1 2006-01-26 Matthias Clasen <mclasen@redhat.com>
3 * glib/gmem.c (g_mem_init_nomessage): Fix C99isms. (#328705,
6 Wed Jan 25 19:16:57 2006 Tim Janik <timj@imendio.com>
8 * fixed buglets reported by Jens Granseuer in #328254.
10 * configure.in: free the memory allocated in posix_memalign() tests.
12 * glib/gslice.c: spelling fixes.
14 Wed Jan 25 16:39:18 2006 Tim Janik <timj@imendio.com>
16 * glib/gslice.c: honour g_mem_gc_friendly settings when freeing
17 slices, make sure g_mem_gc_friendly is properly initialized.
19 * gmem.[hc]: ensure g_mem_gc_friendly is initialized from G_DEBUG upon
20 the first allocation. applied some branching optimizations.
22 * docs/macros.txt: reflected --enable-gc-friendly change and
23 described ENABLE_GC_FRIENDLY_DEFAULT as well as G_DEBUG=gc-friendly.
25 * configure.in: changed --enable-gc-friendly=yes to define
26 ENABLE_GC_FRIENDLY_DEFAULT.
28 * glib/garray.c: changed ENABLE_GC_FRIENDLY macro #ifdef-s to
29 if (G_UNLIKELY (g_mem_gc_friendly)).
32 * glib/ghash.c: removed ENABLE_GC_FRIENDLY code which is now taken
33 care of by g_slice_free1().
35 * tests/slice-test.c: fixed leaks, reported by Kjartan Maraas.
37 Tue Jan 24 17:49:36 2006 Tim Janik <timj@imendio.com>
39 * glib/gslice.c: only use posix_memalign() if it's known to work,
40 revert to memalign() otherwise.
42 * configure.in: check for broken posix_memalign() implementations
45 2006-01-24 Matthias Clasen <mclasen@redhat.com>
47 * tests/unicode-encoding.c: Use UTF-16LE as target encoding
48 on all little-endian systems. (#143380, Marc Moorcroft)
50 2006-01-23 Matthias Clasen <mclasen@redhat.com>
52 * configure.in: Change the shared libary extension for hpux-ia64
53 to so. (#328253, The Written Word)
55 Mon Jan 23 17:30:33 2006 Tim Janik <timj@imendio.com>
57 * glib/gutils.c (_g_getenv_nomalloc): wiped out all the wonderfull
58 G_OS_WIN32 code i wrote ;-[) after tml told me windows has getenv()
61 Mon Jan 23 16:46:20 2006 Tim Janik <timj@imendio.com>
63 * glib/gslice.c (slice_config_init): initialize GSlice config from
64 G_SLICE environment variable. we support G_SLICE=always-malloc
65 currently, which forces all g_slice_*() allocations to use the system
69 g_parse_debug_string(): added a note about not using g_malloc() here.
70 _g_getenv_nomalloc(): getenv() variant that doesn't use g_malloc or
71 g_slice. contains only guesswork in the WIN32 branch.
73 2006-01-18 Matthias Clasen <mclasen@redhat.com>
77 * === Released 2.9.4 ===
81 2006-01-17 Matthias Clasen <mclasen@redhat.com>
83 * glib/gkeyfile.c (g_key_file_get_groups): Correct a g_new0()
84 call. (#327421, Morten Welinder)
86 2006-01-17 Sebastian Wilhelmi <seppi@seppi.de>
88 * glib/gthreadpool.c: To avoid deadlocks get rid of the settings
89 G_LOCK. Use the unused_thread_queue lock instead. Change
90 g_thread_pool_thread_proxy such that threads only wait on
91 non-exlusive pools for at most a 1/2 second. Do not reorder tasks
92 due to superfluous tasks. Global tasks wait at most for
93 max-idle-time milliseconds. Make sure, that no task is woken up
94 twice for the same event via a wakeup_serial. This fixes #324228.
96 * tests/threadpool-test.c: Adapt test accordingly. Do not pass
97 invalid NULL into the thread pools. This as well fixes #327290.
99 2006-01-16 Matthias Clasen <mclasen@redhat.com>
101 * configure.in: Bump version
103 * === Released 2.9.3 ===
107 * glib/gthread.h (g_static_mutex_get_mutex_impl_shortcut):
108 Rework the strict aliasing fix to not break C++, pointed
109 out by Murray Cumming.
111 * glib/gasyncqueue.c (g_async_queue_push_sorted_unlocked):
112 Signal waiting threads, problem noticed by Christian Kellner.
114 2006-01-16 Matthias Clasen <mclasen@redhat.com>
116 Fix bug 326747, Alberto Ruiz:
118 * glib/gconvert.c (make_valid_utf8): Change this function to
119 replace unknown characters by the Unicode replacement character
120 instead of '?', and don't append "(invalid encoding)".
121 (g_filename_display_name, g_filename_display_basename): Document
122 how to determine if the filename was in an invalid encoding.
124 2006-01-14 Matthias Clasen <mclasen@redhat.com>
126 * glib/gtree.c: Replace the simple recursive implementation
127 by a nonrecursive, threaded implementation by Maurizio
130 2006-01-12 Matthias Clasen <mclasen@redhat.com>
132 * tests/asyncqueue-test.c (main): Fix the
133 assert to not always trigger. (#326558,
136 2006-01-11 Matthias Clasen <mclasen@redhat.com>
138 * tests/convert-test.c: Don't test items_read and
139 items_written in error cases where it is not set.
