1 Fri Dec 2 13:08:58 2005 Tim Janik <timj@gtk.org>
3 * glib/gslice.h (g_slice_free): reworked GCC-specific type-safe macro
4 variant into something less verbose, but digestible for gcc-3.4.
6 Fri Dec 2 10:55:07 2005 Tim Janik <timj@imendio.com>
8 * tests/slice-test.c: extended to perform the benchmarking on the old
9 memchunk code if 'O' is selected.
11 * tests/memchunks.c: new file which contains the old GLib mem chunks
12 implementation with prefix old_mem_chunk_.
14 * tests/Makefile.am: added memchunks.c
16 Fri Dec 2 00:16:59 2005 Tim Janik <timj@imendio.com>
18 * glib/gslice.c: improved gettimeofday() timer resolution by maintaining
19 acceptable load distribution of the syscall.
20 (allocator_get_magazine_threshold): reduce minimum/single-thread magazine
21 sizes to approximately page_size/2.5 to avoid excessive startup allocations.
23 2005-12-01 Matthias Clasen <mclasen@redhat.com>
25 * tests/gobject/ifacecheck.c (test_iface_base_init): Don't
26 double-free base interface members.
28 * tests/Makefile.am: Remove duplicate variable.
30 Thu Dec 1 17:32:46 2005 Tim Janik <timj@imendio.com>
32 * glib/gslice.[hc]: new slice allocator implementation.
34 * tests/slice-test.c: added random slice allocation test.
36 * glib/gthread.[hc]: removed newly added private thread mem API.
41 * glib/gmem.c: divided glib threading initialisation into three phases,
42 initialisation where private keys and messaging are not available (only
43 needed by gmem.c), initialisation without messaging but private keys
44 available (gslice.c, gmessage.c), and full fledged initialisers that
45 server the rest of glib. initialisation functions got renamed to reflect
46 the limitations of their corresponding phases.
48 * glib/gmem.c: removed memchunk code, defer allocations to
51 * glib/gmem.[hc]: removed g_slice_* skeletons.
53 * glib/glib.symbols: added g_slice_* symbols.
55 * configure.in: check for availability of posix_memalign(3), memalign(3)
58 * glib/Makefile.am: added gslice.[hc].
60 2005-12-01 Tor Lillqvist <tml@novell.com>
62 * glib/gstdio.c (g_stat): In the Win32 implementation, strip
63 trailing slash(es) for non-root folders. stat() fails if non-root
64 folders are specified with trailing slashes. It's too much hassle
65 to demand that callers strip such slashes themselves, especially
66 as it is easy to get it wrong and strip the slash of a root
68 (g_rename): On NT-based Windows, use MoveFileEx() with
69 MOVEFILE_REPLACE_EXISTING to better match Unix behaviour.
71 2005-11-28 Matthias Clasen <mclasen@redhat.com>
73 Fix G_STMT_START / G_STMT_END on Solaris. (#321972,
76 * configure.in: Check whether do { } while (0) works.
78 * glib/gmacros.h: Use do { } while (0) for G_STMT_START /
79 G_STMT_END if it works.
81 2005-11-28 Matthias Clasen <mclasen@redhat.com>
83 * glib/gthread.c (g_static_rw_lock_wait, g_static_rw_lock_signal):
84 * glib/gnode.c (g_node_depth_traverse_level):
85 * glib/gmem.c (g_allocator_new):
86 * glib/ghash.c (g_hash_table_unref, g_hash_table_destroy)
87 (g_hash_table_foreach_remove_or_steal): Silence compiler
90 2005-11-27 Matthias Clasen <mclasen@redhat.com>
92 * glib/gunicollate.c (g_utf8_collate_key_for_filename):
93 Don't read beyond len. (#322520, Christian Persch)
95 Wed Nov 23 17:34:01 2005 Tim Janik <timj@imendio.com>
97 * glib/gdataset.c: access datalist flags via atomic pointer access
98 functions, instead of acquiring the dataset lock. this is faster and
99 also matches the atomic pointer readouts in gdatalistprivate.h and
100 direct pointer modifications required by gobject.c.
