1 Fri Dec 2 10:55:07 2005 Tim Janik <timj@imendio.com>
3 * tests/slice-test.c: extended to perform the benchmarking on the old
4 memchunk code if 'O' is selected.
6 * tests/memchunks.c: new file which contains the old GLib mem chunks
7 implementation with prefix old_mem_chunk_.
9 * tests/Makefile.am: added memchunks.c
11 Fri Dec 2 00:16:59 2005 Tim Janik <timj@imendio.com>
13 * glib/gslice.c: improved gettimeofday() timer resolution by maintaining
14 acceptable load distribution of the syscall.
15 (allocator_get_magazine_threshold): reduce minimum/single-thread magazine
16 sizes to approximately page_size/2.5 to avoid excessive startup allocations.
18 2005-12-01 Matthias Clasen <mclasen@redhat.com>
20 * tests/gobject/ifacecheck.c (test_iface_base_init): Don't
21 double-free base interface members.
23 * tests/Makefile.am: Remove duplicate variable.
25 Thu Dec 1 17:32:46 2005 Tim Janik <timj@imendio.com>
27 * glib/gslice.[hc]: new slice allocator implementation.
29 * tests/slice-test.c: added random slice allocation test.
31 * glib/gthread.[hc]: removed newly added private thread mem API.
36 * glib/gmem.c: divided glib threading initialisation into three phases,
37 initialisation where private keys and messaging are not available (only
38 needed by gmem.c), initialisation without messaging but private keys
39 available (gslice.c, gmessage.c), and full fledged initialisers that
40 server the rest of glib. initialisation functions got renamed to reflect
41 the limitations of their corresponding phases.
43 * glib/gmem.c: removed memchunk code, defer allocations to
46 * glib/gmem.[hc]: removed g_slice_* skeletons.
48 * glib/glib.symbols: added g_slice_* symbols.
50 * configure.in: check for availability of posix_memalign(3), memalign(3)
53 * glib/Makefile.am: added gslice.[hc].
55 2005-12-01 Tor Lillqvist <tml@novell.com>
57 * glib/gstdio.c (g_stat): In the Win32 implementation, strip
58 trailing slash(es) for non-root folders. stat() fails if non-root
59 folders are specified with trailing slashes. It's too much hassle
60 to demand that callers strip such slashes themselves, especially
61 as it is easy to get it wrong and strip the slash of a root
63 (g_rename): On NT-based Windows, use MoveFileEx() with
64 MOVEFILE_REPLACE_EXISTING to better match Unix behaviour.
66 2005-11-28 Matthias Clasen <mclasen@redhat.com>
68 Fix G_STMT_START / G_STMT_END on Solaris. (#321972,
71 * configure.in: Check whether do { } while (0) works.
73 * glib/gmacros.h: Use do { } while (0) for G_STMT_START /
74 G_STMT_END if it works.
76 2005-11-28 Matthias Clasen <mclasen@redhat.com>
78 * glib/gthread.c (g_static_rw_lock_wait, g_static_rw_lock_signal):
79 * glib/gnode.c (g_node_depth_traverse_level):
80 * glib/gmem.c (g_allocator_new):
81 * glib/ghash.c (g_hash_table_unref, g_hash_table_destroy)
82 (g_hash_table_foreach_remove_or_steal): Silence compiler
85 2005-11-27 Matthias Clasen <mclasen@redhat.com>
87 * glib/gunicollate.c (g_utf8_collate_key_for_filename):
88 Don't read beyond len. (#322520, Christian Persch)
90 Wed Nov 23 17:34:01 2005 Tim Janik <timj@imendio.com>
92 * glib/gdataset.c: access datalist flags via atomic pointer access
93 functions, instead of acquiring the dataset lock. this is faster and
94 also matches the atomic pointer readouts in gdatalistprivate.h and
95 direct pointer modifications required by gobject.c.
97 Wed Nov 23 13:35:31 2005 Tim Janik <timj@gtk.org>
99 * glib/gdataset.c: streamlined initialization code somewhat,
100 removed GData node cache.
102 2005-11-22 Matthias Clasen <mclasen@redhat.com>
104 * glib/ghash.c (g_hash_table_ref):
105 (g_hash_table_unref): Mark these as new API.
107 Avoid double locking in g_intern_string (#322133,
110 * glib/gdataset.c (g_quark_from_string_internal): New
111 internal function which factors out common parts of
112 g_quark_from[_static]_string.
113 (g_quark_from_string, g_quark_from_static_string):
114 Use g_quark_from_string_internal.
115 (g_intern_string, g_intern_static_string): Use
116 g_quark_from_string_internal, and only take the
118 (g_quark_new): Don't store the strings shifted by -1
119 in the g_quarks array.
