Andy Wingo [Wed, 23 May 2012 10:00:23 +0000 (12:00 +0200)]
deprecate lookup closures
* libguile/deprecated.h (SCM_TOP_LEVEL_LOOKUP_CLOSURE):
* libguile/deprecated.c (scm_lookup_closure_module):
(scm_module_lookup_closure):
(scm_current_module_lookup_closure): Deprecate this part of the eval
closure interface. It was unused internally, after the scm_sym2var
refactor.
* libguile/eval.h:
* libguile/modules.c:
* libguile/modules.h: Remove deprecated code.
* libguile/goops.c (scm_ensure_accessor): Use scm_module_variable
instead of calling the lookup closure. However I'm not sure that this
code is used at all.
Andy Wingo [Wed, 23 May 2012 09:46:30 +0000 (11:46 +0200)]
deprecate scm_sym2var
* libguile/deprecated.h:
* libguile/deprecated.c (scm_sym2var): Deprecate this function.
* libguile/modules.h:
* libguile/modules.c (scm_module_ensure_local_variable): New public
function, replacing scm_sym2var with a true definep, without going
through eval closures (which are deprecated).
(scm_current_module): Rework to do something sensible before modules
are booted.
(scm_module_lookup, scm_lookup): Refactor to use scm_module_variable.
(scm_module_define, scm_define): Refactor to use
scm_module_ensure_local_variable.
* libguile/vm-i-system.c (define!): Use scm_define.
* libguile/vm.c (resolve_variable): Use scm_module_lookup.
* libguile/macros.c (scm_make_syntax_transformer): Use
scm_module_variable.
* libguile/gdbint.c (gdb_binding): Use scm_define.
* doc/ref/api-modules.texi (Accessing Modules from C): Add docs for
scm_module_ensure_local_variable.
Andy Wingo [Mon, 21 May 2012 16:06:34 +0000 (18:06 +0200)]
fix arity check for applicable structs
* module/language/tree-il/analyze.scm (validate-arity): Fix for
applicable structs. Applicable structs are procedures, but not every
struct has a first slot, and not every struct with a procedure in its
first slot is applicable. Besides, the approach in this patch gives
better errors.
Andy Wingo [Mon, 21 May 2012 15:48:45 +0000 (17:48 +0200)]
rearrange scheme compilation order
* module/Makefile.am (SOURCES): Move vlists, srfi-1, peval, and cse up
in the compilation order. These are exercised a lot in the compiler.
Andy Wingo [Mon, 21 May 2012 15:31:58 +0000 (17:31 +0200)]
vm.c: remove a useless check
* libguile/vm.c: Remove a check for modules being booted. If the module
passed in is true, that is guaranteed to have been the case. Passes a
fresh bootstrap.
Ludovic Courtès [Tue, 15 May 2012 17:12:39 +0000 (19:12 +0200)]
Add more `%file-port-name-canonicalization' tests.
* test-suite/tests/ports.test ("%file-port-name-canonicalization")["relative
canonicalization with /", "relative canonicalization from ice-9",
"absolute canonicalization from ice-9"]: New tests.
Ludovic Courtès [Tue, 15 May 2012 17:05:37 +0000 (19:05 +0200)]
Fix relative file name canonicalization with empty %LOAD-PATH entries.
* libguile/filesys.c (scm_i_relativize_path): Don't attempt to
canonicalize when encountering an entry of IN_PATH that is the empty
string.
* test-suite/tests/ports.test (with-load-path): New macro.
("%file-port-name-canonicalization"): New test prefix.
Andy Wingo [Tue, 15 May 2012 15:37:57 +0000 (17:37 +0200)]
fix the cse tests
* test-suite/tests/cse.test (pass-if-cse): Fix-letrec and canonicalize
the output, so that unreferenced failure continuations get trimmed.
("cse"): Fix the two tests regarding bailout info.
Andy Wingo [Tue, 15 May 2012 15:23:06 +0000 (17:23 +0200)]
cse passes a lookup procedure to the effects analyzer
* module/language/tree-il/cse.scm (cse): Arrange to pass a lookup
procedure to compute-effects, for better effects analysis.
Andy Wingo [Tue, 15 May 2012 15:22:05 +0000 (17:22 +0200)]
better effects analysis for calls to lexically bound procedures
* module/language/tree-il/effects.scm (make-effects-analyzer): The
analyzer will take an optional second argument, a lookup procedure of
type sym -> exp. This can let the analyzer dig into calls to
lexically bound procedures.
Andy Wingo [Tue, 15 May 2012 15:20:57 +0000 (17:20 +0200)]
fix-letrec tweak
* module/language/tree-il/fix-letrec.scm (make-sequence*, fix-letrec!):
When turning unreferenced bindings into sequences, don't bother
emitting trivially constant expressions in effect position.
Andy Wingo [Tue, 15 May 2012 15:20:01 +0000 (17:20 +0200)]
CSE in tail position
* module/language/tree-il/cse.scm (singly-valued-expression?, cse):
Allow CSE to propagate lexicals to tail positions, if the expression
is singly-valued.
Andy Wingo [Tue, 15 May 2012 10:21:57 +0000 (12:21 +0200)]
stronger conditional optimization
* module/language/tree-il/peval.scm (peval): If we can lift one common
test, see if we can lift others as well.
* test-suite/tests/peval.test: Add a test.
Andy Wingo [Tue, 15 May 2012 10:18:30 +0000 (12:18 +0200)]
optimization for chain of if expressions with common tests
* module/language/tree-il/peval.scm (peval): Optimize common tests in
chains of "if" expressions, like those generated by matchers.
* test-suite/tests/peval.test ("partial evaluation"): Add a test.