141 * tests/asyncqueue-test.c: Set error to NULL.
143 2006-01-08 Matthias Clasen <mclasen@redhat.com>
145 * glib/gtree.c: Break some long lines.
147 * tests/tree-test.c: Add more tests.
149 * glib/gslice.c: Include config.h first, pointed out
152 2005-12-20 Sven Herzberg <herzi@gnome-de.org>
154 * docs/reference/gobject/tmpl/gtype.sgml: explain that
155 G_TYPE_INSTANCE_GET_CLASS() does behave different during
158 2006-01-06 Matthias Clasen <mclasen@redhat.com>
160 * glib/gslice.c: Only define _XOPEN_SOURCE if we know
161 that we have posix_memalign(). (#323937, Bogdan Nicula)
163 2006-01-05 Matthias Clasen <mclasen@redhat.com>
165 * configure.in: Bump version
167 * === Released 2.9.2 ===
169 2006-01-05 Hans Breuer <hans@breuer.org>
171 * glib/makefile.msc.in : added gslice.obj
173 2006-01-05 Matthias Clasen <mclasen@redhat.com>
177 * glib/gmain.c: Clarify the documentation of source ids
178 a bit. (#325874, Dan Williams)
180 * configure.in: Fix another strict aliasing problem.
182 * glib/gthreadpool.c: Avoid use of varargs macro. (#325864,
185 2006-01-04 Manish Singh <yosh@gimp.org>
188 * glib/gstdio.h: don't macroized g_access, g_chdir, and g_unlink
189 either, since they have the same issue as g_rmdir. (Related to
192 Wed Jan 4 13:33:25 2006 Tim Janik <timj@gtk.org>
194 * glib/gslice.c (magazine_cache_pop_magazine): don't reverse chunk
195 order when creating magazines, so we hand out chunks with ascending
198 2006-01-04 Abel Cheung <maddog@linuxhall.org>
200 * configure.in: Added "ml" "zh_HK" to ALL_LINGUAS.
202 2006-01-03 Martyn Russell <martyn@imendio.com>
204 * docs/reference/glib/glib-sections.txt:
206 * glib/gthreadpool.[ch]:
207 - Added new API g_thread_pool_get_idle_time() and
208 g_thread_pool_set_idle_time(). (#324228).
210 * tests/threadpool-test.c:
211 - Updated test case to do thread pool sorting, thread pool with
212 no sorting and a thread pool with idle thread timeouts.
214 2006-01-03 Matthias Clasen <mclasen@redhat.com>
216 * glib/gmain.h: Add new functions here, too.
218 * glib/glib.symbols: Add new functions.
220 2005-12-20 Michael Meeks <michael.meeks@novell.com>
222 * glib/gmain.c (g_main_context_is_owner): new method
223 to determine if the current thread is the owner of the
226 2006-01-02 Matthias Clasen <mclasen@redhat.com>
230 * glib/gstdio.c (g_rmdir): Don't provide g_rmdir() as a macro
231 expanding to rmdir, since rmdir is not declared in a portable
232 system header we can include in gstdio.h. (#325249, Jani Monoses)
234 2006-01-01 Tor Lillqvist <tml@novell.com>
236 * glib/gspawn-win32.c (g_spawn_sync_utf8): Set the GIOChannels for
237 stdout and stderr to unbuffered. Otherwise the giochannel layer
238 will try to read from them regardless whether the
239 g_io_channel_win32_poll() call here has indicated
240 readability or not. (#325310)
242 2005-12-29 Matthias Clasen <mclasen@redhat.com>
244 * glib/gutils.c (glib_check_version): Fix a copy-and-paste error
245 in the docs. (#325273, Declan Naughton)
247 2005-12-27 Manish Singh <yosh@gimp.org>
249 * tests/run-collate-tests.sh: use LC_ALL instead of LC_COLLATE, to
250 make sure we really override things.
252 2005-12-27 Matthias Clasen <mclasen@redhat.com>
254 Fix #316221, Michal Benes, Stanislav Brabec;
256 * configure.in: Fix a strict aliasing problem in
257 g_static_mutex_get_mutex().
258 * glib/gthread.h: ...and in
259 g_static_mutex_get_mutex_impl_shortcut().
261 * glib/gdatasetprivate.h: Add a cast to silence compiler
262 warnings. (#321978, Andrew Paprocki)
264 Partial fix for bug #323937, Bogdan Nicula.
266 * configure.in: Check for malloc.h
267 * glib/gslice.c: Don't include malloc.h unconditionally.
269 2005-12-27 Manish Singh <yosh@gimp.org>
271 * tests/run-collate-tests.sh: export LC_COLLATE so it takes.
272 (#324950, Dan Yefimov)
274 2005-12-26 Matthias Clasen <mclasen@redhat.com>
276 * glib/gslice.c: On Win32, include process.h (#325015, Kazuki
279 2005-12-25 Matthias Clasen <mclasen@redhat.com>
281 * AUTHORS: Update my email
283 * tests/slice-test.c: Fix C99isms. (#324950, Dan Yefimov)
285 2005-12-21 Matthias Clasen <mclasen@redhat.com>
287 * README.in: Add some notes on when 'make check' may fail.
289 2005-12-20 Matthias Clasen <mclasen@redhat.com>
292 * glib/gthreadpool.h:
293 * glib/gthreadpool.c (g_thread_pool_set_sort_function): New function
294 to sort tasks pushed into a threadpool. (#324479, Martyn Russell)
296 * tests/threadpool-test.c: Test this.
