implementation of Common LISP format description
8084 according to the CL reference book `Common LISP' from Guy L.
8085 Steele, Digital Press. Backward compatible to most of the
8086 available Scheme format implementations.
8088 Returns `#t', `#f' or a string; has side effect of printing
8089 according to FORMAT-STRING. If DESTINATION is `#t', the output is
8090 to the current output port and `#t' is returned. If DESTINATION
8091 is `#f', a formatted string is returned as the result of the call.
8092 NEW: If DESTINATION is a string, DESTINATION is regarded as the
8093 format string; FORMAT-STRING is then the first argument and the
8094 output is returned as a string. If DESTINATION is a number, the
8095 output is to the current error port if available by the
8096 implementation. Otherwise DESTINATION must be an output port and
8099 FORMAT-STRING must be a string. In case of a formatting error
8100 format returns `#f' and prints a message on the current output or
8101 error port. Characters are output as if the string were output by
8102 the `display' function with the exception of those prefixed by a
8103 tilde (~). For a detailed description of the FORMAT-STRING syntax
8104 please consult a Common LISP format reference manual. For a test
8105 suite to verify this format implementation load `formatst.scm'.
8106 Please send bug reports to `lutzeb@cs.tu-berlin.de'.
8108 Note: `format' is not reentrant, i.e. only one `format'-call may
8109 be executed at a time.
8112 *** Format Specification (Format version 3.0)
8114 Please consult a Common LISP format reference manual for a detailed
8115 description of the format string syntax. For a demonstration of the
8116 implemented directives see `formatst.scm'.
8118 This implementation supports directive parameters and modifiers (`:'
8119 and `@' characters). Multiple parameters must be separated by a comma
8120 (`,'). Parameters can be numerical parameters (positive or negative),
8121 character parameters (prefixed by a quote character (`''), variable
8122 parameters (`v'), number of rest arguments parameter (`#'), empty and
8123 default parameters. Directive characters are case independent. The
8124 general form of a directive is:
8126 DIRECTIVE ::= ~{DIRECTIVE-PARAMETER,}[:][@]DIRECTIVE-CHARACTER
8128 DIRECTIVE-PARAMETER ::= [ [-|+]{0-9}+ | 'CHARACTER | v | # ]
8130 *** Implemented CL Format Control Directives
8132 Documentation syntax: Uppercase characters represent the
8133 corresponding control directive characters. Lowercase characters
8134 represent control directive parameter descriptions.
8137 Any (print as `display' does).
8141 `~MINCOL,COLINC,MINPAD,PADCHARA'
8145 S-expression (print as `write' does).
8149 `~MINCOL,COLINC,MINPAD,PADCHARS'
8155 print number sign always.
8158 print comma separated.
8160 `~MINCOL,PADCHAR,COMMACHARD'
8166 print number sign always.
8169 print comma separated.
8171 `~MINCOL,PADCHAR,COMMACHARX'
8177 print number sign always.
8180 print comma separated.
8182 `~MINCOL,PADCHAR,COMMACHARO'
8188 print number sign always.
8191 print comma separated.
8193 `~MINCOL,PADCHAR,COMMACHARB'
8198 `~N,MINCOL,PADCHAR,COMMACHARR'
8202 print a number as a Roman numeral.
8205 print a number as an "old fashioned" Roman numeral.
8208 print a number as an ordinal English number.
8211 print a number as a cardinal English number.
8216 prints `y' and `ies'.
8219 as `~P but jumps 1 argument backward.'
8222 as `~@P but jumps 1 argument backward.'
8227 prints a character as the reader can understand it (i.e. `#\'
8231 prints a character as emacs does (eg. `^C' for ASCII 03).
8234 Fixed-format floating-point (prints a flonum like MMM.NNN).
8235 `~WIDTH,DIGITS,SCALE,OVERFLOWCHAR,PADCHARF'
8237 If the number is positive a plus sign is printed.
8240 Exponential floating-point (prints a flonum like MMM.NNN`E'EE).
8241 `~WIDTH,DIGITS,EXPONENTDIGITS,SCALE,OVERFLOWCHAR,PADCHAR,EXPONENTCHARE'
8243 If the number is positive a plus sign is printed.
8246 General floating-point (prints a flonum either fixed or
8248 `~WIDTH,DIGITS,EXPONENTDIGITS,SCALE,OVERFLOWCHAR,PADCHAR,EXPONENTCHARG'
8250 If the number is positive a plus sign is printed.
8253 Dollars floating-point (prints a flonum in fixed with signs
8255 `~DIGITS,SCALE,WIDTH,PADCHAR$'
8257 If the number is positive a plus sign is printed.
8260 A sign is always printed and appears before the padding.
8263 The sign appears before the padding.
8271 print newline if not at the beginning of the output line.
8273 prints `~&' and then N-1 newlines.
8278 print N page separators.
8288 newline is ignored, white space left.
8291 newline is left, white space ignored.
8296 relative tabulation.
8302 Indirection (expects indirect arguments as a list).
8304 extracts indirect arguments from format arguments.
8307 Case conversion (converts by `string-downcase').
8309 converts by `string-capitalize'.
8312 converts by `string-capitalize-first'.
8315 converts by `string-upcase'.
8318 Argument Jumping (jumps 1 argument forward).
8320 jumps N arguments forward.
8323 jumps 1 argument backward.
8326 jumps N arguments backward.
8329 jumps to the 0th argument.
8332 jumps to the Nth argument (beginning from 0)
8334 `~[STR0~;STR1~;...~;STRN~]'
8335 Conditional Expression (numerical clause conditional).
8337 take argument from N.
8340 true test conditional.
8343 if-else-then conditional.
8349 default clause follows.
8352 Iteration (args come from the next argument (a list)).
8354 at most N iterations.
8357 args from next arg (a list of lists).
8360 args from the rest of arguments.
8363 args from the rest args (lists).
8374 aborts if N <= M <= K
8376 *** Not Implemented CL Format Control Directives
8379 print `#f' as an empty list (see below).
8382 print `#f' as an empty list (see below).
8388 (sorry I don't understand its semantics completely)
8390 *** Extended, Replaced and Additional Control Directives
8392 `~MINCOL,PADCHAR,COMMACHAR,COMMAWIDTHD'
8393 `~MINCOL,PADCHAR,COMMACHAR,COMMAWIDTHX'
8394 `~MINCOL,PADCHAR,COMMACHAR,COMMAWIDTHO'
8395 `~MINCOL,PADCHAR,COMMACHAR,COMMAWIDTHB'
8396 `~N,MINCOL,PADCHAR,COMMACHAR,COMMAWIDTHR'
8397 COMMAWIDTH is the number of characters between two comma
8401 print a R4RS complex number as `~F~@Fi' with passed parameters for
8405 Pretty print formatting of an argument for scheme code lists.
8411 Flushes the output if format DESTINATION is a port.
8414 Print a `#\space' character
8416 print N `#\space' characters.
8419 Print a `#\tab' character
8421 print N `#\tab' characters.
8424 Takes N as an integer representation for a character. No arguments
8425 are consumed. N is converted to a character by `integer->char'. N
8426 must be a positive decimal number.
8429 Print out readproof. Prints out internal objects represented as
8430 `#<...>' as strings `"#<...>"' so that the format output can always
8431 be processed by `read'.
8434 Print out readproof. Prints out internal objects represented as
8435 `#<...>' as strings `"#<...>"' so that the format output can always
8436 be processed by `read'.
8439 Prints information and a copyright notice on the format
8442 prints format version.
8445 may also print number strings, i.e. passing a number as a string
8446 and format it accordingly.
8448 *** Configuration Variables
8450 The format module exports some configuration variables to suit the
8451 systems and users needs. There should be no modification necessary for
8452 the configuration that comes with Guile. Format detects automatically
8453 if the running scheme system implements floating point numbers and
8456 format:symbol-case-conv
8457 Symbols are converted by `symbol->string' so the case type of the
8458 printed symbols is implementation dependent.
8459 `format:symbol-case-conv' is a one arg closure which is either
8460 `#f' (no conversion), `string-upcase', `string-downcase' or
8461 `string-capitalize'. (default `#f')
8463 format:iobj-case-conv
8464 As FORMAT:SYMBOL-CASE-CONV but applies for the representation of
8465 implementation internal objects. (default `#f')
8468 The character prefixing the exponent value in `~E' printing.
8471 *** Compatibility With Other Format Implementations
8477 Downward compatible except for padding support and `~A', `~S',
8478 `~P', `~X' uppercase printing. SLIB format 1.4 uses C-style
8479 `printf' padding support which is completely replaced by the CL
8480 `format' padding style.
8483 Downward compatible except for `~', which is not documented
8484 (ignores all characters inside the format string up to a newline
8485 character). (7.1 implements `~a', `~s', ~NEWLINE, `~~', `~%',
8486 numerical and variable parameters and `:/@' modifiers in the CL
8490 Downward compatible except for `~A' and `~S' which print in
8491 uppercase. (Elk implements `~a', `~s', `~~', and `~%' (no
8492 directive parameters or modifiers)).
8495 Downward compatible except for an optional destination parameter:
8496 S2C accepts a format call without a destination which returns a
8497 formatted string. This is equivalent to a #f destination in S2C.
8498 (S2C implements `~a', `~s', `~c', `~%', and `~~' (no directive
8499 parameters or modifiers)).
8502 ** Changes to string-handling functions.
8504 These functions were added to support the (ice-9 format) module, above.
8506 *** New function: string-upcase STRING
8507 *** New function: string-downcase STRING
8509 These are non-destructive versions of the existing string-upcase! and
8510 string-downcase! functions.
8512 *** New function: string-capitalize! STRING
8513 *** New function: string-capitalize STRING
8515 These functions convert the first letter of each word in the string to
8518 (string-capitalize "howdy there")
8521 As with the other functions, string-capitalize! modifies the string in
8522 place, while string-capitalize returns a modified copy of its argument.
8524 *** New function: string-ci->symbol STRING
8526 Return a symbol whose name is STRING, but having the same case as if
8527 the symbol had be read by `read'.
8529 Guile can be configured to be sensitive or insensitive to case
8530 differences in Scheme identifiers. If Guile is case-insensitive, all
8531 symbols are converted to lower case on input. The `string-ci->symbol'
8532 function returns a symbol whose name in STRING, transformed as Guile
8533 would if STRING were input.
8535 *** New function: substring-move! STRING1 START END STRING2 START
8537 Copy the substring of STRING1 from START (inclusive) to END
8538 (exclusive) to STRING2 at START. STRING1 and STRING2 may be the same
8539 string, and the source and destination areas may overlap; in all
8540 cases, the function behaves as if all the characters were copied
8543 *** Extended functions: substring-move-left! substring-move-right!
8545 These functions now correctly copy arbitrarily overlapping substrings;
8546 they are both synonyms for substring-move!.
8549 ** New module (ice-9 getopt-long), with the function `getopt-long'.
8551 getopt-long is a function for parsing command-line arguments in a
8552 manner consistent with other GNU programs.
8554 (getopt-long ARGS GRAMMAR)
8555 Parse the arguments ARGS according to the argument list grammar GRAMMAR.
8557 ARGS should be a list of strings. Its first element should be the
8558 name of the program; subsequent elements should be the arguments
8559 that were passed to the program on the command line. The
8560 `program-arguments' procedure returns a list of this form.
8562 GRAMMAR is a list of the form:
8563 ((OPTION (PROPERTY VALUE) ...) ...)
8565 Each OPTION should be a symbol. `getopt-long' will accept a
8566 command-line option named `--OPTION'.
8567 Each option can have the following (PROPERTY VALUE) pairs:
8569 (single-char CHAR) --- Accept `-CHAR' as a single-character
8570 equivalent to `--OPTION'. This is how to specify traditional
8572 (required? BOOL) --- If BOOL is true, the option is required.
8573 getopt-long will raise an error if it is not found in ARGS.
8574 (value BOOL) --- If BOOL is #t, the option accepts a value; if
8575 it is #f, it does not; and if it is the symbol
8576 `optional', the option may appear in ARGS with or
8578 (predicate FUNC) --- If the option accepts a value (i.e. you
8579 specified `(value #t)' for this option), then getopt
8580 will apply FUNC to the value, and throw an exception
8581 if it returns #f. FUNC should be a procedure which
8582 accepts a string and returns a boolean value; you may
8583 need to use quasiquotes to get it into GRAMMAR.
8585 The (PROPERTY VALUE) pairs may occur in any order, but each
8586 property may occur only once. By default, options do not have
8587 single-character equivalents, are not required, and do not take
8590 In ARGS, single-character options may be combined, in the usual
8591 Unix fashion: ("-x" "-y") is equivalent to ("-xy"). If an option
8592 accepts values, then it must be the last option in the
8593 combination; the value is the next argument. So, for example, using
8594 the following grammar:
8595 ((apples (single-char #\a))
8596 (blimps (single-char #\b) (value #t))
8597 (catalexis (single-char #\c) (value #t)))
8598 the following argument lists would be acceptable:
8599 ("-a" "-b" "bang" "-c" "couth") ("bang" and "couth" are the values
8600 for "blimps" and "catalexis")
8601 ("-ab" "bang" "-c" "couth") (same)
8602 ("-ac" "couth" "-b" "bang") (same)
8603 ("-abc" "couth" "bang") (an error, since `-b' is not the
8604 last option in its combination)
8606 If an option's value is optional, then `getopt-long' decides
8607 whether it has a value by looking at what follows it in ARGS. If
8608 the next element is a string, and it does not appear to be an
8609 option itself, then that string is the option's value.
8611 The value of a long option can appear as the next element in ARGS,
8612 or it can follow the option name, separated by an `=' character.
8613 Thus, using the same grammar as above, the following argument lists
8615 ("--apples" "Braeburn" "--blimps" "Goodyear")
8616 ("--apples=Braeburn" "--blimps" "Goodyear")
8617 ("--blimps" "Goodyear" "--apples=Braeburn")
8619 If the option "--" appears in ARGS, argument parsing stops there;
8620 subsequent arguments are returned as ordinary arguments, even if
8621 they resemble options. So, in the argument list:
8622 ("--apples" "Granny Smith" "--" "--blimp" "Goodyear")
8623 `getopt-long' will recognize the `apples' option as having the
8624 value "Granny Smith", but it will not recognize the `blimp'
8625 option; it will return the strings "--blimp" and "Goodyear" as
8626 ordinary argument strings.
8628 The `getopt-long' function returns the parsed argument list as an
8629 assocation list, mapping option names --- the symbols from GRAMMAR
8630 --- onto their values, or #t if the option does not accept a value.
8631 Unused options do not appear in the alist.
8633 All arguments that are not the value of any option are returned
8634 as a list, associated with the empty list.
8636 `getopt-long' throws an exception if:
8637 - it finds an unrecognized option in ARGS
8638 - a required option is omitted
8639 - an option that requires an argument doesn't get one
8640 - an option that doesn't accept an argument does get one (this can
8641 only happen using the long option `--opt=value' syntax)
8642 - an option predicate fails
8647 `((lockfile-dir (required? #t)
8650 (predicate ,file-is-directory?))
8651 (verbose (required? #f)
8654 (x-includes (single-char #\x))
8655 (rnet-server (single-char #\y)
8656 (predicate ,string?))))
8658 (getopt-long '("my-prog" "-vk" "/tmp" "foo1" "--x-includes=/usr/include"
8659 "--rnet-server=lamprod" "--" "-fred" "foo2" "foo3")
8661 => ((() "foo1" "-fred" "foo2" "foo3")
8662 (rnet-server . "lamprod")
8663 (x-includes . "/usr/include")
8664 (lockfile-dir . "/tmp")
8667 ** The (ice-9 getopt-gnu-style) module is obsolete; use (ice-9 getopt-long).