Andy Wingo [Tue, 15 May 2012 10:14:22 +0000 (12:14 +0200)]
mvoe tree-il=? and tree-il-hash to tree-il.scm
* module/language/tree-il.scm (tree-il=?, tree-il-hash): Move these
helpers here, from cse.scm. Export them.
* module/language/tree-il/cse.scm (cse): Adapt accordingly.
Ludovic Courtès [Mon, 14 May 2012 17:21:35 +0000 (19:21 +0200)]
Honor $(program_transform_name) for the `guile-tools' symlink.
Fixes <http://bugs.gnu.org/11451>.
Reported by Cyprien Nicolas <c.nicolas@gmail.com>.
* meta/Makefile.am (install-data-hook): Honor $(program_transform_name).
Ludovic Courtès [Sat, 12 May 2012 14:11:51 +0000 (16:11 +0200)]
Have `-Wformat' remain quiet for any procedure called `_' or `N_'.
* module/language/tree-il/analyze.scm (proc-ref?)[special?]: New
procedure.
Return #t for any toplevel-ref of `_'.
* test-suite/tests/tree-il.test ("warnings")["format"]("non-literal
format string using gettext as top-level _"): New test.
Ludovic Courtès [Sat, 12 May 2012 13:58:23 +0000 (15:58 +0200)]
Have `-Warity-mismatch' handle applicable structs.
* module/language/tree-il/analyze.scm (arity-analysis): Honor applicable
structs.
* test-suite/tests/tree-il.test ("warnings")["arity
mismatch"]("top-level applicable struct", "top-level applicable struct
with wrong arguments"): New tests.
Ludovic Courtès [Sat, 12 May 2012 13:31:28 +0000 (15:31 +0200)]
Improve special-casing of `_' in `-Wformat'.
* module/language/tree-il/analyze.scm (proc-ref?): Check for
SPECIAL-NAME in the <module-ref> and <lexical-ref> cases too.
* test-suite/tests/tree-il.test ("warnings")["format"]("non-literal
format string using gettext as module-ref _", "non-literal format
string using gettext as lexical _"): New tests.
Andy Wingo [Fri, 11 May 2012 12:30:43 +0000 (14:30 +0200)]
fix more assumptions that the frame-procedure is a procedure
* libguile/frames.c (scm_frame_source, scm_frame_previous):
* libguile/stacks.c (scm_make_stack):
* module/ice-9/boot-9.scm (exception-printers):
* module/system/vm/frame.scm (frame-call-representation): Fix more
assumptions that frame-procedure is a program, or even a procedure.
Andy Wingo [Thu, 10 May 2012 13:55:25 +0000 (15:55 +0200)]
Merge remote-tracking branch 'origin/stable-2.0' into stable-2.0
Andy Wingo [Thu, 10 May 2012 10:43:33 +0000 (12:43 +0200)]
refactor vm application of non-programs; boot continuation refactor
* libguile/frames.c (scm_frame_instruction_pointer):
* module/system/vm/frame.scm (frame-bindings):
(frame-next-source, frame-call-representation): Fix a few locations
that thought that the frame-procedure will always be a VM
procedure. This will not not be the case when traversing the stack of
an application of a non-procedure.
* libguile/vm-i-system.c (call, tail-call, mv-call): Instead of
special-casing structs and smobs at these call sites, just set up the
stack, and jump to a generic apply loop if the proc is not a program.
* libguile/vm-engine.c: The generic apply loop is here. Also, the boot
program is now simply a boot continuation, and can handle any number
of arguments.
* libguile/vm.c (make_boot_program): Update the code that makes the boot
continuation.
Ludovic Courtès [Tue, 8 May 2012 10:43:06 +0000 (12:43 +0200)]
Add missing `SCM_API' for `scm_take_from_input_buffers'.
* libguile/ports.h (scm_take_from_input_buffers): Add `SCM_API'.
Ludovic Courtès [Tue, 8 May 2012 10:43:06 +0000 (12:43 +0200)]
Fix `setvbuf' to leave the line/column number unchanged.
* libguile/fports.c (scm_setvbuf): Use `scm_take_from_input_buffers'
directly instead of `scm_drain_input'; use `scm_unget_byte' instead of
`scm_unread_string' to put the drained input back to PORT. This
leaves PORT's line/column numbers unchanged, whereas they'd previously
be decreased by the `scm_unread_string' call.
* libguile/ports.c (scm_take_from_input_buffers): Update description and
variable names to refer to "bytes", not "chars".
* test-suite/tests/ports.test ("setvbuf"): New test prefix.
Andy Wingo [Tue, 8 May 2012 14:03:54 +0000 (16:03 +0200)]
docbook->texi fix
* module/texinfo/docbook.scm (*sdocbook->stexi-rules*): Only convert
ulink to uref if there is a URL attribute.
Andy Wingo [Mon, 7 May 2012 18:39:14 +0000 (20:39 +0200)]
(texinfo docbook) support for @acronym
* module/texinfo/docbook.scm (*sdocbook->stexi-rules*): Recognize
"acronym" as parsing to the @acronym texinfo command.
Andy Wingo [Mon, 7 May 2012 18:18:56 +0000 (20:18 +0200)]
add support for texinfo parsed arguments, like @acronym
* module/texinfo.scm (texi-command-specs): Add a new kind of texinfo
command, inline-text-args, a sort of a cross between inline-args,
which are unparsed, and inline-text, which is. Perhaps this should
supersede inline-args at some point. In any case, add acronym as an
inline-text-args element.
(inline-content?, arguments->attlist, complete-start-command)
(parse-inline-text-args, make-dom-parser): Adapt for
inline-text-args.
* module/texinfo/serialize.scm (inline-text-args): Add serialization for
@acronym.
* test-suite/tests/texinfo.test ("test-texinfo->stexinfo"): Add some
tests.
