Andy Wingo [Sun, 24 Oct 2010 10:36:17 +0000 (12:36 +0200)]
add HTTP module
* module/web/http.scm: New module, declares known HTTP headers, and
their parsers and unparsers.
* test-suite/tests/web-http.test: Add test suite.
* module/Makefile.am:
* test-suite/Makefile.am: Adapt.
Andy Wingo [Sat, 23 Oct 2010 13:23:42 +0000 (15:23 +0200)]
URI parsing errors throw to `uri-error'
* module/web/uri.scm (uri-error): New proc, throws to 'uri-error.
(validate-uri, uri-decode, uri-encode): Use uri-error.
* test-suite/tests/web-uri.test: Update for uri-error.
Neil Jerram [Thu, 4 Nov 2010 01:13:16 +0000 (01:13 +0000)]
Finish adding api-regex.texi
* doc/ref/Makefile.am (guile_TEXINFOS): Add api-regex.texi, so that it
gets into distributions.
Andreas Rottmann [Tue, 2 Nov 2010 23:19:54 +0000 (00:19 +0100)]
Add implementation of SRFI 38
* module/srfi/srfi-38.scm: New file, partly based on the reference
implementation and on Alex Shinn's public-domain implementation for
Chicken.
* module/Makefile.am (SRFI_SOURCES): Add srfi/srfi-38.scm.
* test-suite/tests/srfi-38.test: New file, minimal test suite for SRFI
38.
* test-suite/Makefile.am (SCM_TESTS): Added tests/srfi-38.test.
* doc/ref/srfi-modules.texi: Add a node for SRFI 38.
Signed-off-by: Ludovic Courtès <ludo@gnu.org>
Andreas Rottmann [Tue, 2 Nov 2010 23:09:57 +0000 (00:09 +0100)]
Use a fluid for the list of the reader's "hash procedures"
This allows customizing the reader behavior for a dynamic extent more easily.
* libguile/read.c (scm_read_hash_procedures): Renamed to
`scm_i_read_hash_procedures'.
(scm_i_read_hash_procedures_ref, scm_i_read_hash_procedures_set_x):
New (internal) accessor functions for the fluid.
(scm_read_hash_extend, scm_get_hash_procedure): Use these accessor
functions.
(scm_init_read): Create the fluid, named `%read-hash-procedures' instead of
the previous plain list `read-hash-procedures'.
* test-suite/tests/reader.test: Adapt the "R6RS/SRFI-30 block comment
syntax overridden" test to make use of the fluid.
* module/ice-9/deprecated.scm (read-hash-procedures):
New identifier macro -- backward-compatibility shim.
Signed-off-by: Ludovic Courtès <ludo@gnu.org>
Neil Jerram [Sun, 31 Oct 2010 08:34:05 +0000 (08:34 +0000)]
Manual sections don't need a page break before
* doc/ref/api-coverage.texi (Code Coverage): Remove @page.
* doc/ref/api-foreign.texi (Foreign Function Interface): Ditto.
* doc/ref/api-lalr.texi: (LALR(1) Parsing): Ditto.
* doc/ref/api-macros.texi (Macros): Ditto.
Neil Jerram [Sun, 31 Oct 2010 08:27:56 +0000 (08:27 +0000)]
Add ref to new location for regex doc
* doc/ref/api-data.texi: Add ref to new location for regex doc. Also
correct other refs to say section rather than chapter.
Neil Jerram [Sun, 31 Oct 2010 08:24:28 +0000 (08:24 +0000)]
Promote regex doc out of the `Simple Data Types' section
Because that probably isn't where people will look for it.
Thanks to Noah Lavine for the idea.
* doc/ref/api-regex.texi (Regular Expressions): New file, containing
the regex doc (promoted one level) that used to be in api-data.texi.
* doc/ref/guile.texi (API Reference): Include new file, and add menu
entry for the new section.
* THANKS: Add Noah.
Ludovic Courtès [Wed, 27 Oct 2010 22:03:54 +0000 (00:03 +0200)]
Fix typo in the (system base lalr) documentation.
* doc/ref/api-lalr.texi (LALR(1) Parsing): "The tokenizer should be a
unary procedure" -> "thunk". Patch by Noah Lavine
<noah.b.lavine@gmail.com>.
Andreas Rottmann [Wed, 27 Oct 2010 21:33:40 +0000 (23:33 +0200)]
Extend the #:replace list of the SRFI 69 module
* module/srfi/srfi-69.scm: Add `make-hash-table' and `hash-table?' to
the #:replace list of the module definition.
Ludovic Courtès [Wed, 27 Oct 2010 21:31:49 +0000 (23:31 +0200)]
Have "frisk" recognize `#:use-module' and `#:autoload'.
* module/scripts/frisk.scm (grok-proc): Support keywords `#:use-module'
and `#:autoload'.
Ludovic Courtès [Wed, 27 Oct 2010 21:28:58 +0000 (23:28 +0200)]
Augment `arithmetic.bm'.