8669 It will be removed in a few releases.
8671 ** New syntax: lambda*
8672 ** New syntax: define*
8673 ** New syntax: define*-public
8674 ** New syntax: defmacro*
8675 ** New syntax: defmacro*-public
8676 Guile now supports optional arguments.
8678 `lambda*', `define*', `define*-public', `defmacro*' and
8679 `defmacro*-public' are identical to the non-* versions except that
8680 they use an extended type of parameter list that has the following BNF
8681 syntax (parentheses are literal, square brackets indicate grouping,
8682 and `*', `+' and `?' have the usual meaning):
8684 ext-param-list ::= ( [identifier]* [#&optional [ext-var-decl]+]?
8685 [#&key [ext-var-decl]+ [#&allow-other-keys]?]?
8686 [[#&rest identifier]|[. identifier]]? ) | [identifier]
8688 ext-var-decl ::= identifier | ( identifier expression )
8690 The semantics are best illustrated with the following documentation
8691 and examples for `lambda*':
8694 lambda extended for optional and keyword arguments
8696 lambda* creates a procedure that takes optional arguments. These
8697 are specified by putting them inside brackets at the end of the
8698 paramater list, but before any dotted rest argument. For example,
8699 (lambda* (a b #&optional c d . e) '())
8700 creates a procedure with fixed arguments a and b, optional arguments c
8701 and d, and rest argument e. If the optional arguments are omitted
8702 in a call, the variables for them are unbound in the procedure. This
8703 can be checked with the bound? macro.
8705 lambda* can also take keyword arguments. For example, a procedure
8707 (lambda* (#&key xyzzy larch) '())
8708 can be called with any of the argument lists (#:xyzzy 11)
8709 (#:larch 13) (#:larch 42 #:xyzzy 19) (). Whichever arguments
8710 are given as keywords are bound to values.
8712 Optional and keyword arguments can also be given default values
8713 which they take on when they are not present in a call, by giving a
8714 two-item list in place of an optional argument, for example in:
8715 (lambda* (foo #&optional (bar 42) #&key (baz 73)) (list foo bar baz))
8716 foo is a fixed argument, bar is an optional argument with default
8717 value 42, and baz is a keyword argument with default value 73.
8718 Default value expressions are not evaluated unless they are needed
8719 and until the procedure is called.
8721 lambda* now supports two more special parameter list keywords.
8723 lambda*-defined procedures now throw an error by default if a
8724 keyword other than one of those specified is found in the actual
8725 passed arguments. However, specifying #&allow-other-keys
8726 immediately after the kyword argument declarations restores the
8727 previous behavior of ignoring unknown keywords. lambda* also now
8728 guarantees that if the same keyword is passed more than once, the
8729 last one passed is the one that takes effect. For example,
8730 ((lambda* (#&key (heads 0) (tails 0)) (display (list heads tails)))
8731 #:heads 37 #:tails 42 #:heads 99)
8732 would result in (99 47) being displayed.
8734 #&rest is also now provided as a synonym for the dotted syntax rest
8735 argument. The argument lists (a . b) and (a #&rest b) are equivalent in
8736 all respects to lambda*. This is provided for more similarity to DSSSL,
8737 MIT-Scheme and Kawa among others, as well as for refugees from other
8740 Further documentation may be found in the optargs.scm file itself.
8742 The optional argument module also exports the macros `let-optional',
8743 `let-optional*', `let-keywords', `let-keywords*' and `bound?'. These
8744 are not documented here because they may be removed in the future, but
8745 full documentation is still available in optargs.scm.
8747 ** New syntax: and-let*
8748 Guile now supports the `and-let*' form, described in the draft SRFI-2.
8750 Syntax: (land* (<clause> ...) <body> ...)
8751 Each <clause> should have one of the following forms:
8752 (<variable> <expression>)
8755 Each <variable> or <bound-variable> should be an identifier. Each
8756 <expression> should be a valid expression. The <body> should be a
8757 possibly empty sequence of expressions, like the <body> of a
8760 Semantics: A LAND* expression is evaluated by evaluating the
8761 <expression> or <bound-variable> of each of the <clause>s from
8762 left to right. The value of the first <expression> or
8763 <bound-variable> that evaluates to a false value is returned; the
8764 remaining <expression>s and <bound-variable>s are not evaluated.
8765 The <body> forms are evaluated iff all the <expression>s and
8766 <bound-variable>s evaluate to true values.
8768 The <expression>s and the <body> are evaluated in an environment
8769 binding each <variable> of the preceding (<variable> <expression>)
8770 clauses to the value of the <expression>. Later bindings
8771 shadow earlier bindings.
8773 Guile's and-let* macro was contributed by Michael Livshin.
8775 ** New sorting functions
8777 *** New function: sorted? SEQUENCE LESS?
8778 Returns `#t' when the sequence argument is in non-decreasing order
8779 according to LESS? (that is, there is no adjacent pair `... x y
8780 ...' for which `(less? y x)').
8782 Returns `#f' when the sequence contains at least one out-of-order
8783 pair. It is an error if the sequence is neither a list nor a
8786 *** New function: merge LIST1 LIST2 LESS?
8787 LIST1 and LIST2 are sorted lists.
8788 Returns the sorted list of all elements in LIST1 and LIST2.
8790 Assume that the elements a and b1 in LIST1 and b2 in LIST2 are "equal"
8791 in the sense that (LESS? x y) --> #f for x, y in {a, b1, b2},
8792 and that a < b1 in LIST1. Then a < b1 < b2 in the result.
8793 (Here "<" should read "comes before".)
8795 *** New procedure: merge! LIST1 LIST2 LESS?
8796 Merges two lists, re-using the pairs of LIST1 and LIST2 to build
8797 the result. If the code is compiled, and LESS? constructs no new
8798 pairs, no pairs at all will be allocated. The first pair of the
8799 result will be either the first pair of LIST1 or the first pair of
8802 *** New function: sort SEQUENCE LESS?
8803 Accepts either a list or a vector, and returns a new sequence
8804 which is sorted. The new sequence is the same type as the input.
8805 Always `(sorted? (sort sequence less?) less?)'. The original
8806 sequence is not altered in any way. The new sequence shares its
8807 elements with the old one; no elements are copied.
8809 *** New procedure: sort! SEQUENCE LESS
8810 Returns its sorted result in the original boxes. No new storage is
8811 allocated at all. Proper usage: (set! slist (sort! slist <))
8813 *** New function: stable-sort SEQUENCE LESS?
8814 Similar to `sort' but stable. That is, if "equal" elements are
8815 ordered a < b in the original sequence, they will have the same order
8818 *** New function: stable-sort! SEQUENCE LESS?
8819 Similar to `sort!' but stable.
8820 Uses temporary storage when sorting vectors.
8822 *** New functions: sort-list, sort-list!
8823 Added for compatibility with scsh.
8825 ** New built-in random number support
8827 *** New function: random N [STATE]
8828 Accepts a positive integer or real N and returns a number of the
8829 same type between zero (inclusive) and N (exclusive). The values
8830 returned have a uniform distribution.
8832 The optional argument STATE must be of the type produced by
8833 `copy-random-state' or `seed->random-state'. It defaults to the value
8834 of the variable `*random-state*'. This object is used to maintain the
8835 state of the pseudo-random-number generator and is altered as a side
8836 effect of the `random' operation.
8838 *** New variable: *random-state*
8839 Holds a data structure that encodes the internal state of the
8840 random-number generator that `random' uses by default. The nature
8841 of this data structure is implementation-dependent. It may be
8842 printed out and successfully read back in, but may or may not
8843 function correctly as a random-number state object in another
8846 *** New function: copy-random-state [STATE]
8847 Returns a new object of type suitable for use as the value of the
8848 variable `*random-state*' and as a second argument to `random'.
8849 If argument STATE is given, a copy of it is returned. Otherwise a
8850 copy of `*random-state*' is returned.
8852 *** New function: seed->random-state SEED
8853 Returns a new object of type suitable for use as the value of the
8854 variable `*random-state*' and as a second argument to `random'.
8855 SEED is a string or a number. A new state is generated and
8856 initialized using SEED.
8858 *** New function: random:uniform [STATE]
8859 Returns an uniformly distributed inexact real random number in the
8860 range between 0 and 1.
8862 *** New procedure: random:solid-sphere! VECT [STATE]
8863 Fills VECT with inexact real random numbers the sum of whose
8864 squares is less than 1.0. Thinking of VECT as coordinates in
8865 space of dimension N = `(vector-length VECT)', the coordinates are
8866 uniformly distributed within the unit N-shere. The sum of the
8867 squares of the numbers is returned. VECT can be either a vector
8868 or a uniform vector of doubles.
8870 *** New procedure: random:hollow-sphere! VECT [STATE]
8871 Fills VECT with inexact real random numbers the sum of whose squares
8872 is equal to 1.0. Thinking of VECT as coordinates in space of
8873 dimension n = `(vector-length VECT)', the coordinates are uniformly
8874 distributed over the surface of the unit n-shere. VECT can be either
8875 a vector or a uniform vector of doubles.
8877 *** New function: random:normal [STATE]
8878 Returns an inexact real in a normal distribution with mean 0 and
8879 standard deviation 1. For a normal distribution with mean M and
8880 standard deviation D use `(+ M (* D (random:normal)))'.
8882 *** New procedure: random:normal-vector! VECT [STATE]
8883 Fills VECT with inexact real random numbers which are independent and
8884 standard normally distributed (i.e., with mean 0 and variance 1).
8885 VECT can be either a vector or a uniform vector of doubles.
8887 *** New function: random:exp STATE
8888 Returns an inexact real in an exponential distribution with mean 1.
8889 For an exponential distribution with mean U use (* U (random:exp)).
8891 ** The range of logand, logior, logxor, logtest, and logbit? have changed.
8893 These functions now operate on numbers in the range of a C unsigned
8896 These functions used to operate on numbers in the range of a C signed
8897 long; however, this seems inappropriate, because Guile integers don't
8900 ** New function: make-guardian
8901 This is an implementation of guardians as described in
8902 R. Kent Dybvig, Carl Bruggeman, and David Eby (1993) "Guardians in a
8903 Generation-Based Garbage Collector" ACM SIGPLAN Conference on
8904 Programming Language Design and Implementation, June 1993
8905 ftp://ftp.cs.indiana.edu/pub/scheme-repository/doc/pubs/guardians.ps.gz
8907 ** New functions: delq1!, delv1!, delete1!
8908 These procedures behave similar to delq! and friends but delete only
8909 one object if at all.
8911 ** New function: unread-string STRING PORT
8912 Unread STRING to PORT, that is, push it back onto the port so that
8913 next read operation will work on the pushed back characters.
8915 ** unread-char can now be called multiple times
8916 If unread-char is called multiple times, the unread characters will be
8917 read again in last-in first-out order.
8919 ** the procedures uniform-array-read! and uniform-array-write! now
8920 work on any kind of port, not just ports which are open on a file.
8922 ** Now 'l' in a port mode requests line buffering.
8924 ** The procedure truncate-file now works on string ports as well
8925 as file ports. If the size argument is omitted, the current
8926 file position is used.
8928 ** new procedure: seek PORT/FDES OFFSET WHENCE
8929 The arguments are the same as for the old fseek procedure, but it
8930 works on string ports as well as random-access file ports.
8932 ** the fseek procedure now works on string ports, since it has been
8933 redefined using seek.
8935 ** the setvbuf procedure now uses a default size if mode is _IOFBF and
8936 size is not supplied.
8938 ** the newline procedure no longer flushes the port if it's not
8939 line-buffered: previously it did if it was the current output port.
8941 ** open-pipe and close-pipe are no longer primitive procedures, but
8942 an emulation can be obtained using `(use-modules (ice-9 popen))'.
8944 ** the freopen procedure has been removed.
8946 ** new procedure: drain-input PORT
8947 Drains PORT's read buffers (including any pushed-back characters)
8948 and returns the contents as a single string.
8950 ** New function: map-in-order PROC LIST1 LIST2 ...
8951 Version of `map' which guarantees that the procedure is applied to the
8952 lists in serial order.
8954 ** Renamed `serial-array-copy!' and `serial-array-map!' to
8955 `array-copy-in-order!' and `array-map-in-order!'. The old names are
8956 now obsolete and will go away in release 1.5.
8958 ** New syntax: collect BODY1 ...
8959 Version of `begin' which returns a list of the results of the body
8960 forms instead of the result of the last body form. In contrast to
8961 `begin', `collect' allows an empty body.
8963 ** New functions: read-history FILENAME, write-history FILENAME
8964 Read/write command line history from/to file. Returns #t on success
8965 and #f if an error occured.
8967 ** `ls' and `lls' in module (ice-9 ls) now handle no arguments.
8969 These procedures return a list of definitions available in the specified
8970 argument, a relative module reference. In the case of no argument,
8971 `(current-module)' is now consulted for definitions to return, instead
8972 of simply returning #f, the former behavior.
8974 ** The #/ syntax for lists is no longer supported.
8976 Earlier versions of Scheme accepted this syntax, but printed a
8979 ** Guile no longer consults the SCHEME_LOAD_PATH environment variable.
8981 Instead, you should set GUILE_LOAD_PATH to tell Guile where to find
8984 * Changes to the gh_ interface
8988 Now takes a second argument which is the result array. If this
8989 pointer is NULL, a new array is malloced (the old behaviour).
8991 ** gh_chars2byvect, gh_shorts2svect, gh_floats2fvect, gh_scm2chars,
8992 gh_scm2shorts, gh_scm2longs, gh_scm2floats
8996 * Changes to the scm_ interface
8998 ** Function: scm_make_named_hook (char* name, int n_args)
9000 Creates a hook in the same way as make-hook above but also
9001 binds a variable named NAME to it.
9003 This is the typical way of creating a hook from C code.
9005 Currently, the variable is created in the "current" module. This
9006 might change when we get the new module system.
9008 ** The smob interface
9010 The interface for creating smobs has changed. For documentation, see
9011 data-rep.info (made from guile-core/doc/data-rep.texi).
9013 *** Deprecated function: SCM scm_newsmob (scm_smobfuns *)
9015 >>> This function will be removed in 1.3.4. <<<
9019 *** Function: SCM scm_make_smob_type (const char *name, scm_sizet size)
9020 This function adds a new smob type, named NAME, with instance size
9021 SIZE to the system. The return value is a tag that is used in
9022 creating instances of the type. If SIZE is 0, then no memory will
9023 be allocated when instances of the smob are created, and nothing
9024 will be freed by the default free function.
9026 *** Function: void scm_set_smob_mark (long tc, SCM (*mark) (SCM))
9027 This function sets the smob marking procedure for the smob type
9028 specified by the tag TC. TC is the tag returned by
9029 `scm_make_smob_type'.
9031 *** Function: void scm_set_smob_free (long tc, SCM (*mark) (SCM))
9032 This function sets the smob freeing procedure for the smob type
9033 specified by the tag TC. TC is the tag returned by
9034 `scm_make_smob_type'.
9036 *** Function: void scm_set_smob_print (tc, print)
9038 - Function: void scm_set_smob_print (long tc,
9039 scm_sizet (*print) (SCM,
9043 This function sets the smob printing procedure for the smob type
9044 specified by the tag TC. TC is the tag returned by
9045 `scm_make_smob_type'.
9047 *** Function: void scm_set_smob_equalp (long tc, SCM (*equalp) (SCM, SCM))
9048 This function sets the smob equality-testing predicate for the
9049 smob type specified by the tag TC. TC is the tag returned by
9050 `scm_make_smob_type'.
9052 *** Macro: void SCM_NEWSMOB (SCM var, long tc, void *data)
9053 Make VALUE contain a smob instance of the type with type code TC and
9054 smob data DATA. VALUE must be previously declared as C type `SCM'.
9056 *** Macro: fn_returns SCM_RETURN_NEWSMOB (long tc, void *data)
9057 This macro expands to a block of code that creates a smob instance
9058 of the type with type code TC and smob data DATA, and returns that
9059 `SCM' value. It should be the last piece of code in a block.
9061 ** The interfaces for using I/O ports and implementing port types
9062 (ptobs) have changed significantly. The new interface is based on
9063 shared access to buffers and a new set of ptob procedures.