298 Tue Dec 20 18:14:14 2005 Tim Janik <timj@imendio.com>
300 * glib/gslice.[hc]: added mem_error() and mem_assert() to test and
301 handle errors without depending on gmessage.c which might not be
302 setup when the error occours.
303 removed G_SLICE_CONFIG_ALWAYS_FREE config option, fixed the code so
304 always freeing can be achieved by adjusting the working set time to
305 0 with G_SLICE_CONFIG_WORKING_SET_MSECS.
306 added G_SLICE_CONFIG_COLOR_INCREMENT to test different color increments
307 (mainly 0 and 1). reduced the minimum block size to 128 bytes, to
308 minimize wastage if small amounts of differently sized structrues are
309 allocated, this does come at a performance cost of roughly 5% though.
310 fixed up block alignment calculation, so it works for varying
311 block sizes. only use strerror() not g_strerror() since the latter
312 depends on working GQuark and GSlice.
313 mem_error(): implemented in terms of fprintf and vfprintf.
315 * tests/slice-color.c: new program to test cache colorization effects.
317 * tests/slice-test.c: trade G_SLICE_CONFIG_ALWAYS_FREE for 0 duration
318 G_SLICE_CONFIG_WORKING_SET_MSECS.
320 2005-12-17 Matthias Clasen <mclasen@redhat.com>
322 * glib/goption.c (parse_short_option): Set an error in all
323 failure cases. (#324332, Tim-Philipp Müller)
325 2005-12-17 Sebastian Wilhelmi <seppi@seppi.de>
327 * glib/gatomic.c: Fix memory barrier position in g_atomic_int_get
328 and g_atomic_pointer_get. Add g_atomic_int_set and
329 g_atomic_pointer_set implementations for the !DEFINE_WITH_MUTEXES &&
330 G_ATOMIC_OP_MEMORY_BARRIER_NEEDED case, as well as defining them
331 as functions (additionally to the macros in the header) for the
332 !G_ATOMIC_OP_MEMORY_BARRIER_NEEDED case.
334 2005-12-16 Matthias Clasen <mclasen@redhat.com>
336 * glib/gmem.c (g_allocator_new): Don't return a pointer to
337 a const struct, since apps expect to be able to modify it.
338 (#324179, J. Ali Harlow)
340 Tue Dec 13 10:13:32 2005 Tim Janik <timj@imendio.com>
342 * glib/gatomic.h: added g_atomic_pointer_set() and g_atomic_int_set()
343 for systems where the initialization of atomic variables requires a
344 write memory barrier.
346 2005-12-09 Matthias Clasen <mclasen@redhat.com>
348 * configure.in: Bump version
350 * === Released 2.9.1 ===
352 2005-12-09 Alexander Larsson <alexl@redhat.com>
356 Use G_GNUC_WARN_UNUSED_RESULT on list functions that return
359 2005-12-08 Matthias Clasen <mclasen@redhat.com>
363 2005-12-07 Martyn Russell <martyn@imendio.com>
365 * glib/gasyncqueue.c:
366 - Call g_queue_insert_sorted() instead of duplicating the code.
367 - Call g_queue_sort() instead of duplicating the code.
368 - Invert sort function results to make sure the same sort function
369 gives the same results across glist, gslist, gqueue and
372 * tests/asyncqueue-test.c:
373 - Updated the sort function to reflect the example in the
374 documentation for gasyncqueue.c.
376 2005-12-07 Martyn Russell <martyn@imendio.com>
378 * docs/reference/glib/glib-sections.txt:
379 * docs/reference/glib/tmpl/linked_lists_double.sgml:
380 * docs/reference/glib/tmpl/linked_lists_single.sgml:
383 - Added g_list_insert_sorted_with_data () and
384 g_slist_insert_sorted_with_data ().
385 - Removed the extra check in g_list_sort() and g_slist_sort() for
386 GCompareDataFunc vs. GCompareFunc.
388 2005-12-07 Tor Lillqvist <tml@novell.com>
390 * glib-gettextize.in: Look up prefix at run-time on Win32,
391 assuming the standard directory structure with glib-gettextize in
394 2005-12-06 Matthias Clasen <mclasen@redhat.com>
396 * glib/gmem.h: Exempt GTK+ from the mem chunk deprecation, since
397 we need GTK+ 2.8 to compile against GLib 2.10.
399 2005-12-05 Manish Singh <yosh@gimp.org>
401 * tests/run-collate-test.sh: set LC_COLLATE, not LANG, to be sure
402 to override any user settings.
404 2005-12-05 Matthias Clasen <mclasen@redhat.com>
406 * configure.in: Revert an accidential commit
408 2005-12-05 Matthias Clasen <mclasen@redhat.com>
410 * glib/gutf8.c: Documentation updates. (#323291, Morten Welinder)
412 * tests/Makefile.am (TESTS_ENVIRONMENT):
413 * tests/gobject/Makefile.am (TESTS_ENVIRONMENT): Set
414 MALLOC_CHECK_ and MALLOC_PERTURB_
416 * tests/run-collate-tests.sh: Run the collation tests explicitly
417 in en_US locale. (#320463)
419 * glib/gmem.h: Really deprecate GMemChunk.
426 * glib/gstrfuncs.c: Add versioned deprecation docs.