Andy Wingo [Wed, 2 May 2012 10:59:11 +0000 (12:59 +0200)]
Add a missing SYNC_ALL in variable-ref
* libguile/vm-i-system.c (variable-ref): Add a missing SYNC_ALL.
Ludovic Courtès [Sun, 6 May 2012 22:32:01 +0000 (00:32 +0200)]
Optimize `scm_read_string'.
According to the new benchmarks, this leads a 5% speed improvement when
reading small strings, and a 27% improvement when reading large strings.
* libguile/read.c (READER_STRING_BUFFER_SIZE): Change to 128; update
comment to mention codepoints.
(scm_read_string): Make `str' a list of strings, instead of a string.
Store characters read in buffer `c_str'. Cons to STR when C_STR is
full, and concatenate/reverse at the end.
* benchmark-suite/benchmarks/read.bm (small, large): New variables.
Set %DEFAULT-PORT-ENCODING to "UTF-8".
("read")["small strings", "large strings"]: New benchmarks.
Ludovic Courtès [Sun, 6 May 2012 20:24:47 +0000 (22:24 +0200)]
Make `scm_unget_byte' public.
* libguile/ports.h (scm_unget_byte): Make `SCM_API' instead of
`SCM_INTERNAL'.
Ludovic Courtès [Sun, 6 May 2012 20:23:58 +0000 (22:23 +0200)]
read: Avoid `void *' pointer arithmetic.
* libguile/read.c (read_complete_token): Make `new_buf' a `char *' to
avoid pointer arithmetic on `void *'.
Ludovic Courtès [Sun, 6 May 2012 20:23:12 +0000 (22:23 +0200)]
Build with `-Wpointer-arith' when available.
* configure.ac (POTENTIAL_GCC_CFLAGS): Add `-Wpointer-arith'.
Ludovic Courtès [Fri, 4 May 2012 22:44:54 +0000 (00:44 +0200)]
Correct typing of assertion in the `wind' instruction.
* libguile/vm-i-system.c (wind): Check `scm_to_bool (scm_thunk_p (x))'
instead of `scm_thunk_p'.
Ludovic Courtès [Fri, 4 May 2012 20:36:27 +0000 (22:36 +0200)]
Simplify the reader's `read_complete_token'.
* libguile/read.c (read_token): Remove unneeded `const' before `size_t'.
(read_complete_token): Remove `overflow_buffer' parameter; return
`char *' instead of `int'. Allocate the overflow buffer with
`scm_gc_malloc_pointerless' instead of `scm_malloc'. Return either
the overflow buffer or BUFFER.
(scm_read_number, scm_read_mixed_case_symbol,
scm_read_number_and_radix): Rename `buffer' to `local_buffer', and
`overflow_buffer' to `buffer'. Remove `overflow'. Adjust code to new
`read_complete_token'.
Andy Wingo [Mon, 30 Apr 2012 18:25:53 +0000 (20:25 +0200)]
push error handlers out of line in the vm
* libguile/vm.c:
(vm_error):
(vm_error_bad_instruction):
(vm_error_unbound):
(vm_error_unbound_fluid):
(vm_error_not_a_variable):
(vm_error_not_a_thunk):
(vm_error_apply_to_non_list):
(vm_error_kwargs_length_not_even):
(vm_error_kwargs_invalid_keyword):
(vm_error_kwargs_unrecognized_keyword):
(vm_error_too_many_args):
(vm_error_wrong_num_args):
(vm_error_wrong_type_apply):
(vm_error_stack_overflow):
(vm_error_stack_underflow):
(vm_error_improper_list):
(vm_error_not_a_pair):
(vm_error_not_a_bytevector):
(vm_error_not_a_struct):
(vm_error_no_values):
(vm_error_not_enough_values):
(vm_error_continuation_not_rewindable):
(vm_error_bad_wide_string_length):
(vm_error_invalid_address):
(vm_error_object):
(vm_error_free_variable): New internal helpers, implementing VM error
handling.
* libguile/vm-engine.h (VM_ASSERT): New helper macro.
(ASSERT, CHECK_OBJECT, CHECK_FREE_VARIABLE):
(PRE_CHECK_UNDERFLOW, PUSH_LIST): Use the new helper.
* libguile/vm-i-loader.c:
* libguile/vm-i-scheme.c:
* libguile/vm-i-system.c: Use VM_ASSERT and the out-of-line error
handlers.
* libguile/vm-engine.c (vm_engine): Remove inline error handlers, and
remove a couple of local vars. Use VM_ASSERT. Have halt handle the
return itself.
Andy Wingo [Mon, 30 Apr 2012 17:51:06 +0000 (19:51 +0200)]
add internal SCM_NOINLINE definition
* libguile/_scm.h (SCM_NOINLINE): New internal define, for things that
we definitely don't want the compiler to inline.
Andy Wingo [Mon, 30 Apr 2012 17:40:52 +0000 (19:40 +0200)]
add scm_c_values helper
* libguile/values.h:
* libguile/values.c (scm_c_values): New public helper.
Andy Wingo [Mon, 23 Apr 2012 19:42:40 +0000 (21:42 +0200)]
vlist performance improvements; allocate vhash data inline
* module/ice-9/vlist.scm (make-block): If we are making a hash table,
allocate it inline with the contents. Otherwise don't even add a
pointer to the block.
(block-hash-table?): New internal accessor.
(block-ref*): Remove. Vhash entries are no longer wrapped.
(block-ref):
(block-hash-table-next-offset):
(block-hash-table-set-next-offset!):
(block-hash-table-ref):
(block-hash-table-set!):
(block-hash-table-add!): Adapt to take content vector explicitly, and
to expect the hash table inline with the contents. Some of these
accessors are new. Adapt callers.