9065 *** scm_newptob has been removed
9069 *** Function: SCM scm_make_port_type (type_name, fill_buffer, write_flush)
9071 - Function: SCM scm_make_port_type (char *type_name,
9072 int (*fill_buffer) (SCM port),
9073 void (*write_flush) (SCM port));
9075 Similarly to the new smob interface, there is a set of function
9076 setters by which the user can customize the behaviour of his port
9077 type. See ports.h (scm_set_port_XXX).
9079 ** scm_strport_to_string: New function: creates a new string from
9080 a string port's buffer.
9082 ** Plug in interface for random number generators
9083 The variable `scm_the_rng' in random.c contains a value and three
9084 function pointers which together define the current random number
9085 generator being used by the Scheme level interface and the random
9086 number library functions.
9088 The user is free to replace the default generator with the generator
9091 *** Variable: size_t scm_the_rng.rstate_size
9092 The size of the random state type used by the current RNG
9095 *** Function: unsigned long scm_the_rng.random_bits (scm_rstate *STATE)
9096 Given the random STATE, return 32 random bits.
9098 *** Function: void scm_the_rng.init_rstate (scm_rstate *STATE, chars *S, int N)
9099 Seed random state STATE using string S of length N.
9101 *** Function: scm_rstate *scm_the_rng.copy_rstate (scm_rstate *STATE)
9102 Given random state STATE, return a malloced copy.
9105 The default RNG is the MWC (Multiply With Carry) random number
9106 generator described by George Marsaglia at the Department of
9107 Statistics and Supercomputer Computations Research Institute, The
9108 Florida State University (http://stat.fsu.edu/~geo).
9110 It uses 64 bits, has a period of 4578426017172946943 (4.6e18), and
9111 passes all tests in the DIEHARD test suite
9112 (http://stat.fsu.edu/~geo/diehard.html). The generation of 32 bits
9113 costs one multiply and one add on platforms which either supports long
9114 longs (gcc does this on most systems) or have 64 bit longs. The cost
9115 is four multiply on other systems but this can be optimized by writing
9116 scm_i_uniform32 in assembler.
9118 These functions are provided through the scm_the_rng interface for use
9119 by libguile and the application.
9121 *** Function: unsigned long scm_i_uniform32 (scm_i_rstate *STATE)
9122 Given the random STATE, return 32 random bits.
9123 Don't use this function directly. Instead go through the plugin
9124 interface (see "Plug in interface" above).
9126 *** Function: void scm_i_init_rstate (scm_i_rstate *STATE, char *SEED, int N)
9127 Initialize STATE using SEED of length N.
9129 *** Function: scm_i_rstate *scm_i_copy_rstate (scm_i_rstate *STATE)
9130 Return a malloc:ed copy of STATE. This function can easily be re-used
9131 in the interfaces to other RNGs.
9133 ** Random number library functions
9134 These functions use the current RNG through the scm_the_rng interface.
9135 It might be a good idea to use these functions from your C code so
9136 that only one random generator is used by all code in your program.
9138 The default random state is stored in:
9140 *** Variable: SCM scm_var_random_state
9141 Contains the vcell of the Scheme variable "*random-state*" which is
9142 used as default state by all random number functions in the Scheme
9147 double x = scm_c_uniform01 (SCM_RSTATE (SCM_CDR (scm_var_random_state)));
9149 *** Function: scm_rstate *scm_c_default_rstate (void)
9150 This is a convenience function which returns the value of
9151 scm_var_random_state. An error message is generated if this value
9152 isn't a random state.
9154 *** Function: scm_rstate *scm_c_make_rstate (char *SEED, int LENGTH)
9155 Make a new random state from the string SEED of length LENGTH.
9157 It is generally not a good idea to use multiple random states in a
9158 program. While subsequent random numbers generated from one random
9159 state are guaranteed to be reasonably independent, there is no such
9160 guarantee for numbers generated from different random states.
9162 *** Macro: unsigned long scm_c_uniform32 (scm_rstate *STATE)
9163 Return 32 random bits.
9165 *** Function: double scm_c_uniform01 (scm_rstate *STATE)
9166 Return a sample from the uniform(0,1) distribution.
9168 *** Function: double scm_c_normal01 (scm_rstate *STATE)
9169 Return a sample from the normal(0,1) distribution.
9171 *** Function: double scm_c_exp1 (scm_rstate *STATE)
9172 Return a sample from the exp(1) distribution.
9174 *** Function: unsigned long scm_c_random (scm_rstate *STATE, unsigned long M)
9175 Return a sample from the discrete uniform(0,M) distribution.
9177 *** Function: SCM scm_c_random_bignum (scm_rstate *STATE, SCM M)
9178 Return a sample from the discrete uniform(0,M) distribution.
9179 M must be a bignum object. The returned value may be an INUM.
9183 Changes in Guile 1.3 (released Monday, October 19, 1998):
9185 * Changes to the distribution
9187 ** We renamed the SCHEME_LOAD_PATH environment variable to GUILE_LOAD_PATH.
9188 To avoid conflicts, programs should name environment variables after
9189 themselves, except when there's a common practice establishing some
9192 For now, Guile supports both GUILE_LOAD_PATH and SCHEME_LOAD_PATH,
9193 giving the former precedence, and printing a warning message if the
9194 latter is set. Guile 1.4 will not recognize SCHEME_LOAD_PATH at all.
9196 ** The header files related to multi-byte characters have been removed.
9197 They were: libguile/extchrs.h and libguile/mbstrings.h. Any C code
9198 which referred to these explicitly will probably need to be rewritten,
9199 since the support for the variant string types has been removed; see
9202 ** The header files append.h and sequences.h have been removed. These
9203 files implemented non-R4RS operations which would encourage
9204 non-portable programming style and less easy-to-read code.
9206 * Changes to the stand-alone interpreter
9208 ** New procedures have been added to implement a "batch mode":
9210 *** Function: batch-mode?
9212 Returns a boolean indicating whether the interpreter is in batch
9215 *** Function: set-batch-mode?! ARG
9217 If ARG is true, switches the interpreter to batch mode. The `#f'
9218 case has not been implemented.
9220 ** Guile now provides full command-line editing, when run interactively.
9221 To use this feature, you must have the readline library installed.
9222 The Guile build process will notice it, and automatically include
9225 The readline library is available via anonymous FTP from any GNU
9226 mirror site; the canonical location is "ftp://prep.ai.mit.edu/pub/gnu".
9228 ** the-last-stack is now a fluid.
9230 * Changes to the procedure for linking libguile with your programs
9232 ** You can now use the `guile-config' utility to build programs that use Guile.
9234 Guile now includes a command-line utility called `guile-config', which
9235 can provide information about how to compile and link programs that
9238 *** `guile-config compile' prints any C compiler flags needed to use Guile.
9239 You should include this command's output on the command line you use
9240 to compile C or C++ code that #includes the Guile header files. It's
9241 usually just a `-I' flag to help the compiler find the Guile headers.
9244 *** `guile-config link' prints any linker flags necessary to link with Guile.
9246 This command writes to its standard output a list of flags which you
9247 must pass to the linker to link your code against the Guile library.
9248 The flags include '-lguile' itself, any other libraries the Guile
9249 library depends upon, and any `-L' flags needed to help the linker
9250 find those libraries.
9252 For example, here is a Makefile rule that builds a program named 'foo'
9253 from the object files ${FOO_OBJECTS}, and links them against Guile:
9256 ${CC} ${CFLAGS} ${FOO_OBJECTS} `guile-config link` -o foo
9258 Previous Guile releases recommended that you use autoconf to detect
9259 which of a predefined set of libraries were present on your system.
9260 It is more robust to use `guile-config', since it records exactly which
9261 libraries the installed Guile library requires.
9263 This was originally called `build-guile', but was renamed to
9264 `guile-config' before Guile 1.3 was released, to be consistent with
9265 the analogous script for the GTK+ GUI toolkit, which is called
9269 ** Use the GUILE_FLAGS macro in your configure.in file to find Guile.
9271 If you are using the GNU autoconf package to configure your program,
9272 you can use the GUILE_FLAGS autoconf macro to call `guile-config'
9273 (described above) and gather the necessary values for use in your
9276 The GUILE_FLAGS macro expands to configure script code which runs the
9277 `guile-config' script, to find out where Guile's header files and
9278 libraries are installed. It sets two variables, marked for
9279 substitution, as by AC_SUBST.
9281 GUILE_CFLAGS --- flags to pass to a C or C++ compiler to build
9282 code that uses Guile header files. This is almost always just a
9285 GUILE_LDFLAGS --- flags to pass to the linker to link a
9286 program against Guile. This includes `-lguile' for the Guile
9287 library itself, any libraries that Guile itself requires (like
9288 -lqthreads), and so on. It may also include a -L flag to tell the
9289 compiler where to find the libraries.
9291 GUILE_FLAGS is defined in the file guile.m4, in the top-level
9292 directory of the Guile distribution. You can copy it into your
9293 package's aclocal.m4 file, and then use it in your configure.in file.
9295 If you are using the `aclocal' program, distributed with GNU automake,
9296 to maintain your aclocal.m4 file, the Guile installation process
9297 installs guile.m4 where aclocal will find it. All you need to do is
9298 use GUILE_FLAGS in your configure.in file, and then run `aclocal';
9299 this will copy the definition of GUILE_FLAGS into your aclocal.m4
9303 * Changes to Scheme functions and syntax
9305 ** Multi-byte strings have been removed, as have multi-byte and wide
9306 ports. We felt that these were the wrong approach to
9307 internationalization support.
9309 ** New function: readline [PROMPT]
9310 Read a line from the terminal, and allow the user to edit it,
9311 prompting with PROMPT. READLINE provides a large set of Emacs-like
9312 editing commands, lets the user recall previously typed lines, and
9313 works on almost every kind of terminal, including dumb terminals.
9315 READLINE assumes that the cursor is at the beginning of the line when
9316 it is invoked. Thus, you can't print a prompt yourself, and then call
9317 READLINE; you need to package up your prompt as a string, pass it to
9318 the function, and let READLINE print the prompt itself. This is
9319 because READLINE needs to know the prompt's screen width.
9321 For Guile to provide this function, you must have the readline
9322 library, version 2.1 or later, installed on your system. Readline is
9323 available via anonymous FTP from prep.ai.mit.edu in pub/gnu, or from
9324 any GNU mirror site.
9326 See also ADD-HISTORY function.
9328 ** New function: add-history STRING
9329 Add STRING as the most recent line in the history used by the READLINE
9330 command. READLINE does not add lines to the history itself; you must
9331 call ADD-HISTORY to make previous input available to the user.
9333 ** The behavior of the read-line function has changed.
9335 This function now uses standard C library functions to read the line,
9336 for speed. This means that it doesn not respect the value of
9337 scm-line-incrementors; it assumes that lines are delimited with
9340 (Note that this is read-line, the function that reads a line of text
9341 from a port, not readline, the function that reads a line from a
9342 terminal, providing full editing capabilities.)
9344 ** New module (ice-9 getopt-gnu-style): Parse command-line arguments.
9346 This module provides some simple argument parsing. It exports one
9349 Function: getopt-gnu-style ARG-LS
9350 Parse a list of program arguments into an alist of option
9353 Each item in the list of program arguments is examined to see if
9354 it meets the syntax of a GNU long-named option. An argument like
9355 `--MUMBLE' produces an element of the form (MUMBLE . #t) in the
9356 returned alist, where MUMBLE is a keyword object with the same
9357 name as the argument. An argument like `--MUMBLE=FROB' produces
9358 an element of the form (MUMBLE . FROB), where FROB is a string.
9360 As a special case, the returned alist also contains a pair whose
9361 car is the symbol `rest'. The cdr of this pair is a list
9362 containing all the items in the argument list that are not options
9363 of the form mentioned above.
9365 The argument `--' is treated specially: all items in the argument
9366 list appearing after such an argument are not examined, and are
9367 returned in the special `rest' list.
9369 This function does not parse normal single-character switches.
9370 You will need to parse them out of the `rest' list yourself.
9372 ** The read syntax for byte vectors and short vectors has changed.
9374 Instead of #bytes(...), write #y(...).
9376 Instead of #short(...), write #h(...).
9378 This may seem nutty, but, like the other uniform vectors, byte vectors
9379 and short vectors want to have the same print and read syntax (and,
9380 more basic, want to have read syntax!). Changing the read syntax to
9381 use multiple characters after the hash sign breaks with the
9382 conventions used in R5RS and the conventions used for the other
9383 uniform vectors. It also introduces complexity in the current reader,
9384 both on the C and Scheme levels. (The Right solution is probably to
9385 change the syntax and prototypes for uniform vectors entirely.)
9388 ** The new module (ice-9 session) provides useful interactive functions.
9390 *** New procedure: (apropos REGEXP OPTION ...)
9392 Display a list of top-level variables whose names match REGEXP, and
9393 the modules they are imported from. Each OPTION should be one of the
9396 value --- Show the value of each matching variable.
9397 shadow --- Show bindings shadowed by subsequently imported modules.
9398 full --- Same as both `shadow' and `value'.
9402 guile> (apropos "trace" 'full)
9403 debug: trace #<procedure trace args>
9404 debug: untrace #<procedure untrace args>
9405 the-scm-module: display-backtrace #<compiled-closure #<primitive-procedure gsubr-apply>>
9406 the-scm-module: before-backtrace-hook ()
9407 the-scm-module: backtrace #<primitive-procedure backtrace>
9408 the-scm-module: after-backtrace-hook ()
9409 the-scm-module: has-shown-backtrace-hint? #f
9412 ** There are new functions and syntax for working with macros.
9414 Guile implements macros as a special object type. Any variable whose
9415 top-level binding is a macro object acts as a macro. The macro object
9416 specifies how the expression should be transformed before evaluation.
9418 *** Macro objects now print in a reasonable way, resembling procedures.
9420 *** New function: (macro? OBJ)
9421 True iff OBJ is a macro object.
9423 *** New function: (primitive-macro? OBJ)
9424 Like (macro? OBJ), but true only if OBJ is one of the Guile primitive
9425 macro transformers, implemented in eval.c rather than Scheme code.
9427 Why do we have this function?
9428 - For symmetry with procedure? and primitive-procedure?,
9429 - to allow custom print procedures to tell whether a macro is
9430 primitive, and display it differently, and
9431 - to allow compilers and user-written evaluators to distinguish
9432 builtin special forms from user-defined ones, which could be
9435 *** New function: (macro-type OBJ)
9436 Return a value indicating what kind of macro OBJ is. Possible return
9439 The symbol `syntax' --- a macro created by procedure->syntax.
9440 The symbol `macro' --- a macro created by procedure->macro.
9441 The symbol `macro!' --- a macro created by procedure->memoizing-macro.
9442 The boolean #f --- if OBJ is not a macro object.
9444 *** New function: (macro-name MACRO)
9445 Return the name of the macro object MACRO's procedure, as returned by
9448 *** New function: (macro-transformer MACRO)
9449 Return the transformer procedure for MACRO.
9451 *** New syntax: (use-syntax MODULE ... TRANSFORMER)
9453 Specify a new macro expander to use in the current module. Each
9454 MODULE is a module name, with the same meaning as in the `use-modules'
9455 form; each named module's exported bindings are added to the current
9456 top-level environment. TRANSFORMER is an expression evaluated in the
9457 resulting environment which must yield a procedure to use as the
9458 module's eval transformer: every expression evaluated in this module
9459 is passed to this function, and the result passed to the Guile
9462 *** macro-eval! is removed. Use local-eval instead.
9464 ** Some magic has been added to the printer to better handle user
9465 written printing routines (like record printers, closure printers).
9467 The problem is that these user written routines must have access to
9468 the current `print-state' to be able to handle fancy things like
9469 detection of circular references. These print-states have to be
9470 passed to the builtin printing routines (display, write, etc) to
9471 properly continue the print chain.
9473 We didn't want to change all existing print code so that it
9474 explicitly passes thru a print state in addition to a port. Instead,
9475 we extented the possible values that the builtin printing routines
9476 accept as a `port'. In addition to a normal port, they now also take
9477 a pair of a normal port and a print-state. Printing will go to the
9478 port and the print-state will be used to control the detection of
9479 circular references, etc. If the builtin function does not care for a
9480 print-state, it is simply ignored.