428 2005-12-05 Martyn Russell <martyn@imendio.com>
430 * docs/reference/glib/glib-sections.txt:
431 * glib/gasyncqueue.[ch]:
432 - Added support for sorting async queues by with _push_sorted(),
433 _push_sorted_unlocked(), _sort() and _sort_unlocked() (#323047).
436 * tests/asyncqueue-test.c:
437 - Added test case for gasyncqueue.c
439 Mon Dec 5 15:53:20 2005 Tim Janik <timj@imendio.com>
441 * glib/gslice.c: implement chain walking for arbitrary ->next pointer
442 offsets in g_slice_free_chain_with_offset() based on a patch by behdad
443 in bug 323178. moved time consuming logic from g_slice_free() out of
444 the inner loop, so g_slice_free_chain_with_offset() provides a real
445 performance benefit over g_slice_free1() now.
447 * glib/gslice.h: renamed g_slice_free_chain() to
448 g_slice_free_chain_with_offset(). implemented g_slice_free_chain() as
449 a type-safe macro as suggested in bug 323178.
450 simplified the macro implementation of g_slice_free() and implemented
451 it in a type safe manner for all compliers as suggested by Morten
452 Welinder <mortenw@gnome.org>.
457 * glib/glib.symbols: s/g_slice_free_chain/g_slice_free_chain_with_offset/
459 2005-12-05 Matthias Clasen <mclasen@redhat.com>
461 * glib/gasyncqueue.c: Add some docs.
463 * tests/libmoduletestplugin_a.c: Fix compiler warnings.
465 * glib/gatomic.c: In the ia64 implementation, use
466 __sync builtin without _si or _di suffix. (#321229,
467 Stanislav Brabec, patch by Andreas Schwab)
469 2005-12-04 Behdad Esfahbod <behdad@gnome.org>
471 * glib/gslice.h: Remove comma at the end of enum.
473 2005-12-04 Matthias Clasen <mclasen@redhat.com>
475 Handle multiple user names with the same UID better.
476 (#319535, Laszlo Peter)
478 * glib/gutils.c (g_get_any_init_do): When determining user
479 data, first look up $LOGNAME. If the UID doesn't match
480 getuid(), fall back to the current behaviour of looking
481 up the user data based on getuid().
483 2005-12-04 Matthias Clasen <mclasen@redhat.com>
485 * glib/gmacros.h (G_GNUC_WARN_UNUSED_RESULT): Add a macro
486 to make gcc warn if a function result is ignored. (#145466,
487 Arjan van de Ven, Alex Larsson)
489 * glib/gmem.h: Add the new attribute to g_realloc and
492 2005-12-03 Matthias Clasen <mclasen@redhat.com>
495 * glib/gcache.h: Deprecate g_cache_value_foreach. (#322956,
498 * glib/gmappedfile.c: Make mapping of empty files
501 * glib/gfileutils.c: Don't fork a new process just to
502 fix the permissions of the created temp file. (#321318,
503 Alexis S. L. Carvalho)
505 2005-12-02 Matthias Clasen <mclasen@redhat.com>
507 * README.in: Add a note about Solaris threads.
510 * configure.in: Remove support for Solaris threads.
511 (#136971, Sebastian Wilhelmi, patch by Andrew Paprocki)
513 2005-12-02 Matthias Clasen <mclasen@redhat.com>
515 * configure.in: Define G_GUINT64_CONSTANT in analogy to
516 G_GINT64_CONSTANT. (#322568, Andrew Paprocki)
518 2005-12-02 Matthias Clasen <mclasen@redhat.com>
520 * glib/gslice.c: Win32 portability fixes and C99ism removal,
521 pointed out by Kazuki Iwamoto. (#323052)
523 2005-12-02 Matthias Clasen <mclasen@redhat.com>
525 * glib/gatomic.c (g_atomic_int_add): Add a missing volatile for
526 the IA64 implementation.
528 Fri Dec 2 16:18:09 2005 Tim Janik <timj@imendio.com>
530 * glib/gslice.c: some naming and type size fixups.
532 Fri Dec 2 13:08:58 2005 Tim Janik <timj@gtk.org>
534 * glib/gslice.h (g_slice_free): reworked GCC-specific type-safe macro
535 variant into something less verbose, but digestible for gcc-3.4.
537 Fri Dec 2 10:55:07 2005 Tim Janik <timj@imendio.com>
539 * tests/slice-test.c: extended to perform the benchmarking on the old
540 memchunk code if 'O' is selected.
542 * tests/memchunks.c: new file which contains the old GLib mem chunks
543 implementation with prefix old_mem_chunk_.
545 * tests/Makefile.am: added memchunks.c
547 Fri Dec 2 00:16:59 2005 Tim Janik <timj@imendio.com>
549 * glib/gslice.c: improved gettimeofday() timer resolution by maintaining
550 acceptable load distribution of the syscall.
551 (allocator_get_magazine_threshold): reduce minimum/single-thread magazine
552 sizes to approximately page_size/2.5 to avoid excessive startup allocations.
554 2005-12-01 Matthias Clasen <mclasen@redhat.com>
556 * tests/gobject/ifacecheck.c (test_iface_base_init): Don't
557 double-free base interface members.
559 * tests/Makefile.am: Remove duplicate variable.
561 Thu Dec 1 17:32:46 2005 Tim Janik <timj@imendio.com>
563 * glib/gslice.[hc]: new slice allocator implementation.
565 * tests/slice-test.c: added random slice allocation test.
567 * glib/gthread.[hc]: removed newly added private thread mem API.