(assert-vlist): New helper.
(vlist-cons): Update comment.
(vhash?): Update scheme to allocate the hash table and chain links
inline with the contents.
(%vhash-fold*, %vhash-assoc): Rewrite to be more performant.
Andy Wingo [Mon, 23 Apr 2012 15:56:28 +0000 (17:56 +0200)]
avoid emitting degenerate aliases in peval
* module/language/tree-il/peval.scm (<operand>, make-operand)
(make-bound-operands, peval): Avoid emitting needless aliases in
degenerate cases of let.
(visit-operand): If we visit an operand with a fresh counter and have
to abort, record that fact.
* test-suite/tests/peval.test ("partial evaluation"): Add a test.
Andy Wingo [Mon, 23 Apr 2012 11:07:34 +0000 (13:07 +0200)]
vlist-cons micro-optimizations
* module/ice-9/vlist.scm (set-block-next-free!): Define this instead of
increment-block-next-free!.
(block-append!): Refactor to take an offset, and only append if the
offset is the next free value, and there is space in the block.
(block-cons): Refactor to not be a loop. The partial evaluator would
have to understand effects analysis in order to be able to unroll it,
and there's at most one recursion.
Recovers the performance loss resulting from the previous commit.
Andy Wingo [Mon, 23 Apr 2012 09:43:01 +0000 (11:43 +0200)]
slight vlist refactor
* module/ice-9/vlist.scm: Use define-inlinable instead of define-inline,
to ensure strict argument evaluation. There is a slight performance
penalty, but I hope subsequent hacks make it up.
Andy Wingo [Mon, 16 Apr 2012 23:25:19 +0000 (16:25 -0700)]
fix replacement of CSE with lexical-ref
* module/language/tree-il/cse.scm (cse): Fix dominator unrolling for
lexical propagation.
* test-suite/tests/cse.test ("cse"): Add test.
Andy Wingo [Mon, 16 Apr 2012 19:42:31 +0000 (12:42 -0700)]
cse hashing tweak
* module/language/tree-il/cse.scm (cse): Minor tweak to hash depth based
on time profile of compiling peval.scm.
Andy Wingo [Sun, 15 Apr 2012 20:41:05 +0000 (13:41 -0700)]
more inlining in effects.scm
* module/language/tree-il/effects.scm (define-effects)
(&no-effects, &all-effects-but-bailout):
(cause, &depends-on, &causes, depends-on-effects?)
(causes-effects?, effects-commute?): Add ham-fisted inlining.
Andy Wingo [Sun, 15 Apr 2012 20:39:56 +0000 (13:39 -0700)]
better primitives support for bit operations
* module/language/tree-il/primitives.scm
(*interesting-primitive-names*): Add lognot.
(*effect-free-primitives*): Add ash, logand, logior, logxor, and
lognot.
(logior, logand): Define associative expanders.
Andy Wingo [Sun, 15 Apr 2012 20:00:30 +0000 (13:00 -0700)]
disable optimizations in goops dispatch procedures
* module/oop/goops/dispatch.scm: Disable peval and cse.
Andy Wingo [Fri, 13 Apr 2012 22:23:20 +0000 (15:23 -0700)]
enable cse
* module/language/tree-il/optimize.scm: Enable CSE unless #:cse? #f is
passed.
* test-suite/tests/tree-il.test: Disable CSE for one test.
Andy Wingo [Fri, 13 Apr 2012 22:23:15 +0000 (15:23 -0700)]
new pass: cse
* module/language/tree-il/cse.scm: New pass, some simple common
subexpression elimination with effects analysis.
* test-suite/tests/cse.test: New test.
* test-suite/Makefile.am:
* module/Makefile.am: Adapt.
Andy Wingo [Fri, 13 Apr 2012 22:13:19 +0000 (15:13 -0700)]
minor tweaks to a peval test
* test-suite/tests/peval.test: Update mutable var test to really ensure
that the function can't inline.
Andy Wingo [Thu, 12 Apr 2012 23:46:18 +0000 (16:46 -0700)]
peval uses effects analysis
* module/language/tree-il/peval.scm: Use effects analysis from (language
tree-il effects) instead of our own constant-expression?. Eagerly
mark assigned lexicals as non-copyable.
Andy Wingo [Thu, 12 Apr 2012 23:25:45 +0000 (16:25 -0700)]
add effects
* module/language/tree-il/effects.scm: New module, for effects
analysis.
* module/Makefile.am: Adapt.
Andy Wingo [Wed, 11 Apr 2012 18:43:00 +0000 (11:43 -0700)]
peval tests into separate file
* test-suite/tests/tree-il.test ("partial evaluation"):
* test-suite/tests/peval.test ("partial evaluation"): Separate peval
* tests.
* test-suite/Makefile.am: Adapt.
Andy Wingo [Tue, 10 Apr 2012 22:47:21 +0000 (15:47 -0700)]
add more primitives and predicates to (language tree-il primitives)
* module/language/tree-il/primitives.scm
(*interesting-primitive-names*): Add number? and char?. Add more
numeric predicates. Add character comparators. Add throw, error, and
scm-error.
(*primitive-accessors*): Remove struct-vtable. Though the vtable's
contents may change (through redefinition), its identity does not
change.
(*effect-free-primitives*): Put struct-vtable, number?, and char?
here.
(*multiply-valued-primitives*): Instead of listing singly-valued
primitives, list multiply-valued primitives.
(*bailout-primitives*): New list.
(*negatable-primitives*): New alist.
(*bailout-primitive-table*, *multiply-valued-primitive-table*)
(*negatable-primitive-table*): New tables.
(singly-valued-primitive?): Adapt to
use *multiply-valued-primitive-table*.