9482 User written callbacks are now called with such a pair as their
9483 `port', but because every function now accepts this pair as a PORT
9484 argument, you don't have to worry about that. In fact, it is probably
9485 safest to not check for these pairs.
9487 However, it is sometimes necessary to continue a print chain on a
9488 different port, for example to get a intermediate string
9489 representation of the printed value, mangle that string somehow, and
9490 then to finally print the mangled string. Use the new function
9492 inherit-print-state OLD-PORT NEW-PORT
9494 for this. It constructs a new `port' that prints to NEW-PORT but
9495 inherits the print-state of OLD-PORT.
9497 ** struct-vtable-offset renamed to vtable-offset-user
9499 ** New constants: vtable-index-layout, vtable-index-vtable, vtable-index-printer
9501 ** There is now a third optional argument to make-vtable-vtable
9502 (and fourth to make-struct) when constructing new types (vtables).
9503 This argument initializes field vtable-index-printer of the vtable.
9505 ** The detection of circular references has been extended to structs.
9506 That is, a structure that -- in the process of being printed -- prints
9507 itself does not lead to infinite recursion.
9509 ** There is now some basic support for fluids. Please read
9510 "libguile/fluid.h" to find out more. It is accessible from Scheme with
9511 the following functions and macros:
9513 Function: make-fluid
9515 Create a new fluid object. Fluids are not special variables or
9516 some other extension to the semantics of Scheme, but rather
9517 ordinary Scheme objects. You can store them into variables (that
9518 are still lexically scoped, of course) or into any other place you
9519 like. Every fluid has a initial value of `#f'.
9521 Function: fluid? OBJ
9523 Test whether OBJ is a fluid.
9525 Function: fluid-ref FLUID
9526 Function: fluid-set! FLUID VAL
9528 Access/modify the fluid FLUID. Modifications are only visible
9529 within the current dynamic root (that includes threads).
9531 Function: with-fluids* FLUIDS VALUES THUNK
9533 FLUIDS is a list of fluids and VALUES a corresponding list of
9534 values for these fluids. Before THUNK gets called the values are
9535 installed in the fluids and the old values of the fluids are
9536 saved in the VALUES list. When the flow of control leaves THUNK
9537 or reenters it, the values get swapped again. You might think of
9538 this as a `safe-fluid-excursion'. Note that the VALUES list is
9539 modified by `with-fluids*'.
9541 Macro: with-fluids ((FLUID VALUE) ...) FORM ...
9543 The same as `with-fluids*' but with a different syntax. It looks
9544 just like `let', but both FLUID and VALUE are evaluated. Remember,
9545 fluids are not special variables but ordinary objects. FLUID
9546 should evaluate to a fluid.
9548 ** Changes to system call interfaces:
9550 *** close-port, close-input-port and close-output-port now return a
9551 boolean instead of an `unspecified' object. #t means that the port
9552 was successfully closed, while #f means it was already closed. It is
9553 also now possible for these procedures to raise an exception if an
9554 error occurs (some errors from write can be delayed until close.)
9556 *** the first argument to chmod, fcntl, ftell and fseek can now be a
9559 *** the third argument to fcntl is now optional.
9561 *** the first argument to chown can now be a file descriptor or a port.
9563 *** the argument to stat can now be a port.
9565 *** The following new procedures have been added (most use scsh
9568 *** procedure: close PORT/FD
9569 Similar to close-port (*note close-port: Closing Ports.), but also
9570 works on file descriptors. A side effect of closing a file
9571 descriptor is that any ports using that file descriptor are moved
9572 to a different file descriptor and have their revealed counts set
9575 *** procedure: port->fdes PORT
9576 Returns the integer file descriptor underlying PORT. As a side
9577 effect the revealed count of PORT is incremented.
9579 *** procedure: fdes->ports FDES
9580 Returns a list of existing ports which have FDES as an underlying
9581 file descriptor, without changing their revealed counts.
9583 *** procedure: fdes->inport FDES
9584 Returns an existing input port which has FDES as its underlying
9585 file descriptor, if one exists, and increments its revealed count.
9586 Otherwise, returns a new input port with a revealed count of 1.
9588 *** procedure: fdes->outport FDES
9589 Returns an existing output port which has FDES as its underlying
9590 file descriptor, if one exists, and increments its revealed count.
9591 Otherwise, returns a new output port with a revealed count of 1.
9593 The next group of procedures perform a `dup2' system call, if NEWFD
9594 (an integer) is supplied, otherwise a `dup'. The file descriptor to be
9595 duplicated can be supplied as an integer or contained in a port. The
9596 type of value returned varies depending on which procedure is used.
9598 All procedures also have the side effect when performing `dup2' that
9599 any ports using NEWFD are moved to a different file descriptor and have
9600 their revealed counts set to zero.
9602 *** procedure: dup->fdes PORT/FD [NEWFD]
9603 Returns an integer file descriptor.
9605 *** procedure: dup->inport PORT/FD [NEWFD]
9606 Returns a new input port using the new file descriptor.
9608 *** procedure: dup->outport PORT/FD [NEWFD]
9609 Returns a new output port using the new file descriptor.
9611 *** procedure: dup PORT/FD [NEWFD]
9612 Returns a new port if PORT/FD is a port, with the same mode as the
9613 supplied port, otherwise returns an integer file descriptor.
9615 *** procedure: dup->port PORT/FD MODE [NEWFD]
9616 Returns a new port using the new file descriptor. MODE supplies a
9617 mode string for the port (*note open-file: File Ports.).
9619 *** procedure: setenv NAME VALUE
9620 Modifies the environment of the current process, which is also the
9621 default environment inherited by child processes.
9623 If VALUE is `#f', then NAME is removed from the environment.
9624 Otherwise, the string NAME=VALUE is added to the environment,
9625 replacing any existing string with name matching NAME.
9627 The return value is unspecified.
9629 *** procedure: truncate-file OBJ SIZE
9630 Truncates the file referred to by OBJ to at most SIZE bytes. OBJ
9631 can be a string containing a file name or an integer file
9632 descriptor or port open for output on the file. The underlying
9633 system calls are `truncate' and `ftruncate'.
9635 The return value is unspecified.
9637 *** procedure: setvbuf PORT MODE [SIZE]
9638 Set the buffering mode for PORT. MODE can be:
9646 block buffered, using a newly allocated buffer of SIZE bytes.
9647 However if SIZE is zero or unspecified, the port will be made
9650 This procedure should not be used after I/O has been performed with
9653 Ports are usually block buffered by default, with a default buffer
9654 size. Procedures e.g., *Note open-file: File Ports, which accept a
9655 mode string allow `0' to be added to request an unbuffered port.
9657 *** procedure: fsync PORT/FD
9658 Copies any unwritten data for the specified output file descriptor
9659 to disk. If PORT/FD is a port, its buffer is flushed before the
9660 underlying file descriptor is fsync'd. The return value is
9663 *** procedure: open-fdes PATH FLAGS [MODES]
9664 Similar to `open' but returns a file descriptor instead of a port.
9666 *** procedure: execle PATH ENV [ARG] ...
9667 Similar to `execl', but the environment of the new process is
9668 specified by ENV, which must be a list of strings as returned by
9669 the `environ' procedure.
9671 This procedure is currently implemented using the `execve' system
9672 call, but we call it `execle' because of its Scheme calling
9675 *** procedure: strerror ERRNO
9676 Returns the Unix error message corresponding to ERRNO, an integer.
9678 *** procedure: primitive-exit [STATUS]
9679 Terminate the current process without unwinding the Scheme stack.
9680 This is would typically be useful after a fork. The exit status
9681 is STATUS if supplied, otherwise zero.
9683 *** procedure: times
9684 Returns an object with information about real and processor time.
9685 The following procedures accept such an object as an argument and
9686 return a selected component:
9689 The current real time, expressed as time units relative to an
9693 The CPU time units used by the calling process.
9696 The CPU time units used by the system on behalf of the
9700 The CPU time units used by terminated child processes of the
9701 calling process, whose status has been collected (e.g., using
9705 Similarly, the CPU times units used by the system on behalf of
9706 terminated child processes.
9708 ** Removed: list-length
9709 ** Removed: list-append, list-append!
9710 ** Removed: list-reverse, list-reverse!
9712 ** array-map renamed to array-map!
9714 ** serial-array-map renamed to serial-array-map!
9716 ** catch doesn't take #f as first argument any longer
9718 Previously, it was possible to pass #f instead of a key to `catch'.
9719 That would cause `catch' to pass a jump buffer object to the procedure
9720 passed as second argument. The procedure could then use this jump
9721 buffer objekt as an argument to throw.
9723 This mechanism has been removed since its utility doesn't motivate the
9724 extra complexity it introduces.
9726 ** The `#/' notation for lists now provokes a warning message from Guile.
9727 This syntax will be removed from Guile in the near future.
9729 To disable the warning message, set the GUILE_HUSH environment
9730 variable to any non-empty value.
9732 ** The newline character now prints as `#\newline', following the
9733 normal Scheme notation, not `#\nl'.
9735 * Changes to the gh_ interface
9737 ** The gh_enter function now takes care of loading the Guile startup files.
9738 gh_enter works by calling scm_boot_guile; see the remarks below.
9740 ** Function: void gh_write (SCM x)
9742 Write the printed representation of the scheme object x to the current
9743 output port. Corresponds to the scheme level `write'.
9745 ** gh_list_length renamed to gh_length.
9747 ** vector handling routines
9749 Several major changes. In particular, gh_vector() now resembles
9750 (vector ...) (with a caveat -- see manual), and gh_make_vector() now
9751 exists and behaves like (make-vector ...). gh_vset() and gh_vref()
9752 have been renamed gh_vector_set_x() and gh_vector_ref(). Some missing
9753 vector-related gh_ functions have been implemented.
9755 ** pair and list routines
9757 Implemented several of the R4RS pair and list functions that were
9760 ** gh_scm2doubles, gh_doubles2scm, gh_doubles2dvect
9762 New function. Converts double arrays back and forth between Scheme
9765 * Changes to the scm_ interface
9767 ** The function scm_boot_guile now takes care of loading the startup files.
9769 Guile's primary initialization function, scm_boot_guile, now takes
9770 care of loading `boot-9.scm', in the `ice-9' module, to initialize
9771 Guile, define the module system, and put together some standard
9772 bindings. It also loads `init.scm', which is intended to hold
9773 site-specific initialization code.
9775 Since Guile cannot operate properly until boot-9.scm is loaded, there
9776 is no reason to separate loading boot-9.scm from Guile's other
9777 initialization processes.
9779 This job used to be done by scm_compile_shell_switches, which didn't
9780 make much sense; in particular, it meant that people using Guile for
9781 non-shell-like applications had to jump through hoops to get Guile
9782 initialized properly.
9784 ** The function scm_compile_shell_switches no longer loads the startup files.
9785 Now, Guile always loads the startup files, whenever it is initialized;
9786 see the notes above for scm_boot_guile and scm_load_startup_files.
9788 ** Function: scm_load_startup_files
9789 This new function takes care of loading Guile's initialization file
9790 (`boot-9.scm'), and the site initialization file, `init.scm'. Since
9791 this is always called by the Guile initialization process, it's
9792 probably not too useful to call this yourself, but it's there anyway.
9794 ** The semantics of smob marking have changed slightly.
9796 The smob marking function (the `mark' member of the scm_smobfuns
9797 structure) is no longer responsible for setting the mark bit on the
9798 smob. The generic smob handling code in the garbage collector will
9799 set this bit. The mark function need only ensure that any other
9800 objects the smob refers to get marked.
9802 Note that this change means that the smob's GC8MARK bit is typically
9803 already set upon entry to the mark function. Thus, marking functions
9804 which look like this:
9807 if (SCM_GC8MARKP (ptr))
9809 SCM_SETGC8MARK (ptr);
9810 ... mark objects to which the smob refers ...
9813 are now incorrect, since they will return early, and fail to mark any
9814 other objects the smob refers to. Some code in the Guile library used
9817 ** The semantics of the I/O port functions in scm_ptobfuns have changed.
9819 If you have implemented your own I/O port type, by writing the
9820 functions required by the scm_ptobfuns and then calling scm_newptob,
9821 you will need to change your functions slightly.
9823 The functions in a scm_ptobfuns structure now expect the port itself
9824 as their argument; they used to expect the `stream' member of the
9825 port's scm_port_table structure. This allows functions in an
9826 scm_ptobfuns structure to easily access the port's cell (and any flags
9827 it its CAR), and the port's scm_port_table structure.
9829 Guile now passes the I/O port itself as the `port' argument in the
9830 following scm_ptobfuns functions:
9832 int (*free) (SCM port);
9833 int (*fputc) (int, SCM port);
9834 int (*fputs) (char *, SCM port);
9835 scm_sizet (*fwrite) SCM_P ((char *ptr,
9839 int (*fflush) (SCM port);
9840 int (*fgetc) (SCM port);
9841 int (*fclose) (SCM port);
9843 The interfaces to the `mark', `print', `equalp', and `fgets' methods
9846 If you have existing code which defines its own port types, it is easy
9847 to convert your code to the new interface; simply apply SCM_STREAM to
9848 the port argument to yield the value you code used to expect.
9850 Note that since both the port and the stream have the same type in the
9851 C code --- they are both SCM values --- the C compiler will not remind
9852 you if you forget to update your scm_ptobfuns functions.
9855 ** Function: int scm_internal_select (int fds,
9859 struct timeval *timeout);
9861 This is a replacement for the `select' function provided by the OS.
9862 It enables I/O blocking and sleeping to happen for one cooperative
9863 thread without blocking other threads. It also avoids busy-loops in
9864 these situations. It is intended that all I/O blocking and sleeping
9865 will finally go through this function. Currently, this function is
9866 only available on systems providing `gettimeofday' and `select'.
9868 ** Function: SCM scm_internal_stack_catch (SCM tag,
9869 scm_catch_body_t body,
9871 scm_catch_handler_t handler,
9874 A new sibling to the other two C level `catch' functions
9875 scm_internal_catch and scm_internal_lazy_catch. Use it if you want
9876 the stack to be saved automatically into the variable `the-last-stack'
9877 (scm_the_last_stack_var) on error. This is necessary if you want to
9878 use advanced error reporting, such as calling scm_display_error and
9879 scm_display_backtrace. (They both take a stack object as argument.)
9881 ** Function: SCM scm_spawn_thread (scm_catch_body_t body,
9883 scm_catch_handler_t handler,
9886 Spawns a new thread. It does a job similar to
9887 scm_call_with_new_thread but takes arguments more suitable when
9888 spawning threads from application C code.
9890 ** The hook scm_error_callback has been removed. It was originally
9891 intended as a way for the user to install his own error handler. But
9892 that method works badly since it intervenes between throw and catch,
9893 thereby changing the semantics of expressions like (catch #t ...).
9894 The correct way to do it is to use one of the C level catch functions
9895 in throw.c: scm_internal_catch/lazy_catch/stack_catch.
9897 ** Removed functions:
9899 scm_obj_length, scm_list_length, scm_list_append, scm_list_append_x,
9900 scm_list_reverse, scm_list_reverse_x
9902 ** New macros: SCM_LISTn where n is one of the integers 0-9.
9904 These can be used for pretty list creation from C. The idea is taken
9905 from Erick Gallesio's STk.
9907 ** scm_array_map renamed to scm_array_map_x
9909 ** mbstrings are now removed
9911 This means that the type codes scm_tc7_mb_string and
9912 scm_tc7_mb_substring has been removed.
9914 ** scm_gen_putc, scm_gen_puts, scm_gen_write, and scm_gen_getc have changed.
9916 Since we no longer support multi-byte strings, these I/O functions
9917 have been simplified, and renamed. Here are their old names, and
9918 their new names and arguments:
9920 scm_gen_putc -> void scm_putc (int c, SCM port);
9921 scm_gen_puts -> void scm_puts (char *s, SCM port);
9922 scm_gen_write -> void scm_lfwrite (char *ptr, scm_sizet size, SCM port);
9923 scm_gen_getc -> void scm_getc (SCM port);
9926 ** The macros SCM_TYP7D and SCM_TYP7SD has been removed.