120 (g_quark_to_string): Adapt to the previous change.
122 Tue Nov 22 14:04:26 2005 Tim Janik <timj@imendio.com>
126 g_hash_table_new_full(): create hash tables with a ref count of 1.
127 g_hash_table_ref(): atomically ref_count+=1
128 g_hash_table_unref(): atomically ref_count-=1, destroys hash table
129 when refcount reaches 0.
130 g_hash_table_destroy(): just destroy keys and values, unref by 1.
131 g_hash_table_insert():
132 g_hash_table_replace(): assert ref_count>0.
135 * glib/gatomic.c: added 'volatile' qualifier to all atomic pointer and
138 2005-11-20 Behdad Esfahbod <behdad@gnome.org>
140 * glib/guniprop.c (g_unichar_get_mirror_char): Remove unused
141 variables. (#321984, Andrew Paprocki)
143 2005-11-18 Matthias Clasen <mclasen@redhat.com>
145 * configure.in: Bump version
147 2005-11-17 Matthias Clasen <mclasen@redhat.com>
149 * === Released 2.9.0 ===
151 * glib/glib.symbols: Updates
153 * glib/gutf8.c (g_utf8_offset_to_pointer): Add some performance
154 hints to the docs. (Paolo Borelli)
158 * glib/gthread.c (g_thread_foreach): Mark as new api.
160 * README.in: Updates.
164 * glib/gdate.c (g_date_set_time_t):
165 (g_date_set_time_val): Add functions to set a date from
166 a time_t and from a GTimeVal.
167 (g_date_set_time): Deprecate. (#314067, Roger Leigh)
170 * tests/utf8-pointer.c: Unit tests for g_utf8_pointer_to_offset
171 and g_utf8_offset_to_pointer.
173 * glib/gutf8.c (g_utf8_pointer_to_offset)
174 (g_utf8_offset_to_pointer): Handle negative offsets, and use
175 "stutter stepping" for going backwards. (#320638, Larry
180 * glib/gthread.c: const correctness fixes, found
181 by Arjan van de Ven and gcc.
183 2005-11-16 Behdad Esfahbod <behdad@gnome.org>
185 * docs/reference/glib/tmpl/unicode.sgml: Correct typo on mentioning
186 @G_UNICODE_BREAK_UNKNOWN that should be #G_UNICODE_BREAK_UNKNOWN.
188 2005-11-09 Behdad Esfahbod <behdad@gnome.org>
190 * glib/guniprop.c: Use bit hacks instead when checking a general
191 category value against multiple values.
193 * glib/gutf8.c: Change ISO10646 to Unicode in docs.
195 2005-11-10 Simos Xenitellis <simos@gnome.org>
197 * configure.in: Added tt (Tatar) to ALL_LINGUAS.
199 2005-11-05 Tor Lillqvist <tml@novell.com>
201 * glib/giowin32.c (g_io_win32_set_flags): Don't call g_warning().
202 (#320688) Instead set error and return failure.
204 2005-11-07 Matthias Clasen <mclasen@redhat.com>
206 * glib/gmain.c (g_main_dispatch): Don't call cb_funcs->unref
207 while holding the context lock. (#320886, Andy Wingo)
209 2005-11-07 Billy Biggs <vektor@dumbterm.net>
211 * tests/refcount/Makefile.am: Add a missing reference to libglib
212 in the LDADD for this test.
214 2005-11-07 Matthias Clasen <mclasen@redhat.com>
216 * glib/gmem.c (g_allocator_new): We need to set n_preallocs to a
217 nonzero value, otherwise GTK+ 2.8 breaks when compiled against
218 GLib 2.9. (#320755, Luca Ferretti)
220 2005-11-04 Matthias Clasen <mclasen@redhat.com>
222 * glib/Makefile.am: Apply a patch from Behdad Esfahbod to
223 use a faster lookup table for g_unichar_get_mirror_char().
224 * glib/gmirroringtable.h: The new table.
225 * glib/gunichartables.h: Remove bidi_mirroring_table.
226 * glib/gen-unicode-tables.pl: Don't generate the mirroring
228 * glib/glib-mirroring-tab/*: Sources for the program
229 which generated gmirroringtable.h.
231 * glib/glist.c: Avoid some code duplication.