102 Wed Nov 23 13:35:31 2005 Tim Janik <timj@gtk.org>
104 * glib/gdataset.c: streamlined initialization code somewhat,
105 removed GData node cache.
107 2005-11-22 Matthias Clasen <mclasen@redhat.com>
109 * glib/ghash.c (g_hash_table_ref):
110 (g_hash_table_unref): Mark these as new API.
112 Avoid double locking in g_intern_string (#322133,
115 * glib/gdataset.c (g_quark_from_string_internal): New
116 internal function which factors out common parts of
117 g_quark_from[_static]_string.
118 (g_quark_from_string, g_quark_from_static_string):
119 Use g_quark_from_string_internal.
120 (g_intern_string, g_intern_static_string): Use
121 g_quark_from_string_internal, and only take the
123 (g_quark_new): Don't store the strings shifted by -1
124 in the g_quarks array.
125 (g_quark_to_string): Adapt to the previous change.
127 Tue Nov 22 14:04:26 2005 Tim Janik <timj@imendio.com>
131 g_hash_table_new_full(): create hash tables with a ref count of 1.
132 g_hash_table_ref(): atomically ref_count+=1
133 g_hash_table_unref(): atomically ref_count-=1, destroys hash table
134 when refcount reaches 0.
135 g_hash_table_destroy(): just destroy keys and values, unref by 1.
136 g_hash_table_insert():
137 g_hash_table_replace(): assert ref_count>0.
140 * glib/gatomic.c: added 'volatile' qualifier to all atomic pointer and
143 2005-11-20 Behdad Esfahbod <behdad@gnome.org>
145 * glib/guniprop.c (g_unichar_get_mirror_char): Remove unused
146 variables. (#321984, Andrew Paprocki)
148 2005-11-18 Matthias Clasen <mclasen@redhat.com>
150 * configure.in: Bump version
152 2005-11-17 Matthias Clasen <mclasen@redhat.com>
154 * === Released 2.9.0 ===
156 * glib/glib.symbols: Updates
158 * glib/gutf8.c (g_utf8_offset_to_pointer): Add some performance
159 hints to the docs. (Paolo Borelli)
163 * glib/gthread.c (g_thread_foreach): Mark as new api.
165 * README.in: Updates.
169 * glib/gdate.c (g_date_set_time_t):
170 (g_date_set_time_val): Add functions to set a date from
171 a time_t and from a GTimeVal.
172 (g_date_set_time): Deprecate. (#314067, Roger Leigh)
175 * tests/utf8-pointer.c: Unit tests for g_utf8_pointer_to_offset
176 and g_utf8_offset_to_pointer.
178 * glib/gutf8.c (g_utf8_pointer_to_offset)
179 (g_utf8_offset_to_pointer): Handle negative offsets, and use
180 "stutter stepping" for going backwards. (#320638, Larry
185 * glib/gthread.c: const correctness fixes, found
186 by Arjan van de Ven and gcc.
188 2005-11-16 Behdad Esfahbod <behdad@gnome.org>
190 * docs/reference/glib/tmpl/unicode.sgml: Correct typo on mentioning
191 @G_UNICODE_BREAK_UNKNOWN that should be #G_UNICODE_BREAK_UNKNOWN.
193 2005-11-09 Behdad Esfahbod <behdad@gnome.org>
195 * glib/guniprop.c: Use bit hacks instead when checking a general
196 category value against multiple values.
198 * glib/gutf8.c: Change ISO10646 to Unicode in docs.
200 2005-11-10 Simos Xenitellis <simos@gnome.org>
202 * configure.in: Added tt (Tatar) to ALL_LINGUAS.