9928 ** The macro SCM_TYP7S has taken the role of the old SCM_TYP7D
9930 SCM_TYP7S now masks away the bit which distinguishes substrings from
9933 ** scm_catch_body_t: Backward incompatible change!
9935 Body functions to scm_internal_catch and friends do not any longer
9936 take a second argument. This is because it is no longer possible to
9937 pass a #f arg to catch.
9939 ** Calls to scm_protect_object and scm_unprotect now nest properly.
9941 The function scm_protect_object protects its argument from being freed
9942 by the garbage collector. scm_unprotect_object removes that
9945 These functions now nest properly. That is, for every object O, there
9946 is a counter which scm_protect_object(O) increments and
9947 scm_unprotect_object(O) decrements, if the counter is greater than
9948 zero. Every object's counter is zero when it is first created. If an
9949 object's counter is greater than zero, the garbage collector will not
9950 reclaim its storage.
9952 This allows you to use scm_protect_object in your code without
9953 worrying that some other function you call will call
9954 scm_unprotect_object, and allow it to be freed. Assuming that the
9955 functions you call are well-behaved, and unprotect only those objects
9956 they protect, you can follow the same rule and have confidence that
9957 objects will be freed only at appropriate times.
9960 Changes in Guile 1.2 (released Tuesday, June 24 1997):
9962 * Changes to the distribution
9964 ** Nightly snapshots are now available from ftp.red-bean.com.
9965 The old server, ftp.cyclic.com, has been relinquished to its rightful
9968 Nightly snapshots of the Guile development sources are now available via
9969 anonymous FTP from ftp.red-bean.com, as /pub/guile/guile-snap.tar.gz.
9971 Via the web, that's: ftp://ftp.red-bean.com/pub/guile/guile-snap.tar.gz
9972 For getit, that's: ftp.red-bean.com:/pub/guile/guile-snap.tar.gz
9974 ** To run Guile without installing it, the procedure has changed a bit.
9976 If you used a separate build directory to compile Guile, you'll need
9977 to include the build directory in SCHEME_LOAD_PATH, as well as the
9978 source directory. See the `INSTALL' file for examples.
9980 * Changes to the procedure for linking libguile with your programs
9982 ** The standard Guile load path for Scheme code now includes
9983 $(datadir)/guile (usually /usr/local/share/guile). This means that
9984 you can install your own Scheme files there, and Guile will find them.
9985 (Previous versions of Guile only checked a directory whose name
9986 contained the Guile version number, so you had to re-install or move
9987 your Scheme sources each time you installed a fresh version of Guile.)
9989 The load path also includes $(datadir)/guile/site; we recommend
9990 putting individual Scheme files there. If you want to install a
9991 package with multiple source files, create a directory for them under
9994 ** Guile 1.2 will now use the Rx regular expression library, if it is
9995 installed on your system. When you are linking libguile into your own
9996 programs, this means you will have to link against -lguile, -lqt (if
9997 you configured Guile with thread support), and -lrx.
9999 If you are using autoconf to generate configuration scripts for your
10000 application, the following lines should suffice to add the appropriate
10001 libraries to your link command:
10003 ### Find Rx, quickthreads and libguile.
10004 AC_CHECK_LIB(rx, main)
10005 AC_CHECK_LIB(qt, main)
10006 AC_CHECK_LIB(guile, scm_shell)
10008 The Guile 1.2 distribution does not contain sources for the Rx
10009 library, as Guile 1.0 did. If you want to use Rx, you'll need to
10010 retrieve it from a GNU FTP site and install it separately.
10012 * Changes to Scheme functions and syntax
10014 ** The dynamic linking features of Guile are now enabled by default.
10015 You can disable them by giving the `--disable-dynamic-linking' option
10018 (dynamic-link FILENAME)
10020 Find the object file denoted by FILENAME (a string) and link it
10021 into the running Guile application. When everything works out,
10022 return a Scheme object suitable for representing the linked object
10023 file. Otherwise an error is thrown. How object files are
10024 searched is system dependent.
10026 (dynamic-object? VAL)
10028 Determine whether VAL represents a dynamically linked object file.
10030 (dynamic-unlink DYNOBJ)
10032 Unlink the indicated object file from the application. DYNOBJ
10033 should be one of the values returned by `dynamic-link'.
10035 (dynamic-func FUNCTION DYNOBJ)
10037 Search the C function indicated by FUNCTION (a string or symbol)
10038 in DYNOBJ and return some Scheme object that can later be used
10039 with `dynamic-call' to actually call this function. Right now,
10040 these Scheme objects are formed by casting the address of the
10041 function to `long' and converting this number to its Scheme
10044 (dynamic-call FUNCTION DYNOBJ)
10046 Call the C function indicated by FUNCTION and DYNOBJ. The
10047 function is passed no arguments and its return value is ignored.
10048 When FUNCTION is something returned by `dynamic-func', call that
10049 function and ignore DYNOBJ. When FUNCTION is a string (or symbol,
10050 etc.), look it up in DYNOBJ; this is equivalent to
10052 (dynamic-call (dynamic-func FUNCTION DYNOBJ) #f)
10054 Interrupts are deferred while the C function is executing (with
10055 SCM_DEFER_INTS/SCM_ALLOW_INTS).
10057 (dynamic-args-call FUNCTION DYNOBJ ARGS)
10059 Call the C function indicated by FUNCTION and DYNOBJ, but pass it
10060 some arguments and return its return value. The C function is
10061 expected to take two arguments and return an `int', just like
10064 int c_func (int argc, char **argv);
10066 ARGS must be a list of strings and is converted into an array of
10067 `char *'. The array is passed in ARGV and its size in ARGC. The
10068 return value is converted to a Scheme number and returned from the
10069 call to `dynamic-args-call'.
10071 When dynamic linking is disabled or not supported on your system,
10072 the above functions throw errors, but they are still available.
10074 Here is a small example that works on GNU/Linux:
10076 (define libc-obj (dynamic-link "libc.so"))
10077 (dynamic-args-call 'rand libc-obj '())
10079 See the file `libguile/DYNAMIC-LINKING' for additional comments.
10081 ** The #/ syntax for module names is depreciated, and will be removed
10082 in a future version of Guile. Instead of
10090 The latter syntax is more consistent with existing Lisp practice.
10092 ** Guile now does fancier printing of structures. Structures are the
10093 underlying implementation for records, which in turn are used to
10094 implement modules, so all of these object now print differently and in
10095 a more informative way.
10097 The Scheme printer will examine the builtin variable *struct-printer*
10098 whenever it needs to print a structure object. When this variable is
10099 not `#f' it is deemed to be a procedure and will be applied to the
10100 structure object and the output port. When *struct-printer* is `#f'
10101 or the procedure return `#f' the structure object will be printed in
10102 the boring #<struct 80458270> form.
10104 This hook is used by some routines in ice-9/boot-9.scm to implement
10105 type specific printing routines. Please read the comments there about
10106 "printing structs".
10108 One of the more specific uses of structs are records. The printing
10109 procedure that could be passed to MAKE-RECORD-TYPE is now actually
10110 called. It should behave like a *struct-printer* procedure (described
10113 ** Guile now supports a new R4RS-compliant syntax for keywords. A
10114 token of the form #:NAME, where NAME has the same syntax as a Scheme
10115 symbol, is the external representation of the keyword named NAME.
10116 Keyword objects print using this syntax as well, so values containing
10117 keyword objects can be read back into Guile. When used in an
10118 expression, keywords are self-quoting objects.
10120 Guile suports this read syntax, and uses this print syntax, regardless
10121 of the current setting of the `keyword' read option. The `keyword'
10122 read option only controls whether Guile recognizes the `:NAME' syntax,
10123 which is incompatible with R4RS. (R4RS says such token represent
10126 ** Guile has regular expression support again. Guile 1.0 included
10127 functions for matching regular expressions, based on the Rx library.
10128 In Guile 1.1, the Guile/Rx interface was removed to simplify the
10129 distribution, and thus Guile had no regular expression support. Guile
10130 1.2 again supports the most commonly used functions, and supports all
10131 of SCSH's regular expression functions.
10133 If your system does not include a POSIX regular expression library,
10134 and you have not linked Guile with a third-party regexp library such as
10135 Rx, these functions will not be available. You can tell whether your
10136 Guile installation includes regular expression support by checking
10137 whether the `*features*' list includes the `regex' symbol.
10139 *** regexp functions
10141 By default, Guile supports POSIX extended regular expressions. That
10142 means that the characters `(', `)', `+' and `?' are special, and must
10143 be escaped if you wish to match the literal characters.
10145 This regular expression interface was modeled after that implemented
10146 by SCSH, the Scheme Shell. It is intended to be upwardly compatible
10147 with SCSH regular expressions.
10149 **** Function: string-match PATTERN STR [START]
10150 Compile the string PATTERN into a regular expression and compare
10151 it with STR. The optional numeric argument START specifies the
10152 position of STR at which to begin matching.
10154 `string-match' returns a "match structure" which describes what,
10155 if anything, was matched by the regular expression. *Note Match
10156 Structures::. If STR does not match PATTERN at all,
10157 `string-match' returns `#f'.
10159 Each time `string-match' is called, it must compile its PATTERN
10160 argument into a regular expression structure. This operation is
10161 expensive, which makes `string-match' inefficient if the same regular
10162 expression is used several times (for example, in a loop). For better
10163 performance, you can compile a regular expression in advance and then
10164 match strings against the compiled regexp.
10166 **** Function: make-regexp STR [FLAGS]
10167 Compile the regular expression described by STR, and return the
10168 compiled regexp structure. If STR does not describe a legal
10169 regular expression, `make-regexp' throws a
10170 `regular-expression-syntax' error.
10172 FLAGS may be the bitwise-or of one or more of the following:
10174 **** Constant: regexp/extended
10175 Use POSIX Extended Regular Expression syntax when interpreting
10176 STR. If not set, POSIX Basic Regular Expression syntax is used.
10177 If the FLAGS argument is omitted, we assume regexp/extended.
10179 **** Constant: regexp/icase
10180 Do not differentiate case. Subsequent searches using the
10181 returned regular expression will be case insensitive.
10183 **** Constant: regexp/newline
10184 Match-any-character operators don't match a newline.
10186 A non-matching list ([^...]) not containing a newline matches a
10189 Match-beginning-of-line operator (^) matches the empty string
10190 immediately after a newline, regardless of whether the FLAGS
10191 passed to regexp-exec contain regexp/notbol.
10193 Match-end-of-line operator ($) matches the empty string
10194 immediately before a newline, regardless of whether the FLAGS
10195 passed to regexp-exec contain regexp/noteol.
10197 **** Function: regexp-exec REGEXP STR [START [FLAGS]]
10198 Match the compiled regular expression REGEXP against `str'. If
10199 the optional integer START argument is provided, begin matching
10200 from that position in the string. Return a match structure
10201 describing the results of the match, or `#f' if no match could be
10204 FLAGS may be the bitwise-or of one or more of the following:
10206 **** Constant: regexp/notbol
10207 The match-beginning-of-line operator always fails to match (but
10208 see the compilation flag regexp/newline above) This flag may be
10209 used when different portions of a string are passed to
10210 regexp-exec and the beginning of the string should not be
10211 interpreted as the beginning of the line.
10213 **** Constant: regexp/noteol
10214 The match-end-of-line operator always fails to match (but see the
10215 compilation flag regexp/newline above)
10217 **** Function: regexp? OBJ
10218 Return `#t' if OBJ is a compiled regular expression, or `#f'
10221 Regular expressions are commonly used to find patterns in one string
10222 and replace them with the contents of another string.
10224 **** Function: regexp-substitute PORT MATCH [ITEM...]
10225 Write to the output port PORT selected contents of the match
10226 structure MATCH. Each ITEM specifies what should be written, and
10227 may be one of the following arguments:
10229 * A string. String arguments are written out verbatim.
10231 * An integer. The submatch with that number is written.
10233 * The symbol `pre'. The portion of the matched string preceding
10234 the regexp match is written.
10236 * The symbol `post'. The portion of the matched string
10237 following the regexp match is written.
10239 PORT may be `#f', in which case nothing is written; instead,
10240 `regexp-substitute' constructs a string from the specified ITEMs
10243 **** Function: regexp-substitute/global PORT REGEXP TARGET [ITEM...]
10244 Similar to `regexp-substitute', but can be used to perform global
10245 substitutions on STR. Instead of taking a match structure as an
10246 argument, `regexp-substitute/global' takes two string arguments: a
10247 REGEXP string describing a regular expression, and a TARGET string
10248 which should be matched against this regular expression.
10250 Each ITEM behaves as in REGEXP-SUBSTITUTE, with the following
10253 * A function may be supplied. When this function is called, it
10254 will be passed one argument: a match structure for a given
10255 regular expression match. It should return a string to be
10256 written out to PORT.
10258 * The `post' symbol causes `regexp-substitute/global' to recurse
10259 on the unmatched portion of STR. This *must* be supplied in
10260 order to perform global search-and-replace on STR; if it is
10261 not present among the ITEMs, then `regexp-substitute/global'
10262 will return after processing a single match.
10264 *** Match Structures
10266 A "match structure" is the object returned by `string-match' and
10267 `regexp-exec'. It describes which portion of a string, if any, matched
10268 the given regular expression. Match structures include: a reference to
10269 the string that was checked for matches; the starting and ending
10270 positions of the regexp match; and, if the regexp included any
10271 parenthesized subexpressions, the starting and ending positions of each
10274 In each of the regexp match functions described below, the `match'
10275 argument must be a match structure returned by a previous call to
10276 `string-match' or `regexp-exec'. Most of these functions return some
10277 information about the original target string that was matched against a
10278 regular expression; we will call that string TARGET for easy reference.
10280 **** Function: regexp-match? OBJ
10281 Return `#t' if OBJ is a match structure returned by a previous
10282 call to `regexp-exec', or `#f' otherwise.
10284 **** Function: match:substring MATCH [N]
10285 Return the portion of TARGET matched by subexpression number N.
10286 Submatch 0 (the default) represents the entire regexp match. If
10287 the regular expression as a whole matched, but the subexpression
10288 number N did not match, return `#f'.
10290 **** Function: match:start MATCH [N]
10291 Return the starting position of submatch number N.
10293 **** Function: match:end MATCH [N]
10294 Return the ending position of submatch number N.
10296 **** Function: match:prefix MATCH
10297 Return the unmatched portion of TARGET preceding the regexp match.
10299 **** Function: match:suffix MATCH
10300 Return the unmatched portion of TARGET following the regexp match.
10302 **** Function: match:count MATCH
10303 Return the number of parenthesized subexpressions from MATCH.
10304 Note that the entire regular expression match itself counts as a
10305 subexpression, and failed submatches are included in the count.
10307 **** Function: match:string MATCH
10308 Return the original TARGET string.
10310 *** Backslash Escapes
10312 Sometimes you will want a regexp to match characters like `*' or `$'
10313 exactly. For example, to check whether a particular string represents
10314 a menu entry from an Info node, it would be useful to match it against
10315 a regexp like `^* [^:]*::'. However, this won't work; because the
10316 asterisk is a metacharacter, it won't match the `*' at the beginning of
10317 the string. In this case, we want to make the first asterisk un-magic.
10319 You can do this by preceding the metacharacter with a backslash
10320 character `\'. (This is also called "quoting" the metacharacter, and
10321 is known as a "backslash escape".) When Guile sees a backslash in a
10322 regular expression, it considers the following glyph to be an ordinary
10323 character, no matter what special meaning it would ordinarily have.
10324 Therefore, we can make the above example work by changing the regexp to
10325 `^\* [^:]*::'. The `\*' sequence tells the regular expression engine
10326 to match only a single asterisk in the target string.