233 * glib/gscanner.h: Include gdataset.h. (#320322)
235 * glib/gdebug.h: add new GDebugFlag for fatal_criticals
236 * glib/gmessages.c: (_g_debug_init): handle G_DEBUG=fatal_criticals,
237 to help find critical warnings in applications. (#320017,
240 2005-11-02 Tor Lillqvist <tml@novell.com>
242 * glib/glib.symbols: Remove large amount of trailing whitespace
243 from one line. Remove #ifdef G_UNDEFINED from around
244 g_hash_table_new and g_hash_table_new_full symbols, what was the
245 purpose of that? Add the g_slice functions.
247 Tue Nov 1 16:24:20 2005 Tim Janik <timj@imendio.com>
249 * glib/gmem.[hc]: prepared deprecation of GMemChunk and GAllocator.
250 added g_slice_*() API to allocate and cache small bits of memory.
251 an actuall allocator implementation for g_slice_*() is still pending.
253 * glib/gthread.[hc]: changes from a patch by Matthias Clasen.
254 changed GRealThread list to use in-structure *next; fields instead
255 of GSList, in order for thread iteration to not depenend on g_slice_*()
257 _g_thread_mem_private_get():
258 _g_thread_mem_private_set(): added accessors for private memory,
259 needed because the ordinary GPrivate implementation relies on GArray
260 and GSList and therefore indirectly on working g_slice_*() allocations.
263 g_thread_foreach(): new public API function to loop over all existing threads.
277 removed GAllocator and free list usages and accompanying locks.
278 use g_slice_*() API to allocate and cache small bits of memory.
280 * glib/ghook.h: removed GMemChunk field from public API.
283 * glib/glist.h: deprecate allocator API, provide _free1() for consistency.
285 * glib/gnode.h: deprecate allocator API.
287 * glib/gmain.c: reordered GPollRec fields so g_slice_free_chain() can
288 be used for poll rec lists.
290 * glib/grel.c: removed mem chunk usage, and allocated tuples via g_slice_*().
291 g_relation_destroy(): free all tuples from the all_tuples hash table,
292 this effectively maintains the life time track keeping of tuples.
293 g_relation_delete_tuple(): free tuples which are removed from the
294 all_tuples hash table. this fixes a temporary leak that was present
295 in the memchunk code until the destruction of the relation.
297 2005-10-29 Matthias Clasen <mclasen@redhat.com>
299 * tests/convert-test.c: Add some tests for conversions between
300 UTF-8, UCS-4 and UTF-16.
302 * glib/gutf8.c (g_utf8_to_ucs4, g_utf8_to_utf16): Fix handling
303 of len == -1, noticed by Morten Welinder.
305 2005-10-27 Erdal Ronahi <erdal.ronahi@gmail.com>
307 * configure.in: Added ku (Kurdish) to ALL_LINGUAS
309 2005-10-26 Matthias Clasen <mclasen@redhat.com>
311 * glib/gutf8.c (g_ucs4_to_utf8): Don't set items_read twice
312 in the error case, and add some documentation. (#319806, Morten
315 2005-10-19 Manish Singh <yosh@gimp.org>
317 * configure.in: Use AC_CHECK_FUNCS for _NSGetEnviron, to get the
318 config.h symbol automatically. Fixes bug #313731.
320 2005-10-19 Tor Lillqvist <tml@novell.com>
322 * glib/gwin32.c (g_win32_get_package_installation_directory):
323 Return a g_strdup()ed copy of the value stored in the hash table,
324 so that it can be g_free()d without leaving a dangling pointer in
325 the hash table. (#319232)
327 2005-10-06 Matthias Clasen <mclasen@redhat.com>
329 * glib/gunicollate.c (g_utf8_collate_key_for_filename): Handle
330 all-zero sequences correctly. (#317930, Sebastien Bacher)
332 2005-10-05 Matthias Clasen <mclasen@redhat.com>
334 * glib/gmarkup.c (g_markup_escape_text): Doc addition
336 * Makefile.am: Add Collation tests.
338 * tests/collate/*: Inputs and expected outputs for collation tests.
340 * tests/run-collate-tests.sh: Script to run collation tests.
342 * tests/unicode-collate.c (main): Rework slightly to make
343 it usable in unit tests. Also test g_utf8_collate_key_for_filename().