* benchmark-suite/benchmarks/arithmetic.bm (repeat): Change the syntax.
Add support for binary OP.
("fixnum")["1+", "1-"]: Adjust accordingly.
["+", "-"]: New benchmarks.
Neil Jerram [Tue, 26 Oct 2010 08:11:52 +0000 (09:11 +0100)]
Fix manual typo
* doc/ref/api-foreign.texi: "prodedures" -> "procedures".
Julian Graham [Sun, 10 Oct 2010 05:35:26 +0000 (01:35 -0400)]
Improve performance of R6RS records implementation
Reimplement record-type descriptors as vtables for record structs, saving
us what was an expensive inspection of a record's vtable layout string to
determine its type.
* module/rnrs/records/inspection.scm (record-field-mutable?): Check
mutability using the bit field stored in the record-type descriptor
instead of the record struct's vtable.
* module/rnrs/records/procedural.scm (record-internal?): Reimplement as a
delegation to a check of the passed struct's vtable against
`record-type-descriptor?'.
(record-type-vtable): Modify to include base vtable layout as a prefix
of the record-type-descriptor layout so that all record-type instances
are now also vtables.
(make-record-type-descriptor): Remove field vtable; build up a mutability
bit field to use for fast mutability checks.
(record-accessor, record-mutator): Use field struct and mutability bit
field.
Ludovic Courtès [Wed, 20 Oct 2010 20:40:10 +0000 (22:40 +0200)]
Tweak `-Wunused-variable' and `-Wunused-toplevel' for special names.
* module/language/tree-il/analyze.scm (gensym?): New procedure.
(unused-variable-analysis): Ignore variables whose name passes
`gensym?' or is `_'.
(unused-toplevel-analysis): Ignore variables whose name passes
`gensym?'.
* test-suite/tests/tree-il.test ("warnings")["unused-variable"]("special
variable names"): New test.
["unused-toplevel"]("special variable names"): New test.
Ludovic Courtès [Tue, 19 Oct 2010 21:36:04 +0000 (23:36 +0200)]
Add `Libs.private' to the `.pc' files.
* meta/guile-2.0-uninstalled.pc.in (Libs): Remove @GUILE_LIBS@.
(Libs.private): New field.
* meta/guile-2.0.pc.in (Libs): Remove @GUILE_LIBS@.
(Libs.private): New field.
Ludovic Courtès [Tue, 19 Oct 2010 21:32:16 +0000 (23:32 +0200)]
configure: Don't needlessly augment $LIBS (and $GUILE_LIBS.)
* configure.ac: Don't augment $LIBS with $LTLIBUNISTRING and
$BDW_GC_LIBS.
* libguile/Makefile.am (libguile_@GUILE_EFFECTIVE_VERSION@_la_LDFLAGS):
Remove @LTLIBINTL@. Add $(BDW_GC_LIBS).
Tristan Colgate-McFarlane [Wed, 20 Oct 2010 08:36:32 +0000 (09:36 +0100)]
Add support for more multicast sockopts.
* libguile/socket.c: Add support for IP_MULTICAST_TTL and IP_MILTICAST_IF sockopts.
Tristan Colgate-McFarlane [Wed, 20 Oct 2010 08:28:30 +0000 (09:28 +0100)]
Fix detection of struct ip_mreq
* configura.ac: netinet/in.h is required to complete definition of
struct ip_mreq.
Andy Wingo [Tue, 19 Oct 2010 20:58:00 +0000 (22:58 +0200)]
rdelim cleanups
* module/ice-9/rdelim.scm: Clean up export list.
(read-line!, read-delimited!, read-delimited, read-line): Use
define*.
Andy Wingo [Mon, 18 Oct 2010 11:35:46 +0000 (13:35 +0200)]
more -Wundef on darwin fixes
* libguile/threads.c:
* libguile/posix.c: Change a couple #if HAVE_FOO to #ifdef HAVE_FOO to
pacify -Wundef. Some vars are defined or not, whereas some are always
defined to 0 or 1. The inconsistency is unfortunate.
Andy Wingo [Mon, 18 Oct 2010 11:29:58 +0000 (13:29 +0200)]
uninitialized var in scm_read_character
* libguile/read.c (scm_read_character): Fix error condition where
charname could be uninitialized.
Andy Wingo [Mon, 18 Oct 2010 11:27:17 +0000 (13:27 +0200)]
_scm GUILE_USE_64_CALLS compilation warning fix
* libguile/_scm.h: Check that GUILE_USE_64_CALLS is defined. Fixes an
error on i386-apple-darwin9.8.0.
Andy Wingo [Fri, 15 Oct 2010 18:56:42 +0000 (20:56 +0200)]
web/uri: reimplement for rfc 3986, add tests
* module/web/uri.scm: Reimplement for RFC 3986.
* module/Makefile.am: Add to build.