204 2005-11-05 Tor Lillqvist <tml@novell.com>
206 * glib/giowin32.c (g_io_win32_set_flags): Don't call g_warning().
207 (#320688) Instead set error and return failure.
209 2005-11-07 Matthias Clasen <mclasen@redhat.com>
211 * glib/gmain.c (g_main_dispatch): Don't call cb_funcs->unref
212 while holding the context lock. (#320886, Andy Wingo)
214 2005-11-07 Billy Biggs <vektor@dumbterm.net>
216 * tests/refcount/Makefile.am: Add a missing reference to libglib
217 in the LDADD for this test.
219 2005-11-07 Matthias Clasen <mclasen@redhat.com>
221 * glib/gmem.c (g_allocator_new): We need to set n_preallocs to a
222 nonzero value, otherwise GTK+ 2.8 breaks when compiled against
223 GLib 2.9. (#320755, Luca Ferretti)
225 2005-11-04 Matthias Clasen <mclasen@redhat.com>
227 * glib/Makefile.am: Apply a patch from Behdad Esfahbod to
228 use a faster lookup table for g_unichar_get_mirror_char().
229 * glib/gmirroringtable.h: The new table.
230 * glib/gunichartables.h: Remove bidi_mirroring_table.
231 * glib/gen-unicode-tables.pl: Don't generate the mirroring
233 * glib/glib-mirroring-tab/*: Sources for the program
234 which generated gmirroringtable.h.
236 * glib/glist.c: Avoid some code duplication.
238 * glib/gscanner.h: Include gdataset.h. (#320322)
240 * glib/gdebug.h: add new GDebugFlag for fatal_criticals
241 * glib/gmessages.c: (_g_debug_init): handle G_DEBUG=fatal_criticals,
242 to help find critical warnings in applications. (#320017,
245 2005-11-02 Tor Lillqvist <tml@novell.com>
247 * glib/glib.symbols: Remove large amount of trailing whitespace
248 from one line. Remove #ifdef G_UNDEFINED from around
249 g_hash_table_new and g_hash_table_new_full symbols, what was the
250 purpose of that? Add the g_slice functions.
252 Tue Nov 1 16:24:20 2005 Tim Janik <timj@imendio.com>
254 * glib/gmem.[hc]: prepared deprecation of GMemChunk and GAllocator.
255 added g_slice_*() API to allocate and cache small bits of memory.
256 an actuall allocator implementation for g_slice_*() is still pending.
258 * glib/gthread.[hc]: changes from a patch by Matthias Clasen.
259 changed GRealThread list to use in-structure *next; fields instead
260 of GSList, in order for thread iteration to not depenend on g_slice_*()
262 _g_thread_mem_private_get():
263 _g_thread_mem_private_set(): added accessors for private memory,
264 needed because the ordinary GPrivate implementation relies on GArray
265 and GSList and therefore indirectly on working g_slice_*() allocations.
268 g_thread_foreach(): new public API function to loop over all existing threads.
282 removed GAllocator and free list usages and accompanying locks.
283 use g_slice_*() API to allocate and cache small bits of memory.
285 * glib/ghook.h: removed GMemChunk field from public API.
288 * glib/glist.h: deprecate allocator API, provide _free1() for consistency.
290 * glib/gnode.h: deprecate allocator API.
292 * glib/gmain.c: reordered GPollRec fields so g_slice_free_chain() can
293 be used for poll rec lists.
295 * glib/grel.c: removed mem chunk usage, and allocated tuples via g_slice_*().
296 g_relation_destroy(): free all tuples from the all_tuples hash table,
297 this effectively maintains the life time track keeping of tuples.
298 g_relation_delete_tuple(): free tuples which are removed from the
299 all_tuples hash table. this fixes a temporary leak that was present
300 in the memchunk code until the destruction of the relation.