(bailout-primitive?, negate-primitive): New exported procedures.
Andy Wingo [Mon, 23 Apr 2012 08:42:09 +0000 (04:42 -0400)]
modernize (benchmark-suite lib)
* benchmark-suite/benchmark-suite/lib.scm: Rewrite to be more modern,
using parameters, records, and higher precision timers. Since this
file was never installed, this is an acceptable interface change.
(run-benchmark): Run the thunk once before going into the benchmark.
Adapt to new `report' interface.
(report): Change to expect only one argument, a <benchmark-result>
object.
(print-result): Adapt. The result is in the same format as before.
(print-user-result): Adapt. The result is different from before, but
as this is just printed on stdout and not logged, there should be no
problem.
(calibrate-benchmark-framework): Pull initialization into a function.
Andy Wingo [Mon, 23 Apr 2012 08:27:34 +0000 (04:27 -0400)]
avoid inexact iteration count in benchmarks
* benchmark-suite/benchmarks/arithmetic.bm:
* benchmark-suite/benchmarks/r6rs-arithmetic.bm: Use #e1e7 for the
iteration count, instead of the flonum 1e7.
Andy Wingo [Sat, 21 Apr 2012 21:06:48 +0000 (17:06 -0400)]
move (test-suite lib) to lower dir; cleans up uninstalled paths.
* check-guile.in:
* test-suite/Makefile.am:
* test-suite/test-suite/lib.scm:
* benchmark-guile.in:
* benchmark-suite/Makefile.am:
* benchmark-suite/benchmark-suite/lib.scm: Lower the lib modules in the
source tree. This lets us remove top_srcdir and top_builddir from the
uninstalled paths.
* test-suite/tests/asm-to-bytecode.test:
* test-suite/tests/brainfuck.test:
* test-suite/tests/compiler.test:
* test-suite/tests/ftw.test:
* test-suite/tests/gc.test:
* test-suite/tests/match.test:
* test-suite/tests/rnrs-libraries.test:
* test-suite/tests/rnrs-test-a.scm:
* test-suite/tests/sxml.match.test: Adapt to not expect that module
names be prefixed with "test-suite".
Ludovic Courtès [Sat, 21 Apr 2012 21:08:49 +0000 (23:08 +0200)]
Thank Klaus.
Ludovic Courtès [Sat, 21 Apr 2012 21:08:49 +0000 (23:08 +0200)]
bytevectors: Fix IEEE-754 endianness conversion.
Fixes <http://bugs.gnu.org/11310>.
Reported by Klaus Stehle <klaus.stehle@uni-tuebingen.de>.
* libguile/ieee-754.h: Remove.
* libguile/Makefile.am (noinst_HEADERS): Remove `ieee-754.h'.
* libguile/bytevectors.c (scm_ieee754_float, scm_ieee754_double): New
unions.
(float_to_foreign_endianness, float_from_foreign_endianness,
double_to_foreign_endianness, double_from_foreign_endianness): Rewrite
in terms of the new unions.
* test-suite/tests/bytevectors.test ("2.8 Operations on IEEE-754
Representations")["single, little endian", "single, big endian",
"double, little endian", "double, big endian"]: New tests.
Ludovic Courtès [Sat, 21 Apr 2012 21:08:49 +0000 (23:08 +0200)]
Fix and clarify documentation of `sorted?'.
Fixed <http://bugs.gnu.org/11262>.
Reported by Alexei Matveev <alexei.matveev@gmail.com>.
* libguile/sort.c (scm_sorted_p): Fix and clarify docstring.
* doc/ref/api-utility.texi (Sorting): Update accordingly.
Ludovic Courtès [Sat, 21 Apr 2012 21:08:49 +0000 (23:08 +0200)]
Remove duplicate test in `tree-il.test'.
* test-suite/tests/tree-il.test ("partial evaluation"): Remove duplicate
test for `(cons 0 (cons 1 (cons 2 (list 3 4 5))))'.
Ludovic Courtès [Thu, 12 Apr 2012 21:16:48 +0000 (23:16 +0200)]
Avoid use of `GC_PTR' in "smob.h".
Commit
c46fee438cf9f4a3449e8d04e7a54805517fd092 removed the "bdw-gc.h"
include from "smob.h", so better avoid `GC_PTR'.
* libguile/smob.c (scm_i_finalize_smob): Use `void *' instead of `GC_PTR'.
* libguile/smob.h (scm_i_finalize_smob): Update declaration accordingly.
Ludovic Courtès [Mon, 9 Apr 2012 22:17:39 +0000 (00:17 +0200)]
SRFI-9: Set the `record-constructor' slot of the RTD.
Fixed <http://bugs.gnu.org/11196>.
Reported by Klaus Stehle <klaus.stehle@uni-tuebingen.de>.
* module/srfi/srfi-9.scm (define-record-type): Define the contructor
before TYPE-NAME. Set RTD's constructor field.
* test-suite/tests/srfi-9.test ("record compatibility"): New test
prefix.
Mark H Weaver [Wed, 4 Apr 2012 22:58:44 +0000 (18:58 -0400)]
Fix scm_to_utf8_stringn once and for all; optimize; add tests
* libguile/strings.c (scm_to_utf8_stringn): Fix another new bug in this
recent comedy of errors: pass the size of the preallocated buffer to
u32_to_u8. Arrange to call 'scm_i_string_wide_chars' and
'scm_i_string_length' only once each. Rename local variables for
improved code clarity.
* test-suite/standalone/test-conversion.c (test_to_utf8_stringn): New
function to test scm_to_utf8_stringn.
Mark H Weaver [Tue, 3 Apr 2012 13:50:20 +0000 (09:50 -0400)]
Fix scm_to_utf8_stringn to NUL-terminate properly
* libguile/strings.c (scm_to_utf8_stringn): NUL-terminate using the
correct pointer.