* test-suite/Makefile.am:
* test-suite/tests/web-uri.test: Add tests.
Andy Wingo [Fri, 15 Oct 2010 18:51:29 +0000 (20:51 +0200)]
import uri.scm from tekuti, from guile-www
Ludovic Courtès [Fri, 15 Oct 2010 08:19:57 +0000 (10:19 +0200)]
Bump version number for 1.9.13.
* GUILE-VERSION (GUILE_MICRO_VERSION): Increment.
Ludovic Courtès [Fri, 15 Oct 2010 08:36:36 +0000 (10:36 +0200)]
Update `cpp-E.syms' and `cpp-SIG.syms'.
* libguile/cpp-E.syms, libguile/cpp-SIG.syms: Update on a Glibc 2.11.1
system.
Ludovic Courtès [Fri, 15 Oct 2010 08:34:33 +0000 (10:34 +0200)]
Fix the `chknew-E' and `chknew-SIG' rules.
* libguile/Makefile.am (chknew-E chknew-SIG): Run gcc without `-undef'
since otherwise it ends up trying to include headers that are not
available (e.g., <gnu/stubs-32.h> on an x86_64-linux-gnu system.)
Michael Gran [Fri, 15 Oct 2010 01:27:41 +0000 (18:27 -0700)]
Unresolved var errors don't work for top-level vars
* libguile/vm-i-system.c (toplevel_ref, long_toplevel_ref): modified
Michael Gran [Fri, 15 Oct 2010 01:21:28 +0000 (18:21 -0700)]
Update charsets to Unicode 6.0.0
* libguile/srfi-14.i.c: regenerated
Andy Wingo [Thu, 14 Oct 2010 14:13:57 +0000 (16:13 +0200)]
re-implement srfi-34's guard with syntax-case
* module/srfi/srfi-34.scm (guard): Re-implement using syntax-case.
Andy Wingo [Thu, 14 Oct 2010 10:37:31 +0000 (12:37 +0200)]
ssax: add procs to define parsed entities
* module/sxml/ssax.scm (reset-parsed-entity-definitions!)
(define-parsed-entity!): New heavy-handed knobs for adding to the set
of parsed entity definitions.
Neil Jerram [Wed, 13 Oct 2010 22:16:29 +0000 (23:16 +0100)]
Nuke GOOPS `Quick Start' section, in favour of the `Tutorial'
These sections are pretty similar in aim, but `Tutorial' is mostly
better material.
* doc/ref/goops-tutorial.texi (Class definition): Add a sentence about
what slots are.
* doc/ref/goops-tutorial.texi (Tutorial): Remove repetition of the Stk
origin, and index entries that are overly general in the context of
the whole Guile manual.
(Generic functions): Add text here about the nature of methods,
previously in Quick Start.
* doc/ref/goops.texi (Quick Start): Move `Built-in classes' subsection
to be part of `Introspection'. Delete the rest, apart from snippets
moved into Tutorial.
Andy Wingo [Wed, 13 Oct 2010 19:18:44 +0000 (21:18 +0200)]
fix bug in with-fluids in a non-tail position
* libguile/vm-i-system.c (VM_DEFINE_INSTRUCTION): Drop the stack items
corresponding the the with-fluids object. Thanks very much to Stefan
Israelsson Tampe for the fix.
Ludovic Courtès [Tue, 12 Oct 2010 23:31:19 +0000 (01:31 +0200)]
Optimize `1+' and `1-' on fixnums.
* libguile/vm-i-scheme.c (INUM_MAX, INUM_MIN): New macros.
(add1, sub1): Add/subtract without untagging the operand. This leads
to a 44% run time improvement compared to the previous
implementation.
* libguile/vm.c: Include <stdint.h>.
* test-suite/tests/numbers.test ("1+", "1-"): Add tests for
MOST-POSITIVE-FIXNUM, resp. MOST-NEGATIVE-FIXNUM, for 32-bit and
34-bit values thereof.
* benchmark-suite/benchmarks/arithmetic.bm: New file.
* benchmark-suite/Makefile.am (SCM_BENCHMARKS): Add it.
Neil Jerram [Tue, 12 Oct 2010 23:08:27 +0000 (00:08 +0100)]
Edit `Adding Methods to Generic Functions'
* doc/ref/goops.texi (Adding Methods to Generic Functions): Move the
bit about no applicable methods to `Invoking Generic Functions'.
Other minor edits.
(Basic Method Definition): Flattened into parent.
(Method Definition Internals): Moved to MOP section at end of
chapter.
Neil Jerram [Tue, 12 Oct 2010 22:41:41 +0000 (23:41 +0100)]
Two very small edits
* doc/ref/goops.texi (User-defined classes): Remove brackets around
`formally', for better flow.
(Creating Instances): Lengthen hyphen.
Neil Jerram [Tue, 12 Oct 2010 22:40:01 +0000 (23:40 +0100)]
Edit `Generic Functions and Accessors'
* doc/ref/goops.texi (Generic Function Internals): Moved to MOP section
at end of chapter.