10328 Since the backslash is itself a metacharacter, you may force a
10329 regexp to match a backslash in the target string by preceding the
10330 backslash with itself. For example, to find variable references in a
10331 TeX program, you might want to find occurrences of the string `\let\'
10332 followed by any number of alphabetic characters. The regular expression
10333 `\\let\\[A-Za-z]*' would do this: the double backslashes in the regexp
10334 each match a single backslash in the target string.
10336 **** Function: regexp-quote STR
10337 Quote each special character found in STR with a backslash, and
10338 return the resulting string.
10340 *Very important:* Using backslash escapes in Guile source code (as
10341 in Emacs Lisp or C) can be tricky, because the backslash character has
10342 special meaning for the Guile reader. For example, if Guile encounters
10343 the character sequence `\n' in the middle of a string while processing
10344 Scheme code, it replaces those characters with a newline character.
10345 Similarly, the character sequence `\t' is replaced by a horizontal tab.
10346 Several of these "escape sequences" are processed by the Guile reader
10347 before your code is executed. Unrecognized escape sequences are
10348 ignored: if the characters `\*' appear in a string, they will be
10349 translated to the single character `*'.
10351 This translation is obviously undesirable for regular expressions,
10352 since we want to be able to include backslashes in a string in order to
10353 escape regexp metacharacters. Therefore, to make sure that a backslash
10354 is preserved in a string in your Guile program, you must use *two*
10355 consecutive backslashes:
10357 (define Info-menu-entry-pattern (make-regexp "^\\* [^:]*"))
10359 The string in this example is preprocessed by the Guile reader before
10360 any code is executed. The resulting argument to `make-regexp' is the
10361 string `^\* [^:]*', which is what we really want.
10363 This also means that in order to write a regular expression that
10364 matches a single backslash character, the regular expression string in
10365 the source code must include *four* backslashes. Each consecutive pair
10366 of backslashes gets translated by the Guile reader to a single
10367 backslash, and the resulting double-backslash is interpreted by the
10368 regexp engine as matching a single backslash character. Hence:
10370 (define tex-variable-pattern (make-regexp "\\\\let\\\\=[A-Za-z]*"))
10372 The reason for the unwieldiness of this syntax is historical. Both
10373 regular expression pattern matchers and Unix string processing systems
10374 have traditionally used backslashes with the special meanings described
10375 above. The POSIX regular expression specification and ANSI C standard
10376 both require these semantics. Attempting to abandon either convention
10377 would cause other kinds of compatibility problems, possibly more severe
10378 ones. Therefore, without extending the Scheme reader to support
10379 strings with different quoting conventions (an ungainly and confusing
10380 extension when implemented in other languages), we must adhere to this
10381 cumbersome escape syntax.
10383 * Changes to the gh_ interface
10385 * Changes to the scm_ interface
10387 * Changes to system call interfaces:
10389 ** The value returned by `raise' is now unspecified. It throws an exception
10390 if an error occurs.
10392 *** A new procedure `sigaction' can be used to install signal handlers
10394 (sigaction signum [action] [flags])
10396 signum is the signal number, which can be specified using the value
10399 If action is omitted, sigaction returns a pair: the CAR is the current
10400 signal hander, which will be either an integer with the value SIG_DFL
10401 (default action) or SIG_IGN (ignore), or the Scheme procedure which
10402 handles the signal, or #f if a non-Scheme procedure handles the
10403 signal. The CDR contains the current sigaction flags for the handler.
10405 If action is provided, it is installed as the new handler for signum.
10406 action can be a Scheme procedure taking one argument, or the value of
10407 SIG_DFL (default action) or SIG_IGN (ignore), or #f to restore
10408 whatever signal handler was installed before sigaction was first used.
10409 Flags can optionally be specified for the new handler (SA_RESTART is
10410 always used if the system provides it, so need not be specified.) The
10411 return value is a pair with information about the old handler as
10414 This interface does not provide access to the "signal blocking"
10415 facility. Maybe this is not needed, since the thread support may
10416 provide solutions to the problem of consistent access to data
10419 *** A new procedure `flush-all-ports' is equivalent to running
10420 `force-output' on every port open for output.
10422 ** Guile now provides information on how it was built, via the new
10423 global variable, %guile-build-info. This variable records the values
10424 of the standard GNU makefile directory variables as an assocation
10425 list, mapping variable names (symbols) onto directory paths (strings).
10426 For example, to find out where the Guile link libraries were
10427 installed, you can say:
10429 guile -c "(display (assq-ref %guile-build-info 'libdir)) (newline)"
10432 * Changes to the scm_ interface
10434 ** The new function scm_handle_by_message_noexit is just like the
10435 existing scm_handle_by_message function, except that it doesn't call
10436 exit to terminate the process. Instead, it prints a message and just
10437 returns #f. This might be a more appropriate catch-all handler for
10438 new dynamic roots and threads.
10441 Changes in Guile 1.1 (released Friday, May 16 1997):
10443 * Changes to the distribution.
10445 The Guile 1.0 distribution has been split up into several smaller
10447 guile-core --- the Guile interpreter itself.
10448 guile-tcltk --- the interface between the Guile interpreter and
10449 Tcl/Tk; Tcl is an interpreter for a stringy language, and Tk
10450 is a toolkit for building graphical user interfaces.
10451 guile-rgx-ctax --- the interface between Guile and the Rx regular
10452 expression matcher, and the translator for the Ctax
10453 programming language. These are packaged together because the
10454 Ctax translator uses Rx to parse Ctax source code.
10456 This NEWS file describes the changes made to guile-core since the 1.0
10459 We no longer distribute the documentation, since it was either out of
10460 date, or incomplete. As soon as we have current documentation, we
10461 will distribute it.
10465 * Changes to the stand-alone interpreter
10467 ** guile now accepts command-line arguments compatible with SCSH, Olin
10468 Shivers' Scheme Shell.
10470 In general, arguments are evaluated from left to right, but there are
10471 exceptions. The following switches stop argument processing, and
10472 stash all remaining command-line arguments as the value returned by
10473 the (command-line) function.
10474 -s SCRIPT load Scheme source code from FILE, and exit
10475 -c EXPR evalute Scheme expression EXPR, and exit
10476 -- stop scanning arguments; run interactively
10478 The switches below are processed as they are encountered.
10479 -l FILE load Scheme source code from FILE
10480 -e FUNCTION after reading script, apply FUNCTION to
10481 command line arguments
10482 -ds do -s script at this point
10483 --emacs enable Emacs protocol (experimental)
10484 -h, --help display this help and exit
10485 -v, --version display version information and exit
10486 \ read arguments from following script lines
10488 So, for example, here is a Guile script named `ekko' (thanks, Olin)
10489 which re-implements the traditional "echo" command:
10491 #!/usr/local/bin/guile -s
10493 (define (main args)
10494 (map (lambda (arg) (display arg) (display " "))
10498 (main (command-line))
10500 Suppose we invoke this script as follows:
10502 ekko a speckled gecko
10504 Through the magic of Unix script processing (triggered by the `#!'
10505 token at the top of the file), /usr/local/bin/guile receives the
10506 following list of command-line arguments:
10508 ("-s" "./ekko" "a" "speckled" "gecko")
10510 Unix inserts the name of the script after the argument specified on
10511 the first line of the file (in this case, "-s"), and then follows that
10512 with the arguments given to the script. Guile loads the script, which
10513 defines the `main' function, and then applies it to the list of
10514 remaining command-line arguments, ("a" "speckled" "gecko").
10516 In Unix, the first line of a script file must take the following form:
10518 #!INTERPRETER ARGUMENT
10520 where INTERPRETER is the absolute filename of the interpreter
10521 executable, and ARGUMENT is a single command-line argument to pass to
10524 You may only pass one argument to the interpreter, and its length is
10525 limited. These restrictions can be annoying to work around, so Guile
10526 provides a general mechanism (borrowed from, and compatible with,
10527 SCSH) for circumventing them.
10529 If the ARGUMENT in a Guile script is a single backslash character,
10530 `\', Guile will open the script file, parse arguments from its second
10531 and subsequent lines, and replace the `\' with them. So, for example,
10532 here is another implementation of the `ekko' script:
10534 #!/usr/local/bin/guile \
10537 (define (main args)
10538 (for-each (lambda (arg) (display arg) (display " "))
10542 If the user invokes this script as follows:
10544 ekko a speckled gecko
10546 Unix expands this into
10548 /usr/local/bin/guile \ ekko a speckled gecko
10550 When Guile sees the `\' argument, it replaces it with the arguments
10551 read from the second line of the script, producing:
10553 /usr/local/bin/guile -e main -s ekko a speckled gecko
10555 This tells Guile to load the `ekko' script, and apply the function
10556 `main' to the argument list ("a" "speckled" "gecko").
10558 Here is how Guile parses the command-line arguments:
10559 - Each space character terminates an argument. This means that two
10560 spaces in a row introduce an empty-string argument.
10561 - The tab character is not permitted (unless you quote it with the
10562 backslash character, as described below), to avoid confusion.
10563 - The newline character terminates the sequence of arguments, and will
10564 also terminate a final non-empty argument. (However, a newline
10565 following a space will not introduce a final empty-string argument;
10566 it only terminates the argument list.)
10567 - The backslash character is the escape character. It escapes
10568 backslash, space, tab, and newline. The ANSI C escape sequences
10569 like \n and \t are also supported. These produce argument
10570 constituents; the two-character combination \n doesn't act like a
10571 terminating newline. The escape sequence \NNN for exactly three
10572 octal digits reads as the character whose ASCII code is NNN. As
10573 above, characters produced this way are argument constituents.
10574 Backslash followed by other characters is not allowed.
10576 * Changes to the procedure for linking libguile with your programs
10578 ** Guile now builds and installs a shared guile library, if your
10579 system support shared libraries. (It still builds a static library on
10580 all systems.) Guile automatically detects whether your system
10581 supports shared libraries. To prevent Guile from buildisg shared
10582 libraries, pass the `--disable-shared' flag to the configure script.
10584 Guile takes longer to compile when it builds shared libraries, because
10585 it must compile every file twice --- once to produce position-
10586 independent object code, and once to produce normal object code.
10588 ** The libthreads library has been merged into libguile.
10590 To link a program against Guile, you now need only link against
10591 -lguile and -lqt; -lthreads is no longer needed. If you are using
10592 autoconf to generate configuration scripts for your application, the
10593 following lines should suffice to add the appropriate libraries to
10596 ### Find quickthreads and libguile.
10597 AC_CHECK_LIB(qt, main)
10598 AC_CHECK_LIB(guile, scm_shell)
10600 * Changes to Scheme functions
10602 ** Guile Scheme's special syntax for keyword objects is now optional,
10603 and disabled by default.
10605 The syntax variation from R4RS made it difficult to port some
10606 interesting packages to Guile. The routines which accepted keyword
10607 arguments (mostly in the module system) have been modified to also
10608 accept symbols whose names begin with `:'.
10610 To change the keyword syntax, you must first import the (ice-9 debug)
10612 (use-modules (ice-9 debug))
10614 Then you can enable the keyword syntax as follows:
10615 (read-set! keywords 'prefix)
10617 To disable keyword syntax, do this:
10618 (read-set! keywords #f)
10620 ** Many more primitive functions accept shared substrings as
10621 arguments. In the past, these functions required normal, mutable
10622 strings as arguments, although they never made use of this
10625 ** The uniform array functions now operate on byte vectors. These
10626 functions are `array-fill!', `serial-array-copy!', `array-copy!',
10627 `serial-array-map', `array-map', `array-for-each', and
10628 `array-index-map!'.
10630 ** The new functions `trace' and `untrace' implement simple debugging
10631 support for Scheme functions.
10633 The `trace' function accepts any number of procedures as arguments,
10634 and tells the Guile interpreter to display each procedure's name and
10635 arguments each time the procedure is invoked. When invoked with no
10636 arguments, `trace' returns the list of procedures currently being
10639 The `untrace' function accepts any number of procedures as arguments,
10640 and tells the Guile interpreter not to trace them any more. When
10641 invoked with no arguments, `untrace' untraces all curretly traced
10644 The tracing in Guile has an advantage over most other systems: we
10645 don't create new procedure objects, but mark the procedure objects
10646 themselves. This means that anonymous and internal procedures can be
10649 ** The function `assert-repl-prompt' has been renamed to
10650 `set-repl-prompt!'. It takes one argument, PROMPT.
10651 - If PROMPT is #f, the Guile read-eval-print loop will not prompt.
10652 - If PROMPT is a string, we use it as a prompt.
10653 - If PROMPT is a procedure accepting no arguments, we call it, and
10654 display the result as a prompt.
10655 - Otherwise, we display "> ".
10657 ** The new function `eval-string' reads Scheme expressions from a
10658 string and evaluates them, returning the value of the last expression
10659 in the string. If the string contains no expressions, it returns an
10662 ** The new function `thunk?' returns true iff its argument is a
10663 procedure of zero arguments.
10665 ** `defined?' is now a builtin function, instead of syntax. This
10666 means that its argument should be quoted. It returns #t iff its
10667 argument is bound in the current module.
10669 ** The new syntax `use-modules' allows you to add new modules to your
10670 environment without re-typing a complete `define-module' form. It
10671 accepts any number of module names as arguments, and imports their
10672 public bindings into the current module.
10674 ** The new function (module-defined? NAME MODULE) returns true iff
10675 NAME, a symbol, is defined in MODULE, a module object.
10677 ** The new function `builtin-bindings' creates and returns a hash
10678 table containing copies of all the root module's bindings.
10680 ** The new function `builtin-weak-bindings' does the same as
10681 `builtin-bindings', but creates a doubly-weak hash table.
10683 ** The `equal?' function now considers variable objects to be
10684 equivalent if they have the same name and the same value.
10686 ** The new function `command-line' returns the command-line arguments
10687 given to Guile, as a list of strings.
10689 When using guile as a script interpreter, `command-line' returns the
10690 script's arguments; those processed by the interpreter (like `-s' or
10691 `-c') are omitted. (In other words, you get the normal, expected
10692 behavior.) Any application that uses scm_shell to process its
10693 command-line arguments gets this behavior as well.
10695 ** The new function `load-user-init' looks for a file called `.guile'
10696 in the user's home directory, and loads it if it exists. This is
10697 mostly for use by the code generated by scm_compile_shell_switches,
10698 but we thought it might also be useful in other circumstances.
10700 ** The new function `log10' returns the base-10 logarithm of its
10703 ** Changes to I/O functions
10705 *** The functions `read', `primitive-load', `read-and-eval!', and
10706 `primitive-load-path' no longer take optional arguments controlling
10707 case insensitivity and a `#' parser.
10709 Case sensitivity is now controlled by a read option called
10710 `case-insensitive'. The user can add new `#' syntaxes with the
10711 `read-hash-extend' function (see below).
10713 *** The new function `read-hash-extend' allows the user to change the
10714 syntax of Guile Scheme in a somewhat controlled way.
10716 (read-hash-extend CHAR PROC)
10717 When parsing S-expressions, if we read a `#' character followed by
10718 the character CHAR, use PROC to parse an object from the stream.
10719 If PROC is #f, remove any parsing procedure registered for CHAR.
10721 The reader applies PROC to two arguments: CHAR and an input port.
10723 *** The new functions read-delimited and read-delimited! provide a
10724 general mechanism for doing delimited input on streams.
10726 (read-delimited DELIMS [PORT HANDLE-DELIM])
10727 Read until we encounter one of the characters in DELIMS (a string),
10728 or end-of-file. PORT is the input port to read from; it defaults to
10729 the current input port. The HANDLE-DELIM parameter determines how
10730 the terminating character is handled; it should be one of the
10733 'trim omit delimiter from result
10734 'peek leave delimiter character in input stream
10735 'concat append delimiter character to returned value
10736 'split return a pair: (RESULT . TERMINATOR)
10738 HANDLE-DELIM defaults to 'peek.
10740 (read-delimited! DELIMS BUF [PORT HANDLE-DELIM START END])
10741 A side-effecting variant of `read-delimited'.
10743 The data is written into the string BUF at the indices in the
10744 half-open interval [START, END); the default interval is the whole
10745 string: START = 0 and END = (string-length BUF). The values of
10746 START and END must specify a well-defined interval in BUF, i.e.
10747 0 <= START <= END <= (string-length BUF).