302 2005-10-29 Matthias Clasen <mclasen@redhat.com>
304 * tests/convert-test.c: Add some tests for conversions between
305 UTF-8, UCS-4 and UTF-16.
307 * glib/gutf8.c (g_utf8_to_ucs4, g_utf8_to_utf16): Fix handling
308 of len == -1, noticed by Morten Welinder.
310 2005-10-27 Erdal Ronahi <erdal.ronahi@gmail.com>
312 * configure.in: Added ku (Kurdish) to ALL_LINGUAS
314 2005-10-26 Matthias Clasen <mclasen@redhat.com>
316 * glib/gutf8.c (g_ucs4_to_utf8): Don't set items_read twice
317 in the error case, and add some documentation. (#319806, Morten
320 2005-10-19 Manish Singh <yosh@gimp.org>
322 * configure.in: Use AC_CHECK_FUNCS for _NSGetEnviron, to get the
323 config.h symbol automatically. Fixes bug #313731.
325 2005-10-19 Tor Lillqvist <tml@novell.com>
327 * glib/gwin32.c (g_win32_get_package_installation_directory):
328 Return a g_strdup()ed copy of the value stored in the hash table,
329 so that it can be g_free()d without leaving a dangling pointer in
330 the hash table. (#319232)
332 2005-10-06 Matthias Clasen <mclasen@redhat.com>
334 * glib/gunicollate.c (g_utf8_collate_key_for_filename): Handle
335 all-zero sequences correctly. (#317930, Sebastien Bacher)
337 2005-10-05 Matthias Clasen <mclasen@redhat.com>
339 * glib/gmarkup.c (g_markup_escape_text): Doc addition
341 * Makefile.am: Add Collation tests.
343 * tests/collate/*: Inputs and expected outputs for collation tests.
345 * tests/run-collate-tests.sh: Script to run collation tests.
347 * tests/unicode-collate.c (main): Rework slightly to make
348 it usable in unit tests. Also test g_utf8_collate_key_for_filename().
350 2005-10-01 Behdad Esfahbod <behdad@gnome.org>
352 * docs/reference/glib/tmpl/unicode.sgml:
353 * glib/gen-unicode-tables.pl:
355 * glib/gunichartables.h:
357 * tests/casefold.txt:
358 * tests/casemap.txt: Updated to Unicode 4.1. There are five new
359 GUnicodeBreakType types. That may break some applications, like
362 2005-09-26 Matthias Clasen <mclasen@redhat.com>
364 * glib/gstrfuncs.c (g_ascii_strtoull): Add details to the
365 docs. (#314393, Matthew F. Barnes)
368 * glib/gprintf.h: Remove g_snprintf() and g_vsnprintf(), since
369 they are already declared in glib.h. This doesn't break documented
370 use of gprintf.h, but should probably be pointed out in the
371 release notes for 2.10. (#314232, Behdad Esfahbod)
373 Tue Sep 20 13:16:04 2005 Tim Janik <timj@imendio.com>
375 * glib/gpattern.c (g_pattern_ph_match): applied significant recursion
376 complexity optimization, based on a patch from Matthias Clasen.
378 * tests/patterntest.c: more tests, mostly from matthias.
380 2005-09-20 Matthias Clasen <mclasen@redhat.com>
382 * glib/gqueue.c (g_queue_insert_sorted): Correct the docs.
383 (#316703, Mark Drago)
385 Mon Sep 19 17:23:23 2005 Tim Janik <timj@imendio.com>
387 * glib/gpattern.c: applied a patch from matthias which checks on the
388 upper bound of GPatternSpec length to optimize matches.
391 * tests/patterntest.c: added more match cases.
393 2005-09-16 Tor Lillqvist <tml@novell.com>
395 * glib/gstrfuncs.c (g_ascii_strcasecmp, g_ascii_strncasecmp): Add
396 warning to doc comment that these functions should not be used on
397 encodings like CP932.