(Basic Generic Function Creation): Flattened into parent, and text
simplified.
(Extending Primitives): Renamed from `Extending Guiles Primitives';
removed `generic-capability?', which is no longer available;
simplified a bit.
(Merging Generics): New subsection for the material about merging;
text simplified a bit.
Ludovic Courtès [Tue, 12 Oct 2010 20:57:08 +0000 (22:57 +0200)]
Allocate complex numbers in contiguous pointer-less memory.
* libguile/numbers.h (SCM_COMPLEX_MEM): Remove.
(SCM_COMPLEX_REAL): Change to just fetch the `real' field of the
pointed-to `scm_t_complex'.
(SCM_COMPLEX_IMAG): Likewise.
(scm_t_complex)[type, pad]: New fields.
* libguile/numbers.c (scm_c_make_rectangular): Allocate the whole
complex contiguously, with `scm_gc_malloc_pointerless'.
Ludovic Courtès [Tue, 12 Oct 2010 20:00:28 +0000 (22:00 +0200)]
Allocate inexact numbers in pointer-less memory.
* libguile/numbers.c (scm_from_double): Use `scm_gc_malloc_pointerless'
instead of `scm_double_cell'.
Ludovic Courtès [Tue, 12 Oct 2010 19:59:53 +0000 (21:59 +0200)]
Placate `-Wformat'.
* module/scripts/compile.scm (fail): Use a fancy format string instead
of a non-literal one.
Andy Wingo [Tue, 12 Oct 2010 20:27:01 +0000 (22:27 +0200)]
update NEWS for 1.9.13
* NEWS: Update for 1.9.13.
Neil Jerram [Tue, 12 Oct 2010 19:00:21 +0000 (20:00 +0100)]
Add TeX form of jao's name
* doc/ref/scheme-using.texi: Add TeX equivalent for José.
Andy Wingo [Tue, 12 Oct 2010 11:58:39 +0000 (13:58 +0200)]
fold old news items into main news body
* NEWS: Fold 1.9.12 items into the main body.
Andy Wingo [Tue, 12 Oct 2010 11:24:46 +0000 (13:24 +0200)]
,frame and related commands handle for-trap? appropriately
* module/system/repl/debug.scm (print-frame): Add #:next-source? arg,
for when print-frame should use frame-next-source instead of
frame-source.
(print-frames): Add #:for-trap? arg. If true, the 0th frame should be
printed with frame-next-source.
* module/system/repl/command.scm (define-stack-command): Introduce
for-trap? into the lexical env.
(backtrace, up, down, frame): Update to do the right thing regarding
#:for-trap?.
Andy Wingo [Tue, 12 Oct 2010 11:11:40 +0000 (13:11 +0200)]
don't warn for (format #t fmt) -- format string actually named fmt
* module/language/tree-il/analyze.scm (format-analysis): Don't warn on
non-literal format string if the format string is a lexical ref to a
variable named "fmt". A slight hack, but effective :)
* module/system/repl/command.scm (display-stat): Rename the format
string to "fmt".
Andy Wingo [Tue, 12 Oct 2010 11:09:48 +0000 (13:09 +0200)]
debug has for-trap? field
* module/system/repl/debug.scm (<debug>): New field, `for-trap?'. True
if the stack is for a trap, and thus the top frame should use
frame-next-source instead of frame-source.
* module/system/repl/command.scm (repl-pop-continuation-resumer)
(repl-next-resumer):
* module/system/repl/error-handling.scm (call-with-error-handling):
Update callers.
Andy Wingo [Tue, 12 Oct 2010 11:01:10 +0000 (13:01 +0200)]
remove gds docs, telling users instead to use geiser and paredit
* doc/ref/scheme-using.texi (Using Guile in Emacs): Remove GDS docs, as
we don't have GDS any more. Instead of surveying the field, be
opinionated and tell users what to use: Geiser and Paredit.
Andy Wingo [Tue, 12 Oct 2010 10:58:36 +0000 (12:58 +0200)]
minor doc fixups
* doc/ref/scheme-scripts.texi (Invoking Guile): Some typos.
* doc/ref/tour.texi (Running Guile Scripts): More typos and PDF
improvements.
* doc/ref/scheme-using.texi (Readline, Value History): Minor rewording.
(Help Commands): Minor fixup.
Andy Wingo [Mon, 11 Oct 2010 21:53:40 +0000 (23:53 +0200)]
add --listen docs
* doc/ref/scheme-scripts.texi (Invoking Guile): Add --listen
documentation.
Andy Wingo [Mon, 11 Oct 2010 21:32:30 +0000 (23:32 +0200)]
preface contributors tweaks
* doc/ref/preface.texi (Contributors): A couple tweaks
Andy Wingo [Mon, 11 Oct 2010 20:50:55 +0000 (22:50 +0200)]
update preface.texi contributors
* doc/ref/preface.texi (Contributors): Rewrite and update.