345 2005-10-01 Behdad Esfahbod <behdad@gnome.org>
347 * docs/reference/glib/tmpl/unicode.sgml:
348 * glib/gen-unicode-tables.pl:
350 * glib/gunichartables.h:
352 * tests/casefold.txt:
353 * tests/casemap.txt: Updated to Unicode 4.1. There are five new
354 GUnicodeBreakType types. That may break some applications, like
357 2005-09-26 Matthias Clasen <mclasen@redhat.com>
359 * glib/gstrfuncs.c (g_ascii_strtoull): Add details to the
360 docs. (#314393, Matthew F. Barnes)
363 * glib/gprintf.h: Remove g_snprintf() and g_vsnprintf(), since
364 they are already declared in glib.h. This doesn't break documented
365 use of gprintf.h, but should probably be pointed out in the
366 release notes for 2.10. (#314232, Behdad Esfahbod)
368 Tue Sep 20 13:16:04 2005 Tim Janik <timj@imendio.com>
370 * glib/gpattern.c (g_pattern_ph_match): applied significant recursion
371 complexity optimization, based on a patch from Matthias Clasen.
373 * tests/patterntest.c: more tests, mostly from matthias.
375 2005-09-20 Matthias Clasen <mclasen@redhat.com>
377 * glib/gqueue.c (g_queue_insert_sorted): Correct the docs.
378 (#316703, Mark Drago)
380 Mon Sep 19 17:23:23 2005 Tim Janik <timj@imendio.com>
382 * glib/gpattern.c: applied a patch from matthias which checks on the
383 upper bound of GPatternSpec length to optimize matches.
386 * tests/patterntest.c: added more match cases.
388 2005-09-16 Tor Lillqvist <tml@novell.com>
390 * glib/gstrfuncs.c (g_ascii_strcasecmp, g_ascii_strncasecmp): Add
391 warning to doc comment that these functions should not be used on
392 encodings like CP932.
394 2005-09-14 Matthias Clasen <mclasen@redhat.com>
396 * tests/keyfile-test.c: Add a test for grup names of length 1.
398 * glib/gkeyfile.c (g_key_file_line_is_group): Accept group names
399 of length 1. (#316309)
401 2005-09-12 Matthias Clasen <mclasen@redhat.com>
403 * glib/gmarkup.c (g_markup_escape_text): Clarify docs.
405 2005-09-11 Sebastian Wilhelmi <seppi@seppi.de>
407 * tests/refcount/Makefile.am (INCLUDES): Link the the refcount
408 tests to the system thread library $(G_THREAD_LIBS). Fixes #313744
411 2005-09-11 Kjartan Maraas <kmaraas@gnome.org>
413 * glib/gmain.c: (g_child_watch_prepare), (g_child_watch_check),
414 (child_watch_helper_thread): Remove some dead code. Closes
417 2005-09-07 Tor Lillqvist <tml@novell.com>
419 * glib/Makefile.am: Create also a console version of the
420 gspawn-win32-helper program, gspawn-win32-helper-console.exe.
421 It's otherwise identical to gspawn-win32-helper.exe, except marked
422 as a console application (linked without the -mwindows option).
424 * glib/gspawn-win32.c (do_spawn_directly, do_spawn_with_pipes):
425 Drop the dont_wait parameter. Its truth value correlated 100% with
426 the NULLness of the exit_status parameter anyway, so it's enough
427 to check whether exit_status is NULL. Invert the sense of the
428 dont_return_handle parameter and rename it to do_return_handle, to
429 make the code easier to read by avoiding double negations.
431 (g_spawn_sync_utf8, g_spawn_async_with_pipes_utf8): Modify calls
432 to do_spawn_with_pipes() accordingly.
434 (do_spawn_with_pipes): If we have a console, use the console
435 version of the helper program, otherwise use the GUI one. This
436 avoids extra console windows opening up in some situations. (In
437 case a console application uses the GUI gspawn-win32-helper.exe to
438 spawn another console application we would get a separate console
439 for the spawned console application).
441 * glib-zip.in: Distribute also gspawn-win32-helper-console.exe.
443 2005-09-05 Matthias Clasen <mclasen@redhat.com>
445 * glib/gmappedfile.c (g_mapped_file_new): Report an error
446 if the file is too large. (#315275, Kjartan Maraas)
448 * glib/gkeyfile.c (g_key_file_load_from_fd): The return value
449 of read() is signed. (#315273, Kjartan Maraas)
451 2005-08-31 Tor Lillqvist <tml@novell.com>
453 * glib/gutils.h: Wrapping atexit() is a bad idea on Windows, where
454 the EXE and each DLL have their own atexit function chains.
456 #define g_atexit as atexit instead. This means it has a
457 better chance of doing what the caller wants. For instance,
458 gtkhtml calls g_atexit() registering a function in gtkhtml
459 itself. This caused a crash when g_atexit() was implemented as a
460 function in the GLib DLL. The gtkhtml DLL was already unloaded by
461 the time the GLib DLL got unloaded.