569 * glib/gthreadinit.h:
572 * glib/gmem.c: divided glib threading initialisation into three phases,
573 initialisation where private keys and messaging are not available (only
574 needed by gmem.c), initialisation without messaging but private keys
575 available (gslice.c, gmessage.c), and full fledged initialisers that
576 server the rest of glib. initialisation functions got renamed to reflect
577 the limitations of their corresponding phases.
579 * glib/gmem.c: removed memchunk code, defer allocations to
582 * glib/gmem.[hc]: removed g_slice_* skeletons.
584 * glib/glib.symbols: added g_slice_* symbols.
586 * configure.in: check for availability of posix_memalign(3), memalign(3)
589 * glib/Makefile.am: added gslice.[hc].
591 2005-12-01 Tor Lillqvist <tml@novell.com>
593 * glib/gstdio.c (g_stat): In the Win32 implementation, strip
594 trailing slash(es) for non-root folders. stat() fails if non-root
595 folders are specified with trailing slashes. It's too much hassle
596 to demand that callers strip such slashes themselves, especially
597 as it is easy to get it wrong and strip the slash of a root
599 (g_rename): On NT-based Windows, use MoveFileEx() with
600 MOVEFILE_REPLACE_EXISTING to better match Unix behaviour.
602 2005-11-28 Matthias Clasen <mclasen@redhat.com>
604 Fix G_STMT_START / G_STMT_END on Solaris. (#321972,
607 * configure.in: Check whether do { } while (0) works.
609 * glib/gmacros.h: Use do { } while (0) for G_STMT_START /
610 G_STMT_END if it works.
612 2005-11-28 Matthias Clasen <mclasen@redhat.com>
614 * glib/gthread.c (g_static_rw_lock_wait, g_static_rw_lock_signal):
615 * glib/gnode.c (g_node_depth_traverse_level):
616 * glib/gmem.c (g_allocator_new):
617 * glib/ghash.c (g_hash_table_unref, g_hash_table_destroy)
618 (g_hash_table_foreach_remove_or_steal): Silence compiler
621 2005-11-27 Matthias Clasen <mclasen@redhat.com>
623 * glib/gunicollate.c (g_utf8_collate_key_for_filename):
624 Don't read beyond len. (#322520, Christian Persch)
626 Wed Nov 23 17:34:01 2005 Tim Janik <timj@imendio.com>
628 * glib/gdataset.c: access datalist flags via atomic pointer access
629 functions, instead of acquiring the dataset lock. this is faster and
630 also matches the atomic pointer readouts in gdatalistprivate.h and
631 direct pointer modifications required by gobject.c.
633 Wed Nov 23 13:35:31 2005 Tim Janik <timj@gtk.org>
635 * glib/gdataset.c: streamlined initialization code somewhat,
636 removed GData node cache.
638 2005-11-22 Matthias Clasen <mclasen@redhat.com>
640 * glib/ghash.c (g_hash_table_ref):
641 (g_hash_table_unref): Mark these as new API.
643 Avoid double locking in g_intern_string (#322133,
646 * glib/gdataset.c (g_quark_from_string_internal): New
647 internal function which factors out common parts of
648 g_quark_from[_static]_string.
649 (g_quark_from_string, g_quark_from_static_string):
650 Use g_quark_from_string_internal.
651 (g_intern_string, g_intern_static_string): Use
652 g_quark_from_string_internal, and only take the
654 (g_quark_new): Don't store the strings shifted by -1
655 in the g_quarks array.
656 (g_quark_to_string): Adapt to the previous change.
658 Tue Nov 22 14:04:26 2005 Tim Janik <timj@imendio.com>
662 g_hash_table_new_full(): create hash tables with a ref count of 1.
663 g_hash_table_ref(): atomically ref_count+=1
664 g_hash_table_unref(): atomically ref_count-=1, destroys hash table
665 when refcount reaches 0.
666 g_hash_table_destroy(): just destroy keys and values, unref by 1.
667 g_hash_table_insert():
668 g_hash_table_replace(): assert ref_count>0.
671 * glib/gatomic.c: added 'volatile' qualifier to all atomic pointer and
674 2005-11-20 Behdad Esfahbod <behdad@gnome.org>
676 * glib/guniprop.c (g_unichar_get_mirror_char): Remove unused
677 variables. (#321984, Andrew Paprocki)
679 2005-11-18 Matthias Clasen <mclasen@redhat.com>
681 * configure.in: Bump version
683 2005-11-17 Matthias Clasen <mclasen@redhat.com>
685 * === Released 2.9.0 ===
687 * glib/glib.symbols: Updates
689 * glib/gutf8.c (g_utf8_offset_to_pointer): Add some performance
690 hints to the docs. (Paolo Borelli)
694 * glib/gthread.c (g_thread_foreach): Mark as new api.
696 * README.in: Updates.
700 * glib/gdate.c (g_date_set_time_t):
701 (g_date_set_time_val): Add functions to set a date from
702 a time_t and from a GTimeVal.
703 (g_date_set_time): Deprecate. (#314067, Roger Leigh)
706 * tests/utf8-pointer.c: Unit tests for g_utf8_pointer_to_offset
707 and g_utf8_offset_to_pointer.
709 * glib/gutf8.c (g_utf8_pointer_to_offset)
710 (g_utf8_offset_to_pointer): Handle negative offsets, and use
711 "stutter stepping" for going backwards. (#320638, Larry
716 * glib/gthread.c: const correctness fixes, found
717 by Arjan van de Ven and gcc.