Ludovic Courtès [Mon, 11 Oct 2010 13:38:06 +0000 (15:38 +0200)]
Allow arbitrary code in ASSOC procedures for weak hash tables (bug #29616).
* libguile/hashtab.c (struct t_assoc_args, do_weak_bucket_assoc):
Remove.
(struct t_fixup_args): New type.
(do_weak_bucket_fixup): New function.
(weak_bucket_assoc): Use it. Keep strong references to BUCKET's
entries in STRONG_REFS. Call ASSOC once the alloc lock has been
released. This fixes bug #29616.
* test-suite/tests/weaks.test ("assoc can do anything"): New test.
Ludovic Courtès [Mon, 11 Oct 2010 13:14:55 +0000 (15:14 +0200)]
Fix a bug in weak hash table bucket fixup.
* libguile/hashtab.c (scm_fixup_weak_alist): Keep the value of PREV
unchanged after a nullified pair is deleted; this fixes a bug whereby
if several successive nullified pairs were encountered, not all of them
would be removed, and the assertion in `weak_bucket_assoc' would be
hit. In addition, remove the `scm_is_pair (pair)' test.
Ludovic Courtès [Sun, 10 Oct 2010 17:09:20 +0000 (19:09 +0200)]
Fix typo.
* module/system/repl/command.scm (repl-pop-continuation-resumer): Remove
extraneous argument to `format'.
Ludovic Courtès [Sun, 10 Oct 2010 17:08:11 +0000 (19:08 +0200)]
Add warnings for obvious syntax errors in format strings.
* module/language/tree-il/analyze.scm (&syntax-error): New variable.
(format-string-argument-count): Throw to &SYNTAX-ERROR when a syntax
error in a format string is encountered.
(format-analysis): Catch &SYNTAX-ERROR and convert as a warning of the
appropriate type.
* module/system/base/message.scm (%warning-types)[format]: Handle
`syntax-error' warnings.
* test-suite/tests/tree-il.test
("warnings")["conditionals"]("unterminated", "unexpected ~;",
"unexpected ~]"): New tests.
["unterminated ~{...~}"]: New test.
Ludovic Courtès [Sun, 10 Oct 2010 16:10:18 +0000 (18:10 +0200)]
Add new `format' warnings.
* module/language/tree-il/analyze.scm (format-analysis): Add new
sub-warnings: `wrong-port', `wrong-format-string',
`non-literal-format-string', and `wrong-num-args'.
* module/system/base/message.scm (%warning-types)[format]: Handle
them.
* test-suite/tests/tree-il.test ("warnings")["wrong port arg",
"wrong format string", "non-literal format string",
"wrong number of args"]: New tests.
Ludovic Courtès [Sun, 10 Oct 2010 15:13:36 +0000 (17:13 +0200)]
Compile with `-Wformat'.
* am/guilec (GUILE_WARNINGS): Add `-Wformat'.
Ludovic Courtès [Sun, 10 Oct 2010 15:13:21 +0000 (17:13 +0200)]
Implement fancy format string analysis.
* module/language/tree-il/analyze.scm (format-string-argument-count):
Return two values, the minimum and maximum number of arguments.
Add support for most of `format' escapes, including conditionals.
(format-analysis): Adjust accordingly.
* module/system/base/message.scm (%warning-types)[format]: Take two
arguments, MIN and MAX, instead of EXPECTED. Display warning
accordingly.
* test-suite/tests/tree-il.test ("warnings")["format"]("~%, ~~, ~&, ~t,
~_, and ~\\n", "~{...~}", "~{...~}, too many args", "~@{...~}",
"~@{...~}, too few args", "~(...~)", "~v", "~v:@y", "~*", "~?",
"complex 1", "complex 2", "complex 3"): New tests.
("conditionals"): New test prefix.
Ludovic Courtès [Sat, 9 Oct 2010 16:06:36 +0000 (18:06 +0200)]
Escape newlines from format strings in warnings.
* module/system/base/message.scm (%warning-types)[format]: Escape
newlines from FMT.
Andy Wingo [Sun, 10 Oct 2010 10:56:53 +0000 (12:56 +0200)]
add --listen command line argument.
* libguile/script.c (scm_shell_usage, scm_compile_shell_switches): Add a
--listen argument to spawn a REPL server, possibly specifying the port
or path to listen on. The goal is for this to be the default way to
allow debugging via Emacs or simply using netcat.
Andy Wingo [Sun, 10 Oct 2010 10:15:34 +0000 (12:15 +0200)]
add (system repl server)
* module/system/repl/server.scm: New module, listens on a socket for
connections, then serves repls to those sockets.
* module/Makefile.am: Add repl server.