463 * glib/gutils.c: #undef the #define mentioned above, to also get a
464 real g_atexit() into the DLL for backward compatibility. Document
465 the Windows behaviour of g_atexit(), and document the varying ways
466 atexit() can behave in the context of dynamically loaded modules
469 2005-08-31 Matthias Clasen <mclasen@redhat.com>
473 * glib/gdataset.c: Add string interning functions.
475 2005-08-28 Matthias Clasen <mclasen@redhat.com>
477 * glib/giochannel.c: Unify some near-duplicate strings. (#314654,
480 2005-08-26 Matthias Clasen <mclasen@redhat.com>
482 * configure.in: Bump version to 2.9.0
484 2005-08-25 Tor Lillqvist <tml@novell.com>
486 Make also the g_spawn*() functions take parameters in the GLib
487 file name encoding, i.e. UTF-8, on Windows. Has no impact on Unix
488 API or ABI. Like the other GLib API that was earlier changed to
489 use UTF-8 on Windows, the names of the functions that take UTF-8
490 have _utf8 suffixes added by using preprocessor macros in the
491 header file. The old names are kept for functions with the old
492 behaviour, taking parameters in the system codepage, for DLL ABI
495 * glib/gspawn.h: On Win32 add the suffix _utf8 to the names of the
496 g_spawn*() functions.
498 * glib/gspawn-win32.c: Use wide-char API on NT-based
499 Windows. Convert parameters from UTF-8 to wide chars (NT) or
500 system codepage (Win9x) and call the C library _wspawn*() or
501 spawn*() functions respectvely. Add DLL ABI stability versions
502 that take parameters in the system codepage.
504 * glib/gspawn-win32-helper.c: On NT-based Windows use the
505 wide-char versions of argv and envp, and use wide-char API to
506 change directory and spawn the program to run. Remove the verbose
507 debugging output, it was too complex to modify for the wide-char
508 features. (Just add temporary debugging printouts if needed, no
509 need to have them permanently in the source.)
511 * glib/gspawn.c: Corresponding documentation updates.
513 * glib/glib.symbols: Corresponding changes: Mark the ABI stability
514 symbols as PRIVATE, add the new _utf8-suffixed ones.
516 2005-08-24 Stepan Kasal <kasal@ucw.cz>
518 * glib/gtypes.h (G_MININT64): Cast the constant to gint64; it is
519 guint64 otherwise and that can produce warnings about comparison
520 between signed and unsigned.
522 2005-08-23 Matthias Clasen <mclasen@redhat.com>
524 * glib/gutils.c: Fix the crt_externs.h include.
526 2005-08-23 Stepan Kasal <kasal@ucw.cz>
528 * NEWS: Fix spelling of my first name.
530 2005-08-23 Matthias Clasen <mclasen@redhat.com>
534 * === Released 2.8.1 ===
538 2005-08-20 Hans Breuer <hans@breuer.org>
540 * glib/makefile.msc.in : link with ws2_32.lib
542 2005-08-18 Tor Lillqvist <tml@novell.com>
544 * configure.in: Check for <sys/wait.h>
546 * glib/gbacktrace.c: Include <sys/wait.h> on if HAVE_SYS_WAIT_H.
548 2005-08-18 Ross Burton <ross@burtonini.com>
551 Optimise single-character insertions.
556 * tests/string-test.c:
557 Add tests for new optimisation, and fix a leak.
559 2005-08-17 Matthias Clasen <mclasen@redhat.com>
561 * configure.in: Check for crt_externs.h and _NSGetEnviron.
563 * glib/gutils.c: On Darwin, include crt-externs.h and
564 define environ using _NSGetEnviron(). (#313731)
566 2005-08-16 Stepan Kasal <kasal@ucw.cz>
568 * glib/gutils.c (g_get_any_init): Move the body of the big if...
569 (g_get_any_init_do): ... to this new function.
570 (g_get_any_init): Declare as inline.
571 (g_get_any_init_locked): New inline function, does the locking.
572 Make use of these two throughout the code.
574 2005-08-15 Matthias Clasen <mclasen@redhat.com>
576 * glib/gbacktrace.c (g_on_error_stack_trace): Wait for
577 the child process and then simply return. This makes
578 The "S" option work as documented in g_on_error_query().
579 (#313125, Matthew F. Barnes)
581 * glib/gunicode.h: Update the link to Unicode category
582 values. (#313369, Behnam Esfahbod)
584 * glib/gqueue.c (g_queue_find_custom): Clarify docs
585 a little. (#311727, Tristan van Berkom)
587 * glib/abicheck.sh, gobject/abicheck.sh: Make the
588 check work on ia64 too, where some symbols ended up
589 in yet another section.