719 2005-11-16 Behdad Esfahbod <behdad@gnome.org>
721 * docs/reference/glib/tmpl/unicode.sgml: Correct typo on mentioning
722 @G_UNICODE_BREAK_UNKNOWN that should be #G_UNICODE_BREAK_UNKNOWN.
724 2005-11-09 Behdad Esfahbod <behdad@gnome.org>
726 * glib/guniprop.c: Use bit hacks instead when checking a general
727 category value against multiple values.
729 * glib/gutf8.c: Change ISO10646 to Unicode in docs.
731 2005-11-10 Simos Xenitellis <simos@gnome.org>
733 * configure.in: Added tt (Tatar) to ALL_LINGUAS.
735 2005-11-05 Tor Lillqvist <tml@novell.com>
737 * glib/giowin32.c (g_io_win32_set_flags): Don't call g_warning().
738 (#320688) Instead set error and return failure.
740 2005-11-07 Matthias Clasen <mclasen@redhat.com>
742 * glib/gmain.c (g_main_dispatch): Don't call cb_funcs->unref
743 while holding the context lock. (#320886, Andy Wingo)
745 2005-11-07 Billy Biggs <vektor@dumbterm.net>
747 * tests/refcount/Makefile.am: Add a missing reference to libglib
748 in the LDADD for this test.
750 2005-11-07 Matthias Clasen <mclasen@redhat.com>
752 * glib/gmem.c (g_allocator_new): We need to set n_preallocs to a
753 nonzero value, otherwise GTK+ 2.8 breaks when compiled against
754 GLib 2.9. (#320755, Luca Ferretti)
756 2005-11-04 Matthias Clasen <mclasen@redhat.com>
758 * glib/Makefile.am: Apply a patch from Behdad Esfahbod to
759 use a faster lookup table for g_unichar_get_mirror_char().
760 * glib/gmirroringtable.h: The new table.
761 * glib/gunichartables.h: Remove bidi_mirroring_table.
762 * glib/gen-unicode-tables.pl: Don't generate the mirroring
764 * glib/glib-mirroring-tab/*: Sources for the program
765 which generated gmirroringtable.h.
767 * glib/glist.c: Avoid some code duplication.
769 * glib/gscanner.h: Include gdataset.h. (#320322)
771 * glib/gdebug.h: add new GDebugFlag for fatal_criticals
772 * glib/gmessages.c: (_g_debug_init): handle G_DEBUG=fatal_criticals,
773 to help find critical warnings in applications. (#320017,
776 2005-11-02 Tor Lillqvist <tml@novell.com>
778 * glib/glib.symbols: Remove large amount of trailing whitespace
779 from one line. Remove #ifdef G_UNDEFINED from around
780 g_hash_table_new and g_hash_table_new_full symbols, what was the
781 purpose of that? Add the g_slice functions.
783 Tue Nov 1 16:24:20 2005 Tim Janik <timj@imendio.com>
785 * glib/gmem.[hc]: prepared deprecation of GMemChunk and GAllocator.
786 added g_slice_*() API to allocate and cache small bits of memory.
787 an actuall allocator implementation for g_slice_*() is still pending.
789 * glib/gthread.[hc]: changes from a patch by Matthias Clasen.
790 changed GRealThread list to use in-structure *next; fields instead
791 of GSList, in order for thread iteration to not depenend on g_slice_*()
793 _g_thread_mem_private_get():
794 _g_thread_mem_private_set(): added accessors for private memory,
795 needed because the ordinary GPrivate implementation relies on GArray
796 and GSList and therefore indirectly on working g_slice_*() allocations.
799 g_thread_foreach(): new public API function to loop over all existing threads.
813 removed GAllocator and free list usages and accompanying locks.
814 use g_slice_*() API to allocate and cache small bits of memory.
816 * glib/ghook.h: removed GMemChunk field from public API.
819 * glib/glist.h: deprecate allocator API, provide _free1() for consistency.
821 * glib/gnode.h: deprecate allocator API.
823 * glib/gmain.c: reordered GPollRec fields so g_slice_free_chain() can
824 be used for poll rec lists.
826 * glib/grel.c: removed mem chunk usage, and allocated tuples via g_slice_*().
827 g_relation_destroy(): free all tuples from the all_tuples hash table,
828 this effectively maintains the life time track keeping of tuples.
829 g_relation_delete_tuple(): free tuples which are removed from the
830 all_tuples hash table. this fixes a temporary leak that was present
831 in the memchunk code until the destruction of the relation.
833 2005-10-29 Matthias Clasen <mclasen@redhat.com>
835 * tests/convert-test.c: Add some tests for conversions between
836 UTF-8, UCS-4 and UTF-16.
838 * glib/gutf8.c (g_utf8_to_ucs4, g_utf8_to_utf16): Fix handling
839 of len == -1, noticed by Morten Welinder.
841 2005-10-27 Erdal Ronahi <erdal.ronahi@gmail.com>
843 * configure.in: Added ku (Kurdish) to ALL_LINGUAS
845 2005-10-26 Matthias Clasen <mclasen@redhat.com>
847 * glib/gutf8.c (g_ucs4_to_utf8): Don't set items_read twice
848 in the error case, and add some documentation. (#319806, Morten
851 2005-10-19 Manish Singh <yosh@gimp.org>
853 * configure.in: Use AC_CHECK_FUNCS for _NSGetEnviron, to get the
854 config.h symbol automatically. Fixes bug #313731.