Andy Wingo [Sun, 10 Oct 2010 10:13:04 +0000 (12:13 +0200)]
fix segfaults when closing the current input port
* libguile/ports.c (scm_char_ready_p, scm_peek_char, scm_unread_char)
(scm_unread_string): Always validate the port, even in the case that
we get it the default current-input-port. Otherwise the following
causes a segfault:
(begin (close-port (current-input-port)) (peek-char))
Andy Wingo [Sun, 10 Oct 2010 09:49:50 +0000 (11:49 +0200)]
readline repl-reader falls back to boot-9 definition for other ports
* guile-readline/ice-9/readline.scm (readline-repl-reader): Pull
definition out of activate-readline. If the current input port is not
the readline port, fall back to the boot-9 repl reader.
(activate-readline): Adapt.
Andy Wingo [Sun, 10 Oct 2010 09:24:29 +0000 (11:24 +0200)]
interrupted syscalls run asyncs before throwing syserror
* libguile/error.c (scm_syserror, scm_syserror_msg): Run pending pending
asyncs before throwing the error, as one of the asyncs might be a
signal handler. But there is unfortunately a race here, as noted in a
comment.
Andy Wingo [Fri, 8 Oct 2010 17:27:45 +0000 (19:27 +0200)]
further repl tweaks
* module/system/repl/error-handling.scm (error-string): Refactor a
little.
(call-with-error-handling): Ensure a trailing newline when printing
the error-msg.
* module/system/repl/repl.scm (run-repl): We don't know the name of the
meta-command here.
Ludovic Courtès [Fri, 8 Oct 2010 13:24:15 +0000 (15:24 +0200)]
Improve pretty-printing of tree-il objects.
* module/language/tree-il.scm (print-tree-il): Print the AST with ~S
instead of ~A.
Ludovic Courtès [Fri, 8 Oct 2010 14:25:32 +0000 (16:25 +0200)]
Add `-Wformat'.
* module/language/tree-il/analyze.scm (format-string-argument-count):
New procedure.
(format-analysis): New tree analysis.
* module/language/tree-il/compile-glil.scm (%warning-passes): Add
`format'.
* module/system/base/message.scm (%warning-types): Add `format'.
* test-suite/tests/tree-il.test (%opts-w-format): New variable.
("warnings")["format"]: New test prefix.
* doc/ref/api-evaluation.texi (Compilation): Mention `format' warnings.
Ludovic Courtès [Fri, 8 Oct 2010 11:50:24 +0000 (13:50 +0200)]
Fixlets for REPL error handling.
* module/system/repl/error-handling.scm (error-string): Don't call
`display-error' when STACK is empty.
(call-with-error-handling): Display ERROR-MSG instead of using
`format', since ERROR-MSG may contain `format' escapes.
* module/system/repl/repl.scm (run-repl): Add missing argument to
`format'.
Ludovic Courtès [Fri, 8 Oct 2010 11:48:02 +0000 (13:48 +0200)]
SRFI-1: Rewrite `split-at' and `split-at!' in Scheme.
This partially reverts commit
bb560b9c16893f762699ba5a3109c8367fff8dfc
(Tue Mar 15 2005).
* module/srfi/srfi-1.scm (out-of-range, split-at, split-at!): New
procedures.
* libguile/srfi-1.c (scm_srfi1_split_at, scm_srfi1_split_at_x): Remove.
* libguile/srfi-1.h (scm_srfi1_split_at, scm_srfi1_split_at_x): Ditto.
Ludovic Courtès [Fri, 8 Oct 2010 09:03:51 +0000 (11:03 +0200)]
SRFI-1: Rewrite `filter-map' in Scheme.
This partially reverts commit
c16359466bcc3f2ebf6d750c069f787f629fc625
(Thu Mar 17 2005).
* libguile/srfi-1.c (scm_srfi1_filter_map): Remove.
* libguile/srfi-1.h (scm_srfi1_filter_map): Ditto.
* module/srfi/srfi-1.scm (filter-map): New procedure.
Ludovic Courtès [Fri, 8 Oct 2010 08:43:59 +0000 (10:43 +0200)]
SRFI-1: Make `fold-right' tail-recursive.
* module/srfi/srfi-1.scm (fold-right): Make tail-recursive.
* test-suite/tests/srfi-1.test ("fold-right"): New test prefix.
Ludovic Courtès [Fri, 8 Oct 2010 08:23:52 +0000 (10:23 +0200)]
SRFI-1: Make `unfold' tail-recursive (fix bug #30071).
* module/srfi/srfi-1.scm (unfold): Make tail-recursive, following a
suggestion by Szavai Gyula.
* test-suite/tests/srfi-1.test ("unfold"): New test prefix.
Ludovic Courtès [Fri, 8 Oct 2010 07:49:12 +0000 (09:49 +0200)]
Register `scm_init_r6rs_ports' as an extension.
* libguile/r6rs-ports.c (scm_register_r6rs_ports): New function.
* libguile/r6rs-ports.h (scm_register_r6rs_ports): New declaration.