10749 It returns NBYTES, the number of bytes read. If the buffer filled
10750 up without a delimiter character being found, it returns #f. If the
10751 port is at EOF when the read starts, it returns the EOF object.
10753 If an integer is returned (i.e., the read is successfully terminated
10754 by reading a delimiter character), then the HANDLE-DELIM parameter
10755 determines how to handle the terminating character. It is described
10756 above, and defaults to 'peek.
10758 (The descriptions of these functions were borrowed from the SCSH
10759 manual, by Olin Shivers and Brian Carlstrom.)
10761 *** The `%read-delimited!' function is the primitive used to implement
10762 `read-delimited' and `read-delimited!'.
10764 (%read-delimited! DELIMS BUF GOBBLE? [PORT START END])
10766 This returns a pair of values: (TERMINATOR . NUM-READ).
10767 - TERMINATOR describes why the read was terminated. If it is a
10768 character or the eof object, then that is the value that terminated
10769 the read. If it is #f, the function filled the buffer without finding
10770 a delimiting character.
10771 - NUM-READ is the number of characters read into BUF.
10773 If the read is successfully terminated by reading a delimiter
10774 character, then the gobble? parameter determines what to do with the
10775 terminating character. If true, the character is removed from the
10776 input stream; if false, the character is left in the input stream
10777 where a subsequent read operation will retrieve it. In either case,
10778 the character is also the first value returned by the procedure call.
10780 (The descriptions of this function was borrowed from the SCSH manual,
10781 by Olin Shivers and Brian Carlstrom.)
10783 *** The `read-line' and `read-line!' functions have changed; they now
10784 trim the terminator by default; previously they appended it to the
10785 returned string. For the old behavior, use (read-line PORT 'concat).
10787 *** The functions `uniform-array-read!' and `uniform-array-write!' now
10788 take new optional START and END arguments, specifying the region of
10789 the array to read and write.
10791 *** The `ungetc-char-ready?' function has been removed. We feel it's
10792 inappropriate for an interface to expose implementation details this
10795 ** Changes to the Unix library and system call interface
10797 *** The new fcntl function provides access to the Unix `fcntl' system
10800 (fcntl PORT COMMAND VALUE)
10801 Apply COMMAND to PORT's file descriptor, with VALUE as an argument.
10802 Values for COMMAND are:
10804 F_DUPFD duplicate a file descriptor
10805 F_GETFD read the descriptor's close-on-exec flag
10806 F_SETFD set the descriptor's close-on-exec flag to VALUE
10807 F_GETFL read the descriptor's flags, as set on open
10808 F_SETFL set the descriptor's flags, as set on open to VALUE
10809 F_GETOWN return the process ID of a socket's owner, for SIGIO
10810 F_SETOWN set the process that owns a socket to VALUE, for SIGIO
10811 FD_CLOEXEC not sure what this is
10813 For details, see the documentation for the fcntl system call.
10815 *** The arguments to `select' have changed, for compatibility with
10816 SCSH. The TIMEOUT parameter may now be non-integral, yielding the
10817 expected behavior. The MILLISECONDS parameter has been changed to
10818 MICROSECONDS, to more closely resemble the underlying system call.
10819 The RVEC, WVEC, and EVEC arguments can now be vectors; the type of the
10820 corresponding return set will be the same.
10822 *** The arguments to the `mknod' system call have changed. They are
10825 (mknod PATH TYPE PERMS DEV)
10826 Create a new file (`node') in the file system. PATH is the name of
10827 the file to create. TYPE is the kind of file to create; it should
10828 be 'fifo, 'block-special, or 'char-special. PERMS specifies the
10829 permission bits to give the newly created file. If TYPE is
10830 'block-special or 'char-special, DEV specifies which device the
10831 special file refers to; its interpretation depends on the kind of
10832 special file being created.
10834 *** The `fork' function has been renamed to `primitive-fork', to avoid
10835 clashing with various SCSH forks.
10837 *** The `recv' and `recvfrom' functions have been renamed to `recv!'
10838 and `recvfrom!'. They no longer accept a size for a second argument;
10839 you must pass a string to hold the received value. They no longer
10840 return the buffer. Instead, `recv' returns the length of the message
10841 received, and `recvfrom' returns a pair containing the packet's length
10842 and originating address.
10844 *** The file descriptor datatype has been removed, as have the
10845 `read-fd', `write-fd', `close', `lseek', and `dup' functions.
10846 We plan to replace these functions with a SCSH-compatible interface.
10848 *** The `create' function has been removed; it's just a special case
10851 *** There are new functions to break down process termination status
10852 values. In the descriptions below, STATUS is a value returned by
10855 (status:exit-val STATUS)
10856 If the child process exited normally, this function returns the exit
10857 code for the child process (i.e., the value passed to exit, or
10858 returned from main). If the child process did not exit normally,
10859 this function returns #f.
10861 (status:stop-sig STATUS)
10862 If the child process was suspended by a signal, this function
10863 returns the signal that suspended the child. Otherwise, it returns
10866 (status:term-sig STATUS)
10867 If the child process terminated abnormally, this function returns
10868 the signal that terminated the child. Otherwise, this function
10871 POSIX promises that exactly one of these functions will return true on
10872 a valid STATUS value.
10874 These functions are compatible with SCSH.
10876 *** There are new accessors and setters for the broken-out time vectors
10877 returned by `localtime', `gmtime', and that ilk. They are:
10879 Component Accessor Setter
10880 ========================= ============ ============
10881 seconds tm:sec set-tm:sec
10882 minutes tm:min set-tm:min
10883 hours tm:hour set-tm:hour
10884 day of the month tm:mday set-tm:mday
10885 month tm:mon set-tm:mon
10886 year tm:year set-tm:year
10887 day of the week tm:wday set-tm:wday
10888 day in the year tm:yday set-tm:yday
10889 daylight saving time tm:isdst set-tm:isdst
10890 GMT offset, seconds tm:gmtoff set-tm:gmtoff
10891 name of time zone tm:zone set-tm:zone
10893 *** There are new accessors for the vectors returned by `uname',
10894 describing the host system:
10897 ============================================== ================
10898 name of the operating system implementation utsname:sysname
10899 network name of this machine utsname:nodename
10900 release level of the operating system utsname:release
10901 version level of the operating system utsname:version
10902 machine hardware platform utsname:machine
10904 *** There are new accessors for the vectors returned by `getpw',
10905 `getpwnam', `getpwuid', and `getpwent', describing entries from the
10906 system's user database:
10909 ====================== =================
10910 user name passwd:name
10911 user password passwd:passwd
10913 group id passwd:gid
10914 real name passwd:gecos
10915 home directory passwd:dir
10916 shell program passwd:shell
10918 *** There are new accessors for the vectors returned by `getgr',
10919 `getgrnam', `getgrgid', and `getgrent', describing entries from the
10920 system's group database:
10923 ======================= ============
10924 group name group:name
10925 group password group:passwd
10927 group members group:mem
10929 *** There are new accessors for the vectors returned by `gethost',
10930 `gethostbyaddr', `gethostbyname', and `gethostent', describing
10934 ========================= ===============
10935 official name of host hostent:name
10936 alias list hostent:aliases
10937 host address type hostent:addrtype
10938 length of address hostent:length
10939 list of addresses hostent:addr-list
10941 *** There are new accessors for the vectors returned by `getnet',
10942 `getnetbyaddr', `getnetbyname', and `getnetent', describing internet
10946 ========================= ===============
10947 official name of net netent:name
10948 alias list netent:aliases
10949 net number type netent:addrtype
10950 net number netent:net
10952 *** There are new accessors for the vectors returned by `getproto',
10953 `getprotobyname', `getprotobynumber', and `getprotoent', describing
10954 internet protocols:
10957 ========================= ===============
10958 official protocol name protoent:name
10959 alias list protoent:aliases
10960 protocol number protoent:proto
10962 *** There are new accessors for the vectors returned by `getserv',
10963 `getservbyname', `getservbyport', and `getservent', describing
10964 internet protocols:
10967 ========================= ===============
10968 official service name servent:name
10969 alias list servent:aliases
10970 port number servent:port
10971 protocol to use servent:proto
10973 *** There are new accessors for the sockaddr structures returned by
10974 `accept', `getsockname', `getpeername', `recvfrom!':
10977 ======================================== ===============
10978 address format (`family') sockaddr:fam
10979 path, for file domain addresses sockaddr:path
10980 address, for internet domain addresses sockaddr:addr
10981 TCP or UDP port, for internet sockaddr:port
10983 *** The `getpwent', `getgrent', `gethostent', `getnetent',
10984 `getprotoent', and `getservent' functions now return #f at the end of
10985 the user database. (They used to throw an exception.)
10987 Note that calling MUMBLEent function is equivalent to calling the
10988 corresponding MUMBLE function with no arguments.
10990 *** The `setpwent', `setgrent', `sethostent', `setnetent',
10991 `setprotoent', and `setservent' routines now take no arguments.
10993 *** The `gethost', `getproto', `getnet', and `getserv' functions now
10994 provide more useful information when they throw an exception.
10996 *** The `lnaof' function has been renamed to `inet-lnaof'.
10998 *** Guile now claims to have the `current-time' feature.
11000 *** The `mktime' function now takes an optional second argument ZONE,
11001 giving the time zone to use for the conversion. ZONE should be a
11002 string, in the same format as expected for the "TZ" environment variable.
11004 *** The `strptime' function now returns a pair (TIME . COUNT), where
11005 TIME is the parsed time as a vector, and COUNT is the number of
11006 characters from the string left unparsed. This function used to
11007 return the remaining characters as a string.
11009 *** The `gettimeofday' function has replaced the old `time+ticks' function.
11010 The return value is now (SECONDS . MICROSECONDS); the fractional
11011 component is no longer expressed in "ticks".
11013 *** The `ticks/sec' constant has been removed, in light of the above change.
11015 * Changes to the gh_ interface
11017 ** gh_eval_str() now returns an SCM object which is the result of the
11020 ** gh_scm2str() now copies the Scheme data to a caller-provided C
11023 ** gh_scm2newstr() now makes a C array, copies the Scheme data to it,
11024 and returns the array
11026 ** gh_scm2str0() is gone: there is no need to distinguish
11027 null-terminated from non-null-terminated, since gh_scm2newstr() allows
11028 the user to interpret the data both ways.
11030 * Changes to the scm_ interface
11032 ** The new function scm_symbol_value0 provides an easy way to get a
11033 symbol's value from C code:
11035 SCM scm_symbol_value0 (char *NAME)
11036 Return the value of the symbol named by the null-terminated string
11037 NAME in the current module. If the symbol named NAME is unbound in
11038 the current module, return SCM_UNDEFINED.
11040 ** The new function scm_sysintern0 creates new top-level variables,
11041 without assigning them a value.
11043 SCM scm_sysintern0 (char *NAME)
11044 Create a new Scheme top-level variable named NAME. NAME is a
11045 null-terminated string. Return the variable's value cell.
11047 ** The function scm_internal_catch is the guts of catch. It handles
11048 all the mechanics of setting up a catch target, invoking the catch
11049 body, and perhaps invoking the handler if the body does a throw.
11051 The function is designed to be usable from C code, but is general
11052 enough to implement all the semantics Guile Scheme expects from throw.
11054 TAG is the catch tag. Typically, this is a symbol, but this function
11055 doesn't actually care about that.
11057 BODY is a pointer to a C function which runs the body of the catch;
11058 this is the code you can throw from. We call it like this:
11059 BODY (BODY_DATA, JMPBUF)
11061 BODY_DATA is just the BODY_DATA argument we received; we pass it
11062 through to BODY as its first argument. The caller can make
11063 BODY_DATA point to anything useful that BODY might need.
11064 JMPBUF is the Scheme jmpbuf object corresponding to this catch,
11065 which we have just created and initialized.
11067 HANDLER is a pointer to a C function to deal with a throw to TAG,
11068 should one occur. We call it like this:
11069 HANDLER (HANDLER_DATA, THROWN_TAG, THROW_ARGS)
11071 HANDLER_DATA is the HANDLER_DATA argument we recevied; it's the
11072 same idea as BODY_DATA above.
11073 THROWN_TAG is the tag that the user threw to; usually this is
11074 TAG, but it could be something else if TAG was #t (i.e., a
11075 catch-all), or the user threw to a jmpbuf.
11076 THROW_ARGS is the list of arguments the user passed to the THROW
11079 BODY_DATA is just a pointer we pass through to BODY. HANDLER_DATA
11080 is just a pointer we pass through to HANDLER. We don't actually
11081 use either of those pointers otherwise ourselves. The idea is
11082 that, if our caller wants to communicate something to BODY or
11083 HANDLER, it can pass a pointer to it as MUMBLE_DATA, which BODY and
11084 HANDLER can then use. Think of it as a way to make BODY and
11085 HANDLER closures, not just functions; MUMBLE_DATA points to the
11086 enclosed variables.
11088 Of course, it's up to the caller to make sure that any data a
11089 MUMBLE_DATA needs is protected from GC. A common way to do this is
11090 to make MUMBLE_DATA a pointer to data stored in an automatic
11091 structure variable; since the collector must scan the stack for
11092 references anyway, this assures that any references in MUMBLE_DATA
11095 ** The new function scm_internal_lazy_catch is exactly like
11096 scm_internal_catch, except:
11098 - It does not unwind the stack (this is the major difference).
11099 - If handler returns, its value is returned from the throw.
11100 - BODY always receives #f as its JMPBUF argument (since there's no
11101 jmpbuf associated with a lazy catch, because we don't unwind the
11104 ** scm_body_thunk is a new body function you can pass to
11105 scm_internal_catch if you want the body to be like Scheme's `catch'
11106 --- a thunk, or a function of one argument if the tag is #f.
11108 BODY_DATA is a pointer to a scm_body_thunk_data structure, which
11109 contains the Scheme procedure to invoke as the body, and the tag
11110 we're catching. If the tag is #f, then we pass JMPBUF (created by
11111 scm_internal_catch) to the body procedure; otherwise, the body gets
11114 ** scm_handle_by_proc is a new handler function you can pass to
11115 scm_internal_catch if you want the handler to act like Scheme's catch
11116 --- call a procedure with the tag and the throw arguments.
11118 If the user does a throw to this catch, this function runs a handler
11119 procedure written in Scheme. HANDLER_DATA is a pointer to an SCM
11120 variable holding the Scheme procedure object to invoke. It ought to
11121 be a pointer to an automatic variable (i.e., one living on the stack),
11122 or the procedure object should be otherwise protected from GC.
11124 ** scm_handle_by_message is a new handler function to use with
11125 `scm_internal_catch' if you want Guile to print a message and die.
11126 It's useful for dealing with throws to uncaught keys at the top level.
11128 HANDLER_DATA, if non-zero, is assumed to be a char * pointing to a
11129 message header to print; if zero, we use "guile" instead. That
11130 text is followed by a colon, then the message described by ARGS.
11132 ** The return type of scm_boot_guile is now void; the function does
11133 not return a value, and indeed, never returns at all.
11135 ** The new function scm_shell makes it easy for user applications to
11136 process command-line arguments in a way that is compatible with the
11137 stand-alone guile interpreter (which is in turn compatible with SCSH,
11140 To use the scm_shell function, first initialize any guile modules
11141 linked into your application, and then call scm_shell with the values
11142 of ARGC and ARGV your `main' function received. scm_shell will add
11143 any SCSH-style meta-arguments from the top of the script file to the
11144 argument vector, and then process the command-line arguments. This
11145 generally means loading a script file or starting up an interactive
11146 command interpreter. For details, see "Changes to the stand-alone
11147 interpreter" above.
11149 ** The new functions scm_get_meta_args and scm_count_argv help you
11150 implement the SCSH-style meta-argument, `\'.