856 2005-10-19 Tor Lillqvist <tml@novell.com>
858 * glib/gwin32.c (g_win32_get_package_installation_directory):
859 Return a g_strdup()ed copy of the value stored in the hash table,
860 so that it can be g_free()d without leaving a dangling pointer in
861 the hash table. (#319232)
863 2005-10-06 Matthias Clasen <mclasen@redhat.com>
865 * glib/gunicollate.c (g_utf8_collate_key_for_filename): Handle
866 all-zero sequences correctly. (#317930, Sebastien Bacher)
868 2005-10-05 Matthias Clasen <mclasen@redhat.com>
870 * glib/gmarkup.c (g_markup_escape_text): Doc addition
872 * Makefile.am: Add Collation tests.
874 * tests/collate/*: Inputs and expected outputs for collation tests.
876 * tests/run-collate-tests.sh: Script to run collation tests.
878 * tests/unicode-collate.c (main): Rework slightly to make
879 it usable in unit tests. Also test g_utf8_collate_key_for_filename().
881 2005-10-01 Behdad Esfahbod <behdad@gnome.org>
883 * docs/reference/glib/tmpl/unicode.sgml:
884 * glib/gen-unicode-tables.pl:
886 * glib/gunichartables.h:
888 * tests/casefold.txt:
889 * tests/casemap.txt: Updated to Unicode 4.1. There are five new
890 GUnicodeBreakType types. That may break some applications, like
893 2005-09-26 Matthias Clasen <mclasen@redhat.com>
895 * glib/gstrfuncs.c (g_ascii_strtoull): Add details to the
896 docs. (#314393, Matthew F. Barnes)
899 * glib/gprintf.h: Remove g_snprintf() and g_vsnprintf(), since
900 they are already declared in glib.h. This doesn't break documented
901 use of gprintf.h, but should probably be pointed out in the
902 release notes for 2.10. (#314232, Behdad Esfahbod)
904 Tue Sep 20 13:16:04 2005 Tim Janik <timj@imendio.com>
906 * glib/gpattern.c (g_pattern_ph_match): applied significant recursion
907 complexity optimization, based on a patch from Matthias Clasen.
909 * tests/patterntest.c: more tests, mostly from matthias.
911 2005-09-20 Matthias Clasen <mclasen@redhat.com>
913 * glib/gqueue.c (g_queue_insert_sorted): Correct the docs.
914 (#316703, Mark Drago)
916 Mon Sep 19 17:23:23 2005 Tim Janik <timj@imendio.com>
918 * glib/gpattern.c: applied a patch from matthias which checks on the
919 upper bound of GPatternSpec length to optimize matches.
922 * tests/patterntest.c: added more match cases.
924 2005-09-16 Tor Lillqvist <tml@novell.com>
926 * glib/gstrfuncs.c (g_ascii_strcasecmp, g_ascii_strncasecmp): Add
927 warning to doc comment that these functions should not be used on
928 encodings like CP932.
930 2005-09-14 Matthias Clasen <mclasen@redhat.com>
932 * tests/keyfile-test.c: Add a test for grup names of length 1.
934 * glib/gkeyfile.c (g_key_file_line_is_group): Accept group names
935 of length 1. (#316309)
937 2005-09-12 Matthias Clasen <mclasen@redhat.com>
939 * glib/gmarkup.c (g_markup_escape_text): Clarify docs.
941 2005-09-11 Sebastian Wilhelmi <seppi@seppi.de>
943 * tests/refcount/Makefile.am (INCLUDES): Link the the refcount
944 tests to the system thread library $(G_THREAD_LIBS). Fixes #313744
947 2005-09-11 Kjartan Maraas <kmaraas@gnome.org>
949 * glib/gmain.c: (g_child_watch_prepare), (g_child_watch_check),
950 (child_watch_helper_thread): Remove some dead code. Closes
953 2005-09-07 Tor Lillqvist <tml@novell.com>
955 * glib/Makefile.am: Create also a console version of the
956 gspawn-win32-helper program, gspawn-win32-helper-console.exe.
957 It's otherwise identical to gspawn-win32-helper.exe, except marked
958 as a console application (linked without the -mwindows option).
960 * glib/gspawn-win32.c (do_spawn_directly, do_spawn_with_pipes):
961 Drop the dont_wait parameter. Its truth value correlated 100% with
962 the NULLness of the exit_status parameter anyway, so it's enough
963 to check whether exit_status is NULL. Invert the sense of the
964 dont_return_handle parameter and rename it to do_return_handle, to
965 make the code easier to read by avoiding double negations.
967 (g_spawn_sync_utf8, g_spawn_async_with_pipes_utf8): Modify calls
968 to do_spawn_with_pipes() accordingly.
970 (do_spawn_with_pipes): If we have a console, use the console
971 version of the helper program, otherwise use the GUI one. This
972 avoids extra console windows opening up in some situations. (In
973 case a console application uses the GUI gspawn-win32-helper.exe to
974 spawn another console application we would get a separate console
975 for the spawned console application).
977 * glib-zip.in: Distribute also gspawn-win32-helper-console.exe.
979 2005-09-05 Matthias Clasen <mclasen@redhat.com>
981 * glib/gmappedfile.c (g_mapped_file_new): Report an error
982 if the file is too large. (#315275, Kjartan Maraas)
984 * glib/gkeyfile.c (g_key_file_load_from_fd): The return value
985 of read() is signed. (#315273, Kjartan Maraas)
987 2005-08-31 Tor Lillqvist <tml@novell.com>
989 * glib/gutils.h: Wrapping atexit() is a bad idea on Windows, where
990 the EXE and each DLL have their own atexit function chains.