399 2005-09-14 Matthias Clasen <mclasen@redhat.com>
401 * tests/keyfile-test.c: Add a test for grup names of length 1.
403 * glib/gkeyfile.c (g_key_file_line_is_group): Accept group names
404 of length 1. (#316309)
406 2005-09-12 Matthias Clasen <mclasen@redhat.com>
408 * glib/gmarkup.c (g_markup_escape_text): Clarify docs.
410 2005-09-11 Sebastian Wilhelmi <seppi@seppi.de>
412 * tests/refcount/Makefile.am (INCLUDES): Link the the refcount
413 tests to the system thread library $(G_THREAD_LIBS). Fixes #313744
416 2005-09-11 Kjartan Maraas <kmaraas@gnome.org>
418 * glib/gmain.c: (g_child_watch_prepare), (g_child_watch_check),
419 (child_watch_helper_thread): Remove some dead code. Closes
422 2005-09-07 Tor Lillqvist <tml@novell.com>
424 * glib/Makefile.am: Create also a console version of the
425 gspawn-win32-helper program, gspawn-win32-helper-console.exe.
426 It's otherwise identical to gspawn-win32-helper.exe, except marked
427 as a console application (linked without the -mwindows option).
429 * glib/gspawn-win32.c (do_spawn_directly, do_spawn_with_pipes):
430 Drop the dont_wait parameter. Its truth value correlated 100% with
431 the NULLness of the exit_status parameter anyway, so it's enough
432 to check whether exit_status is NULL. Invert the sense of the
433 dont_return_handle parameter and rename it to do_return_handle, to
434 make the code easier to read by avoiding double negations.
436 (g_spawn_sync_utf8, g_spawn_async_with_pipes_utf8): Modify calls
437 to do_spawn_with_pipes() accordingly.
439 (do_spawn_with_pipes): If we have a console, use the console
440 version of the helper program, otherwise use the GUI one. This
441 avoids extra console windows opening up in some situations. (In
442 case a console application uses the GUI gspawn-win32-helper.exe to
443 spawn another console application we would get a separate console
444 for the spawned console application).
446 * glib-zip.in: Distribute also gspawn-win32-helper-console.exe.
448 2005-09-05 Matthias Clasen <mclasen@redhat.com>
450 * glib/gmappedfile.c (g_mapped_file_new): Report an error
451 if the file is too large. (#315275, Kjartan Maraas)
453 * glib/gkeyfile.c (g_key_file_load_from_fd): The return value
454 of read() is signed. (#315273, Kjartan Maraas)
456 2005-08-31 Tor Lillqvist <tml@novell.com>
458 * glib/gutils.h: Wrapping atexit() is a bad idea on Windows, where
459 the EXE and each DLL have their own atexit function chains.
461 #define g_atexit as atexit instead. This means it has a
462 better chance of doing what the caller wants. For instance,
463 gtkhtml calls g_atexit() registering a function in gtkhtml
464 itself. This caused a crash when g_atexit() was implemented as a
465 function in the GLib DLL. The gtkhtml DLL was already unloaded by
466 the time the GLib DLL got unloaded.
468 * glib/gutils.c: #undef the #define mentioned above, to also get a
469 real g_atexit() into the DLL for backward compatibility. Document
470 the Windows behaviour of g_atexit(), and document the varying ways
471 atexit() can behave in the context of dynamically loaded modules
474 2005-08-31 Matthias Clasen <mclasen@redhat.com>
478 * glib/gdataset.c: Add string interning functions.