11152 char **scm_get_meta_args (int ARGC, char **ARGV)
11153 If the second element of ARGV is a string consisting of a single
11154 backslash character (i.e. "\\" in Scheme notation), open the file
11155 named by the following argument, parse arguments from it, and return
11156 the spliced command line. The returned array is terminated by a
11159 For details of argument parsing, see above, under "guile now accepts
11160 command-line arguments compatible with SCSH..."
11162 int scm_count_argv (char **ARGV)
11163 Count the arguments in ARGV, assuming it is terminated by a null
11166 For an example of how these functions might be used, see the source
11167 code for the function scm_shell in libguile/script.c.
11169 You will usually want to use scm_shell instead of calling this
11172 ** The new function scm_compile_shell_switches turns an array of
11173 command-line arguments into Scheme code to carry out the actions they
11174 describe. Given ARGC and ARGV, it returns a Scheme expression to
11175 evaluate, and calls scm_set_program_arguments to make any remaining
11176 command-line arguments available to the Scheme code. For example,
11177 given the following arguments:
11179 -e main -s ekko a speckled gecko
11181 scm_set_program_arguments will return the following expression:
11183 (begin (load "ekko") (main (command-line)) (quit))
11185 You will usually want to use scm_shell instead of calling this
11188 ** The function scm_shell_usage prints a usage message appropriate for
11189 an interpreter that uses scm_compile_shell_switches to handle its
11190 command-line arguments.
11192 void scm_shell_usage (int FATAL, char *MESSAGE)
11193 Print a usage message to the standard error output. If MESSAGE is
11194 non-zero, write it before the usage message, followed by a newline.
11195 If FATAL is non-zero, exit the process, using FATAL as the
11196 termination status. (If you want to be compatible with Guile,
11197 always use 1 as the exit status when terminating due to command-line
11200 You will usually want to use scm_shell instead of calling this
11203 ** scm_eval_0str now returns SCM_UNSPECIFIED if the string contains no
11204 expressions. It used to return SCM_EOL. Earth-shattering.
11206 ** The macros for declaring scheme objects in C code have been
11207 rearranged slightly. They are now:
11209 SCM_SYMBOL (C_NAME, SCHEME_NAME)
11210 Declare a static SCM variable named C_NAME, and initialize it to
11211 point to the Scheme symbol whose name is SCHEME_NAME. C_NAME should
11212 be a C identifier, and SCHEME_NAME should be a C string.
11214 SCM_GLOBAL_SYMBOL (C_NAME, SCHEME_NAME)
11215 Just like SCM_SYMBOL, but make C_NAME globally visible.
11217 SCM_VCELL (C_NAME, SCHEME_NAME)
11218 Create a global variable at the Scheme level named SCHEME_NAME.
11219 Declare a static SCM variable named C_NAME, and initialize it to
11220 point to the Scheme variable's value cell.
11222 SCM_GLOBAL_VCELL (C_NAME, SCHEME_NAME)
11223 Just like SCM_VCELL, but make C_NAME globally visible.
11225 The `guile-snarf' script writes initialization code for these macros
11226 to its standard output, given C source code as input.
11228 The SCM_GLOBAL macro is gone.
11230 ** The scm_read_line and scm_read_line_x functions have been replaced
11231 by Scheme code based on the %read-delimited! procedure (known to C
11232 code as scm_read_delimited_x). See its description above for more
11235 ** The function scm_sys_open has been renamed to scm_open. It now
11236 returns a port instead of an FD object.
11238 * The dynamic linking support has changed. For more information, see
11239 libguile/DYNAMIC-LINKING.
11244 User-visible changes from Thursday, September 5, 1996 until Guile 1.0
11247 * Changes to the 'guile' program:
11249 ** Guile now loads some new files when it starts up. Guile first
11250 searches the load path for init.scm, and loads it if found. Then, if
11251 Guile is not being used to execute a script, and the user's home
11252 directory contains a file named `.guile', Guile loads that.
11254 ** You can now use Guile as a shell script interpreter.
11256 To paraphrase the SCSH manual:
11258 When Unix tries to execute an executable file whose first two
11259 characters are the `#!', it treats the file not as machine code to
11260 be directly executed by the native processor, but as source code
11261 to be executed by some interpreter. The interpreter to use is
11262 specified immediately after the #! sequence on the first line of
11263 the source file. The kernel reads in the name of the interpreter,
11264 and executes that instead. It passes the interpreter the source
11265 filename as its first argument, with the original arguments
11266 following. Consult the Unix man page for the `exec' system call
11267 for more information.
11269 Now you can use Guile as an interpreter, using a mechanism which is a
11270 compatible subset of that provided by SCSH.
11272 Guile now recognizes a '-s' command line switch, whose argument is the
11273 name of a file of Scheme code to load. It also treats the two
11274 characters `#!' as the start of a comment, terminated by `!#'. Thus,
11275 to make a file of Scheme code directly executable by Unix, insert the
11276 following two lines at the top of the file:
11278 #!/usr/local/bin/guile -s
11281 Guile treats the argument of the `-s' command-line switch as the name
11282 of a file of Scheme code to load, and treats the sequence `#!' as the
11283 start of a block comment, terminated by `!#'.
11285 For example, here's a version of 'echo' written in Scheme:
11287 #!/usr/local/bin/guile -s
11289 (let loop ((args (cdr (program-arguments))))
11292 (display (car args))
11293 (if (pair? (cdr args))
11295 (loop (cdr args)))))
11298 Why does `#!' start a block comment terminated by `!#', instead of the
11299 end of the line? That is the notation SCSH uses, and although we
11300 don't yet support the other SCSH features that motivate that choice,
11301 we would like to be backward-compatible with any existing Guile
11302 scripts once we do. Furthermore, if the path to Guile on your system
11303 is too long for your kernel, you can start the script with this
11307 exec /really/long/path/to/guile -s "$0" ${1+"$@"}
11310 Note that some very old Unix systems don't support the `#!' syntax.
11313 ** You can now run Guile without installing it.
11315 Previous versions of the interactive Guile interpreter (`guile')
11316 couldn't start up unless Guile's Scheme library had been installed;
11317 they used the value of the environment variable `SCHEME_LOAD_PATH'
11318 later on in the startup process, but not to find the startup code
11319 itself. Now Guile uses `SCHEME_LOAD_PATH' in all searches for Scheme
11322 To run Guile without installing it, build it in the normal way, and
11323 then set the environment variable `SCHEME_LOAD_PATH' to a
11324 colon-separated list of directories, including the top-level directory
11325 of the Guile sources. For example, if you unpacked Guile so that the
11326 full filename of this NEWS file is /home/jimb/guile-1.0b3/NEWS, then
11329 export SCHEME_LOAD_PATH=/home/jimb/my-scheme:/home/jimb/guile-1.0b3
11332 ** Guile's read-eval-print loop no longer prints #<unspecified>
11333 results. If the user wants to see this, she can evaluate the
11334 expression (assert-repl-print-unspecified #t), perhaps in her startup
11337 ** Guile no longer shows backtraces by default when an error occurs;
11338 however, it does display a message saying how to get one, and how to
11339 request that they be displayed by default. After an error, evaluate
11341 to see a backtrace, and
11342 (debug-enable 'backtrace)
11343 to see them by default.
11347 * Changes to Guile Scheme:
11349 ** Guile now distinguishes between #f and the empty list.
11351 This is for compatibility with the IEEE standard, the (possibly)
11352 upcoming Revised^5 Report on Scheme, and many extant Scheme
11355 Guile used to have #f and '() denote the same object, to make Scheme's
11356 type system more compatible with Emacs Lisp's. However, the change
11357 caused too much trouble for Scheme programmers, and we found another
11358 way to reconcile Emacs Lisp with Scheme that didn't require this.
11361 ** Guile's delq, delv, delete functions, and their destructive
11362 counterparts, delq!, delv!, and delete!, now remove all matching
11363 elements from the list, not just the first. This matches the behavior
11364 of the corresponding Emacs Lisp functions, and (I believe) the Maclisp
11365 functions which inspired them.
11367 I recognize that this change may break code in subtle ways, but it
11368 seems best to make the change before the FSF's first Guile release,
11372 ** The compiled-library-path function has been deleted from libguile.
11374 ** The facilities for loading Scheme source files have changed.
11376 *** The variable %load-path now tells Guile which directories to search
11377 for Scheme code. Its value is a list of strings, each of which names
11380 *** The variable %load-extensions now tells Guile which extensions to
11381 try appending to a filename when searching the load path. Its value
11382 is a list of strings. Its default value is ("" ".scm").
11384 *** (%search-load-path FILENAME) searches the directories listed in the
11385 value of the %load-path variable for a Scheme file named FILENAME,
11386 with all the extensions listed in %load-extensions. If it finds a
11387 match, then it returns its full filename. If FILENAME is absolute, it
11388 returns it unchanged. Otherwise, it returns #f.
11390 %search-load-path will not return matches that refer to directories.
11392 *** (primitive-load FILENAME :optional CASE-INSENSITIVE-P SHARP)
11393 uses %seach-load-path to find a file named FILENAME, and loads it if
11394 it finds it. If it can't read FILENAME for any reason, it throws an
11397 The arguments CASE-INSENSITIVE-P and SHARP are interpreted as by the
11400 *** load uses the same searching semantics as primitive-load.
11402 *** The functions %try-load, try-load-with-path, %load, load-with-path,
11403 basic-try-load-with-path, basic-load-with-path, try-load-module-with-
11404 path, and load-module-with-path have been deleted. The functions
11405 above should serve their purposes.
11407 *** If the value of the variable %load-hook is a procedure,
11408 `primitive-load' applies its value to the name of the file being
11409 loaded (without the load path directory name prepended). If its value
11410 is #f, it is ignored. Otherwise, an error occurs.
11412 This is mostly useful for printing load notification messages.
11415 ** The function `eval!' is no longer accessible from the scheme level.
11416 We can't allow operations which introduce glocs into the scheme level,
11417 because Guile's type system can't handle these as data. Use `eval' or
11418 `read-and-eval!' (see below) as replacement.
11420 ** The new function read-and-eval! reads an expression from PORT,
11421 evaluates it, and returns the result. This is more efficient than
11422 simply calling `read' and `eval', since it is not necessary to make a
11423 copy of the expression for the evaluator to munge.
11425 Its optional arguments CASE_INSENSITIVE_P and SHARP are interpreted as
11426 for the `read' function.
11429 ** The function `int?' has been removed; its definition was identical
11430 to that of `integer?'.
11432 ** The functions `<?', `<?', `<=?', `=?', `>?', and `>=?'. Code should
11433 use the R4RS names for these functions.
11435 ** The function object-properties no longer returns the hash handle;
11436 it simply returns the object's property list.
11438 ** Many functions have been changed to throw errors, instead of
11439 returning #f on failure. The point of providing exception handling in
11440 the language is to simplify the logic of user code, but this is less
11441 useful if Guile's primitives don't throw exceptions.
11443 ** The function `fileno' has been renamed from `%fileno'.
11445 ** The function primitive-mode->fdes returns #t or #f now, not 1 or 0.
11448 * Changes to Guile's C interface:
11450 ** The library's initialization procedure has been simplified.
11451 scm_boot_guile now has the prototype:
11453 void scm_boot_guile (int ARGC,
11455 void (*main_func) (),
11458 scm_boot_guile calls MAIN_FUNC, passing it CLOSURE, ARGC, and ARGV.
11459 MAIN_FUNC should do all the work of the program (initializing other
11460 packages, reading user input, etc.) before returning. When MAIN_FUNC
11461 returns, call exit (0); this function never returns. If you want some
11462 other exit value, MAIN_FUNC may call exit itself.
11464 scm_boot_guile arranges for program-arguments to return the strings
11465 given by ARGC and ARGV. If MAIN_FUNC modifies ARGC/ARGV, should call
11466 scm_set_program_arguments with the final list, so Scheme code will
11467 know which arguments have been processed.
11469 scm_boot_guile establishes a catch-all catch handler which prints an
11470 error message and exits the process. This means that Guile exits in a
11471 coherent way when system errors occur and the user isn't prepared to
11472 handle it. If the user doesn't like this behavior, they can establish
11473 their own universal catcher in MAIN_FUNC to shadow this one.
11475 Why must the caller do all the real work from MAIN_FUNC? The garbage
11476 collector assumes that all local variables of type SCM will be above
11477 scm_boot_guile's stack frame on the stack. If you try to manipulate
11478 SCM values after this function returns, it's the luck of the draw
11479 whether the GC will be able to find the objects you allocate. So,
11480 scm_boot_guile function exits, rather than returning, to discourage
11481 people from making that mistake.
11483 The IN, OUT, and ERR arguments were removed; there are other
11484 convenient ways to override these when desired.
11486 The RESULT argument was deleted; this function should never return.
11488 The BOOT_CMD argument was deleted; the MAIN_FUNC argument is more
11492 ** Guile's header files should no longer conflict with your system's
11495 In order to compile code which #included <libguile.h>, previous
11496 versions of Guile required you to add a directory containing all the
11497 Guile header files to your #include path. This was a problem, since
11498 Guile's header files have names which conflict with many systems'
11501 Now only <libguile.h> need appear in your #include path; you must
11502 refer to all Guile's other header files as <libguile/mumble.h>.
11503 Guile's installation procedure puts libguile.h in $(includedir), and
11504 the rest in $(includedir)/libguile.
11507 ** Two new C functions, scm_protect_object and scm_unprotect_object,
11508 have been added to the Guile library.
11510 scm_protect_object (OBJ) protects OBJ from the garbage collector.
11511 OBJ will not be freed, even if all other references are dropped,
11512 until someone does scm_unprotect_object (OBJ). Both functions
11515 Note that calls to scm_protect_object do not nest. You can call
11516 scm_protect_object any number of times on a given object, and the
11517 next call to scm_unprotect_object will unprotect it completely.
11519 Basically, scm_protect_object and scm_unprotect_object just
11520 maintain a list of references to things. Since the GC knows about
11521 this list, all objects it mentions stay alive. scm_protect_object
11522 adds its argument to the list; scm_unprotect_object remove its
11523 argument from the list.
11526 ** scm_eval_0str now returns the value of the last expression
11529 ** The new function scm_read_0str reads an s-expression from a
11530 null-terminated string, and returns it.
11532 ** The new function `scm_stdio_to_port' converts a STDIO file pointer
11533 to a Scheme port object.
11535 ** The new function `scm_set_program_arguments' allows C code to set
11536 the value returned by the Scheme `program-arguments' function.
11541 * Guile no longer includes sophisticated Tcl/Tk support.
11543 The old Tcl/Tk support was unsatisfying to us, because it required the
11544 user to link against the Tcl library, as well as Tk and Guile. The
11545 interface was also un-lispy, in that it preserved Tcl/Tk's practice of
11546 referring to widgets by names, rather than exporting widgets to Scheme
11547 code as a special datatype.
11549 In the Usenix Tk Developer's Workshop held in July 1996, the Tcl/Tk
11550 maintainers described some very interesting changes in progress to the
11551 Tcl/Tk internals, which would facilitate clean interfaces between lone
11552 Tk and other interpreters --- even for garbage-collected languages
11553 like Scheme. They expected the new Tk to be publicly available in the
11556 Since it seems that Guile might soon have a new, cleaner interface to
11557 lone Tk, and that the old Guile/Tk glue code would probably need to be
11558 completely rewritten, we (Jim Blandy and Richard Stallman) have
11559 decided not to support the old code. We'll spend the time instead on
11560 a good interface to the newer Tk, as soon as it is available.
11562 Until then, gtcltk-lib provides trivial, low-maintenance functionality.
11565 Copyright information:
11567 Copyright (C) 1996, 1997, 1998, 1999, 2000, 2001, 2006 Free Software Foundation, Inc.
11569 Permission is granted to anyone to make or distribute verbatim copies
11570 of this document as received, in any medium, provided that the
11571 copyright notice and this permission notice are preserved,
11572 thus giving the recipient permission to redistribute in turn.
11574 Permission is granted to distribute modified versions
11575 of this document, or of portions of it,
11576 under the above conditions, provided also that they
11577 carry prominent notices stating who last changed them.
11582 paragraph-separate: "[
\f]*$"