992 #define g_atexit as atexit instead. This means it has a
993 better chance of doing what the caller wants. For instance,
994 gtkhtml calls g_atexit() registering a function in gtkhtml
995 itself. This caused a crash when g_atexit() was implemented as a
996 function in the GLib DLL. The gtkhtml DLL was already unloaded by
997 the time the GLib DLL got unloaded.
999 * glib/gutils.c: #undef the #define mentioned above, to also get a
1000 real g_atexit() into the DLL for backward compatibility. Document
1001 the Windows behaviour of g_atexit(), and document the varying ways
1002 atexit() can behave in the context of dynamically loaded modules
1005 2005-08-31 Matthias Clasen <mclasen@redhat.com>
1007 * glib/glib.symbols:
1009 * glib/gdataset.c: Add string interning functions.
1011 2005-08-28 Matthias Clasen <mclasen@redhat.com>
1013 * glib/giochannel.c: Unify some near-duplicate strings. (#314654,
1016 2005-08-26 Matthias Clasen <mclasen@redhat.com>
1018 * configure.in: Bump version to 2.9.0
1020 2005-08-25 Tor Lillqvist <tml@novell.com>
1022 Make also the g_spawn*() functions take parameters in the GLib
1023 file name encoding, i.e. UTF-8, on Windows. Has no impact on Unix
1024 API or ABI. Like the other GLib API that was earlier changed to
1025 use UTF-8 on Windows, the names of the functions that take UTF-8
1026 have _utf8 suffixes added by using preprocessor macros in the
1027 header file. The old names are kept for functions with the old
1028 behaviour, taking parameters in the system codepage, for DLL ABI
1031 * glib/gspawn.h: On Win32 add the suffix _utf8 to the names of the
1032 g_spawn*() functions.
1034 * glib/gspawn-win32.c: Use wide-char API on NT-based
1035 Windows. Convert parameters from UTF-8 to wide chars (NT) or
1036 system codepage (Win9x) and call the C library _wspawn*() or
1037 spawn*() functions respectvely. Add DLL ABI stability versions
1038 that take parameters in the system codepage.
1040 * glib/gspawn-win32-helper.c: On NT-based Windows use the
1041 wide-char versions of argv and envp, and use wide-char API to
1042 change directory and spawn the program to run. Remove the verbose
1043 debugging output, it was too complex to modify for the wide-char
1044 features. (Just add temporary debugging printouts if needed, no
1045 need to have them permanently in the source.)
1047 * glib/gspawn.c: Corresponding documentation updates.
1049 * glib/glib.symbols: Corresponding changes: Mark the ABI stability
1050 symbols as PRIVATE, add the new _utf8-suffixed ones.
1052 2005-08-24 Stepan Kasal <kasal@ucw.cz>
1054 * glib/gtypes.h (G_MININT64): Cast the constant to gint64; it is
1055 guint64 otherwise and that can produce warnings about comparison
1056 between signed and unsigned.
1058 2005-08-23 Matthias Clasen <mclasen@redhat.com>
1060 * glib/gutils.c: Fix the crt_externs.h include.
1062 2005-08-23 Stepan Kasal <kasal@ucw.cz>
1064 * NEWS: Fix spelling of my first name.
1066 2005-08-23 Matthias Clasen <mclasen@redhat.com>
1070 * === Released 2.8.1 ===
1074 2005-08-20 Hans Breuer <hans@breuer.org>
1076 * glib/makefile.msc.in : link with ws2_32.lib
1078 2005-08-18 Tor Lillqvist <tml@novell.com>
1080 * configure.in: Check for <sys/wait.h>
1082 * glib/gbacktrace.c: Include <sys/wait.h> on if HAVE_SYS_WAIT_H.
1084 2005-08-18 Ross Burton <ross@burtonini.com>
1087 Optimise single-character insertions.
1092 * tests/string-test.c:
1093 Add tests for new optimisation, and fix a leak.
1095 2005-08-17 Matthias Clasen <mclasen@redhat.com>
1097 * configure.in: Check for crt_externs.h and _NSGetEnviron.
1099 * glib/gutils.c: On Darwin, include crt-externs.h and
1100 define environ using _NSGetEnviron(). (#313731)
1102 2005-08-16 Stepan Kasal <kasal@ucw.cz>
1104 * glib/gutils.c (g_get_any_init): Move the body of the big if...
1105 (g_get_any_init_do): ... to this new function.
1106 (g_get_any_init): Declare as inline.
1107 (g_get_any_init_locked): New inline function, does the locking.
1108 Make use of these two throughout the code.
1110 2005-08-15 Matthias Clasen <mclasen@redhat.com>
1112 * glib/gbacktrace.c (g_on_error_stack_trace): Wait for
1113 the child process and then simply return. This makes
1114 The "S" option work as documented in g_on_error_query().
1115 (#313125, Matthew F. Barnes)
1117 * glib/gunicode.h: Update the link to Unicode category
1118 values. (#313369, Behnam Esfahbod)
1120 * glib/gqueue.c (g_queue_find_custom): Clarify docs
1121 a little. (#311727, Tristan van Berkom)
1123 * glib/abicheck.sh, gobject/abicheck.sh: Make the
1124 check work on ia64 too, where some symbols ended up
1125 in yet another section.