Mark H Weaver [Mon, 2 Apr 2012 22:55:45 +0000 (18:55 -0400)]
Fix scm_to_utf8_stringn to return the length in bytes, et al
* libguile/strings.c (u32_u8_length_in_bytes): Internal static function
renamed from u32_u8_strlen, whose name was potentially confusing. For
added safety, handle everything that can be encoded in the more
general UTF-8 encoding: up to six bytes for each code point, with code
points up to 2^31-1.
(scm_to_utf8_stringn): NUL-terminate only if (lenp == NULL).
If (lenp != NULL) return the length in bytes in *lenp.
Andy Wingo [Sun, 1 Apr 2012 18:29:41 +0000 (11:29 -0700)]
fix scm_to_utf8_stringn for wide strings to NUL-append the strings
* libguile/strings.c (u32_u8_strlen, scm_to_utf8_stringn): Sigh, fix the
u32-to-u8 optimization to ensure that the utf8 string is
NUL-terminated.
BT Templeton [Sat, 10 Mar 2012 08:31:58 +0000 (03:31 -0500)]
primitive resolution for public refs
* module/language/tree-il/primitives.scm (resolve-primitives!): Resolve
public module-refs to primitives.
* test-suite/tests/tree-il.test: New tests for primitive resolution.
Andy Wingo [Sun, 11 Mar 2012 09:57:05 +0000 (10:57 +0100)]
add bad-request printer
* module/web/request.scm (bad-request-printer): Add printer for these
exceptions.
Andy Wingo [Sun, 11 Mar 2012 09:24:08 +0000 (10:24 +0100)]
add exception printers for bad-header, bad-header-component
* module/web/http.scm (bad-header-component): Throw
'bad-header-component instead of 'bad-header.
(bad-header-printer, bad-header-component-printer): Add exception
printers.
Andy Wingo [Wed, 7 Mar 2012 11:39:30 +0000 (12:39 +0100)]
micro-optimizations to string-trim-both, and to (web http)
* libguile/srfi-13.c (scm_string_trim, scm_string_trim_right)
(scm_string_trim_both): Take the whitespace fast-path if the char_pred
is scm_char_set_whitespace.
* module/web/http.scm (read-header, split-and-trim, parse-quality-list):
(parse-param-component, parse-credentials, "Content-Type"):
(read-request-line, read-response-line): Use char-set:whitespace
instead of char-whitespace?. It avoids recursing into the VM.
Andy Wingo [Sun, 18 Mar 2012 19:04:28 +0000 (20:04 +0100)]
make applicable smob calls cheaper, and fix a memory leak
* libguile/vm.c (prepare_smob_call): New helper. Now, instead of making
a per-smob trampoline, we will shuffle the smob into the args and use
a gsubr. This prevents a memory leak in which the trampolines, which
were values in a weak-key table, were preventing the smobs from being
collected.
* libguile/vm-i-system.c (call, tail-call, mv-call): Adapt to new smob
application mechanism.
(smob-call): Remove this instruction.
* libguile/smob.h (scm_smob_descriptor): Rename apply_trampoline_objcode
to apply_trampoline.
* libguile/smob.c: Remove our own objcode trampolines in favor of using
scm_c_make_gsubr.
(scm_smob_prehistory): No more trampoline weak map.
* libguile/procprop.c (scm_i_procedure_arity): Adapt to applicable smob
representation change.
Andy Wingo [Mon, 12 Mar 2012 15:58:15 +0000 (16:58 +0100)]
web http: fix Ma -> Mar
* module/web/http.scm (write-date): Fix serialization of Mar. Oops.
Andy Wingo [Sun, 11 Mar 2012 09:29:06 +0000 (10:29 +0100)]
web server http: 400 Bad Request on bad requests
* module/web/server/http.scm (bad-request, http-read): If an exception
is raised while reading a response, write out a 400 Bad Request
response before closing the port.
Andy Wingo [Thu, 8 Mar 2012 22:50:41 +0000 (23:50 +0100)]
more libgc 7.1 compat
* configure.ac:
* libguile/gc.c (GC_set_finalize_on_demand): Check for this function,
and shim if it isn't present.
Andy Wingo [Fri, 24 Feb 2012 12:18:48 +0000 (13:18 +0100)]
run finalizers asynchronously in asyncs
* libguile/finalizers.c: New excitement! We'll be running finalizers
asynchronously, from asyncs. This will make it safer to allocate
while holding a mutex.
(GC_set_finalizer_notifier): Add back-compat shim.
* libguile/init.c (scm_i_init_guile): Init the async finalizer mechanism
during boot.
* libguile/gc.c (scm_storage_prehistory): Tell libgc we'll be finalizing
on demand.
(scm_gc): Explicitly run finalizers here.
* libguile/threads.c (guilify_self_2): Run finalizers here if
queue_finalizer_async happened to run during guilify_self_1.
* configure.ac: Add check for GC_set_finalizer_notifier.
Andy Wingo [Sun, 19 Feb 2012 11:23:29 +0000 (12:23 +0100)]
use the new finalizer helpers
* libguile/foreign.c (scm_set_pointer_finalizer_x)
* libguile/ports.c (finalize_port)
* libguile/smob.c (scm_i_new_smob, scm_i_new_double_smob)
* libguile/struct.c (scm_i_alloc_struct)
* libguile/numbers.c (make_bignum): Use the new API.
Andy Wingo [Sun, 19 Feb 2012 11:22:12 +0000 (12:22 +0100)]
add scm_i_set_finalizer, scm_i_add_finalizer, scm_i_add_resuscitator
* libguile/finalizers.h:
* libguile/finalizers.c: New files.