480 2005-08-28 Matthias Clasen <mclasen@redhat.com>
482 * glib/giochannel.c: Unify some near-duplicate strings. (#314654,
485 2005-08-26 Matthias Clasen <mclasen@redhat.com>
487 * configure.in: Bump version to 2.9.0
489 2005-08-25 Tor Lillqvist <tml@novell.com>
491 Make also the g_spawn*() functions take parameters in the GLib
492 file name encoding, i.e. UTF-8, on Windows. Has no impact on Unix
493 API or ABI. Like the other GLib API that was earlier changed to
494 use UTF-8 on Windows, the names of the functions that take UTF-8
495 have _utf8 suffixes added by using preprocessor macros in the
496 header file. The old names are kept for functions with the old
497 behaviour, taking parameters in the system codepage, for DLL ABI
500 * glib/gspawn.h: On Win32 add the suffix _utf8 to the names of the
501 g_spawn*() functions.
503 * glib/gspawn-win32.c: Use wide-char API on NT-based
504 Windows. Convert parameters from UTF-8 to wide chars (NT) or
505 system codepage (Win9x) and call the C library _wspawn*() or
506 spawn*() functions respectvely. Add DLL ABI stability versions
507 that take parameters in the system codepage.
509 * glib/gspawn-win32-helper.c: On NT-based Windows use the
510 wide-char versions of argv and envp, and use wide-char API to
511 change directory and spawn the program to run. Remove the verbose
512 debugging output, it was too complex to modify for the wide-char
513 features. (Just add temporary debugging printouts if needed, no
514 need to have them permanently in the source.)
516 * glib/gspawn.c: Corresponding documentation updates.
518 * glib/glib.symbols: Corresponding changes: Mark the ABI stability
519 symbols as PRIVATE, add the new _utf8-suffixed ones.
521 2005-08-24 Stepan Kasal <kasal@ucw.cz>
523 * glib/gtypes.h (G_MININT64): Cast the constant to gint64; it is
524 guint64 otherwise and that can produce warnings about comparison
525 between signed and unsigned.
527 2005-08-23 Matthias Clasen <mclasen@redhat.com>
529 * glib/gutils.c: Fix the crt_externs.h include.
531 2005-08-23 Stepan Kasal <kasal@ucw.cz>
533 * NEWS: Fix spelling of my first name.
535 2005-08-23 Matthias Clasen <mclasen@redhat.com>
539 * === Released 2.8.1 ===
543 2005-08-20 Hans Breuer <hans@breuer.org>
545 * glib/makefile.msc.in : link with ws2_32.lib
547 2005-08-18 Tor Lillqvist <tml@novell.com>
549 * configure.in: Check for <sys/wait.h>
551 * glib/gbacktrace.c: Include <sys/wait.h> on if HAVE_SYS_WAIT_H.
553 2005-08-18 Ross Burton <ross@burtonini.com>
556 Optimise single-character insertions.
561 * tests/string-test.c:
562 Add tests for new optimisation, and fix a leak.
564 2005-08-17 Matthias Clasen <mclasen@redhat.com>
566 * configure.in: Check for crt_externs.h and _NSGetEnviron.
568 * glib/gutils.c: On Darwin, include crt-externs.h and
569 define environ using _NSGetEnviron(). (#313731)
571 2005-08-16 Stepan Kasal <kasal@ucw.cz>
573 * glib/gutils.c (g_get_any_init): Move the body of the big if...
574 (g_get_any_init_do): ... to this new function.
575 (g_get_any_init): Declare as inline.
576 (g_get_any_init_locked): New inline function, does the locking.
577 Make use of these two throughout the code.
579 2005-08-15 Matthias Clasen <mclasen@redhat.com>
581 * glib/gbacktrace.c (g_on_error_stack_trace): Wait for
582 the child process and then simply return. This makes
583 The "S" option work as documented in g_on_error_query().
584 (#313125, Matthew F. Barnes)
586 * glib/gunicode.h: Update the link to Unicode category
587 values. (#313369, Behnam Esfahbod)
589 * glib/gqueue.c (g_queue_find_custom): Clarify docs
590 a little. (#311727, Tristan van Berkom)
592 * glib/abicheck.sh, gobject/abicheck.sh: Make the
593 check work on ia64 too, where some symbols ended up
594 in yet another section.