* libguile/init.c (scm_i_init_guile): Call it.
Ludovic Courtès [Fri, 8 Oct 2010 07:46:39 +0000 (09:46 +0200)]
Always run at least the ASCII regexp tests.
* test-suite/tests/regexp.test (with-ascii-or-latin1-locale): New macro.
("regexp-quote"): Use it instead of `with-latin1-locale'.
Andy Wingo [Fri, 8 Oct 2010 10:29:54 +0000 (12:29 +0200)]
fix a bug in `finish'
* module/system/repl/command.scm (repl-pop-continuation-resumer)
(finish): Fix a bug printing return values.
Andy Wingo [Fri, 8 Oct 2010 10:23:34 +0000 (12:23 +0200)]
stepping traps use frame-next-source, not frame-source
* module/system/vm/trap-state.scm (add-ephemeral-stepping-trap!): Use
frame-next-source in stepping traps.
Andy Wingo [Fri, 8 Oct 2010 10:21:20 +0000 (12:21 +0200)]
add program-sources-pre-retire to core and define frame-next-source
* libguile/programs.h:
* libguile/programs.c (scm_program_source): Add an optional arg, the
sources table to traverse. Defaults to the result of
scm_program_sources.
* module/system/vm/program.scm (program-sources-pre-retire): Move
definition here from (system vm traps), and export.
* module/system/vm/traps.scm: Adapt.
* module/system/vm/frame.scm (frame-next-source): New exported binding,
returns the source line corresponding to the next instruction instead
of the previous instruction.
Andy Wingo [Fri, 8 Oct 2010 09:53:59 +0000 (11:53 +0200)]
fix error in frame-return-values
* module/system/vm/frame.scm (frame-return-values): Fix off-by-one
error.
Andy Wingo [Fri, 8 Oct 2010 09:52:38 +0000 (11:52 +0200)]
fix tracing of mv returns
* module/system/vm/trace.scm (print-return): Fix multiple-value-return
printing.
Andy Wingo [Fri, 8 Oct 2010 09:43:15 +0000 (11:43 +0200)]
more inlining in psyntax
* module/ice-9/psyntax.scm (make-wrap, wrap-marks, wrap-subst): Use
identifier syntax here too.
* module/ice-9/psyntax-pp.scm: Regenerated.
Andy Wingo [Fri, 8 Oct 2010 09:30:17 +0000 (11:30 +0200)]
inline symbol? and vector? to opcodes, and a psyntax inlining tweak
* libguile/vm-i-scheme.c (symbol?, vector?): New
instructions. Renumbered the rest.
* libguile/vm-i-system.c: Renumber instructions.
* libguile/_scm.h (SCM_OBJCODE_MINOR_VERSION): Bump.
* module/ice-9/psyntax.scm (binding-type, binding-value): Define using
macros so that we inline to car and cdr opcodes. Oh, for an inliner :)
* module/language/tree-il/compile-glil.scm (*primcall-ops*)
* module/language/tree-il/primitives.scm
(*interesting-primitive-names*, *effect-free-primitives*)
(*effect+exception-free-primitives*): Add symbol? and vector?
inlines.
Andy Wingo [Fri, 8 Oct 2010 09:13:19 +0000 (11:13 +0200)]
add repl debugging command docs
* doc/ref/scheme-using.texi (Debug Commands): Add docs for new debugging
commands (break, step, registers, etc).
Andy Wingo [Fri, 8 Oct 2010 08:54:03 +0000 (10:54 +0200)]
proof-reading fixen in api-debug
* doc/ref/api-debug.texi (Trap States, High-Level Traps): A couple of
proof-reading fixes.
Andy Wingo [Thu, 7 Oct 2010 22:00:16 +0000 (00:00 +0200)]
finish traps documentation
* doc/ref/api-debug.texi (Low-Level Traps, Tracing Traps, Trap States):
Add notes on using modules.
(High-Level Traps): Combine "Trap Handlers" and "Setting Traps"
here. Flesh out docs.
Andy Wingo [Thu, 7 Oct 2010 20:57:07 +0000 (22:57 +0200)]
remove old debugging examples from api-debug
* doc/ref/api-debug.texi (Debugging Examples): Remove section, as the
tracing bits are adequately covered in tracing, and the breakpoints
and such will get covered in the debugging meta-commands section.
Andy Wingo [Thu, 7 Oct 2010 20:50:33 +0000 (22:50 +0200)]
document trap states
* module/system/vm/trap-state.scm: Export add-trap!.
* doc/ref/api-debug.texi (Trap States): Document.
Andy Wingo [Thu, 7 Oct 2010 20:27:00 +0000 (22:27 +0200)]
document tracing traps
* doc/ref/api-debug.texi (Tracing Traps): Document the traps.
Andy Wingo [Thu, 7 Oct 2010 20:15:35 +0000 (22:15 +0200)]
finish documenting low-level traps, other api-debug.texi fixes
* doc/ref/api-debug.texi (Stack Capture): Rename from "Capturing the
Stack or Innermost Stack Frame". Move start-stack docs here.