* libguile.h:
* libguile/Makefile.am: Add to build.
Andy Wingo [Thu, 26 May 2011 14:49:47 +0000 (16:49 +0200)]
scm_new_smob, scm_new_double_smob inline functions
* libguile/smob.h (scm_new_smob, scm_new_double_smob): New constructors,
which do what SCM_NEWSMOB / SCM_NEWSMOB3 had done, but with inline
functions instead of macros. They also bail to scm_i_new_smob /
scm_i_new_double_smob in either the mark or the free case, so that the
inline definition doesn't reference other internal details like libgc
stuff.
(SCM_SMOB_TYPE_MASK et al): Move definitions up so the new_smob see
them as already being declared.
(SCM_NEWSMOB, SCM_RETURN_NEWSMOB, SCM_NEWSMOB2, SCM_RETURN_NEWSMOB2):
(SCM_NEWSMOB3, SCM_RETURN_NEWSMOB3): Reimplement in terms of the new
inline functions.
Remove now-unneeded bdw-gc include.
* libguile/smob.c (finalize_smob): Rename from scm_i_finalize_smob, and
make static.
(scm_i_new_smob, scm_i_new_double_smob): Slow-path allocators.
(scm_i_finalize_smob, scm_i_new_smob_with_mark_proc): Add
back-compatibility shims to preserve ABI.
* libguile/inline.c: Include smob.h, so as to reify scm_new_smob and
scm_new_double_smob.
Andy Wingo [Thu, 26 May 2011 14:08:02 +0000 (16:08 +0200)]
scm_cell, scm_double_cell, scm_words back to gc.h
* libguile/inline.h:
* libguile/gc.h (scm_cell, scm_double_cell, scm_words): Move
declarations and definitions back here, from inline.h. It's more
natural.
* libguile/inline.c: Include gc.h as well.
Andy Wingo [Thu, 26 May 2011 13:53:02 +0000 (15:53 +0200)]
simplify inline function infrastructure
* libguile/__scm.h (SCM_C_EXTERN_INLINE): Move this definition here,
from inline.h. We'd like to support inline function definitions in
more header files: not just inline.h.
(SCM_CAN_INLINE, SCM_INLINE, SCM_INLINE_IMPLEMENTATION): New
definitions.
* libguile/gc.h (SCM_GC_MALLOC, SCM_GC_MALLOC_POINTERLESS): Define these
wrappers, which redirect to the GC_MALLOC macros when building Guile,
and the scm_gc_malloc functions otherwise. A step towards getting
BDW-GC out of Guile's API.
* libguile/inline.h: Simplify, using SCM_INLINE,
SCM_INLINE_IMPLEMENTATION, and SCM_IMPLEMENT_INLINES. Also use the
new SCM_GC_MALLOC macros.
Ian Price [Fri, 27 Jan 2012 06:38:09 +0000 (06:38 +0000)]
Fix flush on soft ports, so that it actually runs.
* libguile/vports.c (sf_flush): Remove conditional testing the
position in the port's write_buf, as it is no longer used.
Ludovic Courtès [Wed, 7 Mar 2012 20:57:58 +0000 (21:57 +0100)]
Improve port benchmark.
* benchmark-suite/benchmarks/ports.bm (sequence): New macro, formerly
local to the "rdelim" benchmark prefix.
(large-string): New procedure.
(%latin1-port, %utf8/ascii-port, %utf8/wide-port): Use it.
("peek-char", "char-ready?", "read-char"): Use `sequence'.
Giuseppe Scrivano [Tue, 21 Feb 2012 14:55:44 +0000 (15:55 +0100)]
Fix a segfault when /dev/urandom is not accessible
* libguile/init.c (scm_i_init_guile): Call `scm_init_random' before
`scm_init_macros'.
Andy Wingo [Mon, 5 Mar 2012 22:52:00 +0000 (23:52 +0100)]
deprecate close-io-port
* module/ice-9/r4rs.scm:
* module/ice-9/deprecated.scm (close-io-port): Deprecate.
Mark H Weaver [Mon, 5 Mar 2012 03:47:04 +0000 (22:47 -0500)]
Don't call SYMBOL_STRINGBUF on a string
* libguile/strings.c (scm_i_make_symbol): Use STRING_STRINGBUF
instead of SYMBOL_STRINGBUF to get the stringbuf of a string.
Mark H Weaver [Mon, 5 Mar 2012 03:12:48 +0000 (22:12 -0500)]
Remove documentation of internal static string comparison functions
* doc/ref/api-data.texi (String Comparison): Remove documentation for
scm_i_string{,_ci}_{equal,less,leq,gr,geq}_p, which are not only
internal functions but static, and thus unusable by external code.
Mark H Weaver [Sun, 4 Mar 2012 17:01:10 +0000 (12:01 -0500)]
Comment out unused definitions of 'do' and 'case' in psyntax.scm
* module/ice-9/psyntax.scm (do, case): Comment out these definitions,
which are never used and immediately replaced by definitions in
boot-9.scm.
* module/ice-9/psyntax-pp.scm: Regenerate.
Mark H Weaver [Sun, 4 Mar 2012 16:53:23 +0000 (11:53 -0500)]
Fix syntax wrap comments; remove old unused cruft
* module/ice-9/psyntax.scm (subst-rename?, rename-old, rename-new,
rename-marks, make-rename): Remove these unused syntax rules.
Fix description of syntax wraps to match current reality.
Mark H Weaver [Fri, 2 Mar 2012 23:40:43 +0000 (18:40 -0500)]
Minimize size of embedded syntax objects in psyntax-pp.scm
* module/ice-9/compile-psyntax.scm: Minimize syntax object literals
embedded in psyntax-pp.scm.
* module/ice-9/psyntax.scm: Rename a few variables so that syntax
objects embedded in macros have no lexical bindings, so that their
minimized syntax objects will have no embedded labels. These labels
were the last remaining gensym counters in psyntax-pp.scm.
* module/ice-9/psyntax-pp.scm: Regenerate. It is now less than one
quarter of its previous size! More importantly, this file no longer
contains any gensym counters, which means that in the future, local
changes to psyntax.scm will usually result in only local changes to
psyntax-pp.scm.
Mark H Weaver [Thu, 1 Mar 2012 22:56:14 +0000 (17:56 -0500)]
tree-il->scheme improvements
* module/language/tree-il.scm (tree-il->scheme): New implementation that
simply calls 'decompile-tree-il'.
* module/language/scheme/decompile-tree-il.scm (choose-output-names,
do-decompile): New internal procedures.
(decompile-tree-il): New and improved implementation. Print source
identifiers where possible, otherwise add minimal numeric suffixes.
Previously we printed the gensyms. Avoid 'begin' in contexts that
provide an implicit 'begin'. Produce 'cond', 'case', 'and', 'or',
'let*', named let, and internal defines where appropriate. Recognize
keyword arguments in 'opts' to disable the production of these derived
syntactic forms, and to optionally strip numeric suffixes from
variable names.
* module/ice-9/compile-psyntax.scm: Disable partial evaluation, letrec
fixing, and primitive expansion when producing psyntax-pp.scm, in
order to produce output as close to the original source as practical.
Disable production of derived syntactic forms as needed during
bootstrap. Strip numeric suffixes from variable names. Adjust
pretty-printing parameters.
* module/ice-9/psyntax-pp.scm: Regenerate. It is now less than half
of the original size.
Mark H Weaver [Sun, 26 Feb 2012 20:58:30 +0000 (15:58 -0500)]
pretty-print: allow max-expr-width to be set; recognize more keywords
* module/ice-9/pretty-print.scm (pretty-print): Add new keyword argument
'#:max-expr-width'.
(generic-write): Add new argument 'max-expr-width'. Previously this
was internally defined to the constant value 50.
Mark H Weaver [Fri, 2 Mar 2012 22:54:23 +0000 (17:54 -0500)]
Fix typos in psyntax.scm
* module/ice-9/psyntax.scm (gen-syntax-case): #'pad --> #'pat
(define*): #'x --> #'id
* module/ice-9/psyntax-pp.scm: Regenerate
Andy Wingo [Fri, 2 Mar 2012 16:46:28 +0000 (17:46 +0100)]
tweaks to -Wformat's gettext detection
* module/language/tree-il/analyze.scm (proc-ref?): Change to use less
false-if-exception and more variable-bound?. If a variable is present
in the local module but not bound, assume that it is gettext if it has
the right name. This is to allow for (define _ gettext).
* test-suite/tests/tree-il.test ("warnings"): Update (_ "foo") example.
Andy Wingo [Fri, 2 Mar 2012 14:51:05 +0000 (15:51 +0100)]
peval: inline applications of lambda to rest args
* module/language/tree-il/peval.scm (peval): Add optimization to
hoist the inner procedure out of e.g.
(lambda args (apply (lambda ...) args))
This commit restores the ability to detect escape-only prompts at
compile-time.
* test-suite/tests/tree-il.test: Update test for prompt with a lambda,
and add a specific test for lambda application.
Andy Wingo [Fri, 2 Mar 2012 15:39:56 +0000 (16:39 +0100)]
tree-il: fix `canonicalize!' for prompts
* module/language/tree-il/canonicalize.scm (canonicalize!): Fix a bug in
which the sense of `escape-only?' was reversed. We never saw this
though, because for other reasons, no prompts were being identified as
escape-only.
Andy Wingo [Fri, 2 Mar 2012 12:02:19 +0000 (13:02 +0100)]
more general treatment of call-with-prompt
* module/language/tree-il/primitives.scm (*primitive-expand-table*):
Don't limit the call-with-prompt to <prompt> transition to lambda
expressions. Instead we can lexically bind the handler, and rely on
peval to propagate a lambda expression.
Mark H Weaver [Sun, 26 Feb 2012 20:53:11 +0000 (15:53 -0500)]
pmatch: always wrap with let, even if the expression appears atomic
* module/system/base/pmatch.scm (pmatch): Always wrap with 'let', even
if the expression appears atomic, because in the presence of
'identifier-syntax', we cannot know what an atomic expression will
later expand to. Also use '#:export-syntax' instead of '#:export'
to export 'pmatch'.
Mark H Weaver [Thu, 1 Mar 2012 21:07:28 +0000 (16:07 -0500)]
Fix <TAG>vector-length when applied to other uniform vector types
* module/srfi/srfi-4.scm, module/srfi/srfi-4/gnu.scm
(define-bytevector-type): Fix definition of <TAG>vector-length when
applied to uniform vectors of different element sizes. Thanks to
Tobias Brandt <tob.brandt@googlemail.com> for reporting this bug.
* test-suite/tests/srfi-4.test: Add tests.
Ludovic Courtès [Sun, 26 Feb 2012 23:51:09 +0000 (00:51 +0100)]
Avoid failure when `format-analysis' stumbles upon unbound variables.
* module/language/tree-il/analyze.scm (proc-ref?): Wrap `variable-ref'
in `false-if-exception'.
* test-suite/tests/tree-il.test ("warnings")["format"]("non-literal
format string with forward declaration"): New test.
Andy Wingo [Fri, 24 Feb 2012 19:01:47 +0000 (20:01 +0100)]
srfi-18 test fix
* test-suite/tests/srfi-18.test: Enclose the tests in a begin instead of
an and. Before, they were not being run, for some reason I don't
fully understand.