(Frames): Document accessors for fp, sp, ip, et al.
(Source Properties): Raise to a subsection.
(VM Hooks): Add notes about the VM trace level within hook firing.
(Low-Level Traps): Flesh out.
Andy Wingo [Thu, 7 Oct 2010 11:06:57 +0000 (13:06 +0200)]
update traps documentation (unfinished)
* doc/ref/Makefile.am:
* doc/ref/guile.texi:
* doc/ref/scheme-debugging.texi: Remove scheme-debugging.texi, which
only described tracing. Tracing documentation is now in
api-debugging.
* doc/ref/scheme-using.texi (Evaluating Scheme Code): Remove reference
to source traps, as that section is going away.
* doc/ref/api-modules.texi (Included Guile Modules): Remove reference to
Tracing. This section is a little silly, anyway...
* doc/ref/api-evaluation.texi (VM Behaviour): Remove section, it is in
api-debugging now.
* doc/ref/api-debug.texi (Stacks, Frames): Rename sections from
"Examining the Stack" and "Examining Stack Frames", respectively.
(Traps): Update for current API. A big and not-quite-finished update.
Andy Wingo [Thu, 7 Oct 2010 10:56:43 +0000 (12:56 +0200)]
update docs regarding --debug
* doc/ref/scheme-scripts.texi (Invoking Guile): Update docs on --debug.
Andy Wingo [Thu, 7 Oct 2010 10:55:37 +0000 (12:55 +0200)]
rename vm-trace to call-with-trace
* module/system/vm/trace.scm (print-application, print-return): Change
to add more whitespace, as (ice-9 debug) did.
(call-with-trace): Rename from vm-trace, and make the vm a keyword
argument.
* module/system/repl/command.scm: Don't autoload (system vm profile).
(trace): Update for call-with-trace name change.
Andy Wingo [Wed, 6 Oct 2010 19:19:49 +0000 (21:19 +0200)]
update (system repl debug) todo
* module/system/repl/debug.scm: Update todo.
Andy Wingo [Wed, 6 Oct 2010 19:19:08 +0000 (21:19 +0200)]
add ,step ,stepi ,next and ,nexti
* module/system/vm/traps.scm (trap-matching-instructions): New trap,
just installs a next hook and runs the handler when a predicate
succeeds.
* module/system/vm/trap-state.scm (add-ephemeral-stepping-trap!): New
procedure, uses trap-matching-instructions with an appropriate
predicate to handle step, stepi, next, and nexti repl metacommands.
* module/system/repl/command.scm (step, step-instruction, next)
(next-instruction): New repl debugger commands.
Andy Wingo [Wed, 6 Oct 2010 19:17:06 +0000 (21:17 +0200)]
cleanups to ,finish
* module/system/repl/command.scm (repl-pop-continuation-resumer): Factor
out of finish.
(finish): Adapt.
* module/system/vm/trap-state.scm (add-ephemeral-trap-at-frame-finish!):
Rename to add "ephemeral" to the name.
* module/system/vm/traps.scm (trap-calls-to-procedure): Remove unused
#:width kwarg.
Andy Wingo [Wed, 6 Oct 2010 08:24:45 +0000 (10:24 +0200)]
build guile-readline/ after module/
* Makefile.am (SUBDIRS): Build module/ before guile-readline/. Reformat.
Andy Wingo [Tue, 5 Oct 2010 19:53:58 +0000 (21:53 +0200)]
add ,finish repl meta-command
* module/system/repl/command.scm (finish): New REPL meta command. Uses
fancy prompt stuff.
Andy Wingo [Tue, 5 Oct 2010 19:53:29 +0000 (21:53 +0200)]
(system vm trap-state): add-trap-at-frame-finish!
* module/system/vm/traps.scm: Fix a comment.
* module/system/vm/trap-state.scm (<trap-state>): Add next-ephemeral-idx
slot.
(wrapper-at-index): Use eqv? instead of = to avoid type errors in user
inputs.
(next-ephemeral-index!, ephemeral-handler-for-index): New functions,
allocate ephemeral trap ids for functions to be called only once.
(add-trap-at-frame-finish!): New export, traps when a frame finishes.
Andy Wingo [Tue, 5 Oct 2010 19:51:44 +0000 (21:51 +0200)]
(system vm frame): frame-return-values
* module/system/vm/frame.scm (frame-return-values): New exported
function, gives the return values for a frame.
* module/system/vm/trace.scm: Remove frame-return-values from here.
Andy Wingo [Tue, 5 Oct 2010 19:50:57 +0000 (21:50 +0200)]
debug-trap-handler ephemeral trap enhancement
* module/system/repl/error-handling.scm (call-with-error-handling): If
the given index is false, assume this was an ephemeral trap, and don't
print a welcome message or reference the trap by index.