1 2000-08-31 Tor Lillqvist <tml@iki.fi>
5 * giowin32.c (g_io_channel_win32_make_pollfd): New function, to
6 make a GPollFD from a GIOChannel. Creates the events and starts
7 the reader thread if necessary.
10 * giowin32.c (g_io_channel_win32_poll): No use for separate
13 * gmain.c (g_get_current_time): (Win32): Simplify, use
14 GetSystemTimeAsFileTime().
16 2000-08-27 Tor Lillqvist <tml@iki.fi>
18 * giowin32.c (g_io_channel_win32_poll): New function, otherwise
19 like g_io_channel_win32_wait_for_condition(), but accept several
21 (g_io_channel_win32_wait_for_condition): Call
22 g_io_channel_win32_poll().
24 * glib.h: Declare g_io_channel_win32_poll().
26 * gwin32.c (g_win32_error_message): Don't believe return value
29 2000-08-25 Elliot Lee <sopwith@redhat.com>
30 * glib.h, gunicode.h, gmodule/gmodule.h:
31 : Mark the following functions G_GNUC_CONST (to allow
32 optimization) because their results are a function of only their parameters:
33 g_int_hash, g_int_equal, g_direct_hash, g_direct_equal, g_quark_to_string,
34 g_date_is_leap_year, g_date_days_in_month, g_date_monday_weeks_in_year,
35 g_date_sunday_weeks_in_year, g_spaced_primes_closest, g_unichar_is*,
36 g_unichar_to*, g_unichar_*digit_value, g_unichar_type
38 2000-08-21 Elliot Lee <sopwith@redhat.com>
39 * gobject/Makefile.am, gobject/gobject-query.c, gobject/gparamspecs.c: Fix inclusion of
42 Mon Aug 21 14:46:23 2000 Owen Taylor <otaylor@redhat.com>
44 * tests/gio-test.c: Fix a couple of trivial bugs that
45 were causing warnings.
47 Mon Aug 21 14:39:36 2000 Owen Taylor <otaylor@redhat.com>
49 * glib.h: Use C99 varargs macros where possible
50 (check __STDC_VERSION__), otherwise, on gcc, use an alternate
51 form of gcc varargs which is more likely
52 to be supported going forward. (Based on some code
55 2000-08-17 Darin Adler <darin@eazel.com>
58 * garray.c: (g_array_free), (g_ptr_array_free),
59 (g_byte_array_free): Return the data left behind.
60 * gstring.c: (g_string_free): Return the data left behind.
62 Changed the free calls that leave data behind so they
63 return a pointer to the left-behind data, NULL if told not
64 to leave anything behind. This makes these calls easier
65 to use correctly, without any incompatible API change for
66 callers that don't know about the return value. Of course,
67 it would be even clearer if the free calls weren't dual-purpose
70 2000-08-12 Tor Lillqvist <tml@iki.fi>
72 * giowin32.c: Some indentation and spacing fixes. Add some more
74 (g_io_win32_add_watch): New function, with common code from
75 g_io_win32_fd_add_watch and g_io_win32_sock_add_watch. Don't start
76 more than one reader thread for a GIOChannel. We should obviously
77 have just one reader thread reading a file descriptor or socket.
79 2000-08-10 Havoc Pennington <hp@redhat.com>
81 * gthread-2.0.pc.in (Cflags): don't duplicate glib Cflags
83 * gmodule-2.0.pc.in (Cflags): don't duplicate glib Cflags
85 * gobject-2.0.pc.in (Cflags): don't duplicate Cflags from glib
88 2000-08-10 Havoc Pennington <hp@redhat.com>
90 * glib-2.0.pc.in (Cflags): Look in glib-2.0/include for
93 2000-08-07 Tor Lillqvist <tml@iki.fi>
95 * tests/gio-test.c (shutdown_source): New function, that calls
96 g_source_remove(). Check return value of g_source_remove(), and
97 decrement running subprocess counter only if g_source_remove()
98 actually did remove the source.
100 (recv_message): Call shutdown_source() on EOF condition. Return
101 FALSE on G_IO_HUP and G_IO_ERR condition. Fix printf format typo.
103 Sun Aug 6 20:06:02 2000 Tim Janik <timj@gtk.org>
105 * gmessages.c (g_log_domain_check_free): keep *last updated while
106 running through the domain list, so we don't screw up the removal,
107 patch provided by Gady Kozma <gadykozma@hotmail.com>.
109 Sun Aug 6 20:03:41 2000 Tim Janik <timj@gtk.org>
111 * gmessages.c (g_log_remove_handler): keep *last updated while running
112 through the handler list, so we don't screw up the removal.
114 Sun Jul 30 16:54:13 2000 Owen Taylor <otaylor@redhat.com>
116 * gunicode.h: Fix stray character
118 * gutf8.c (g_unichar_to_utf8): Allow outbuf to be NULL, in
119 which case we just compute the length.
121 2000-07-31 Havoc Pennington <hp@redhat.com>
123 * Makefile.am (EXTRA_DIST): forgot to add .pc.in to EXTRA_DIST
125 2000-07-31 Havoc Pennington <hp@redhat.com>
127 * glib-2.0.pc.in, gobject-2.0.pc.in, gmodule-2.0.pc.in,
128 gobject-2.0.pc.in: pkg-config data files
130 * Makefile.am: Install/dist the .pc files
132 * configure.in: Output the .pc files
134 2000-07-31 Tor Lillqvist <tml@iki.fi>
136 * giowin32.c (buffer_read): The code didn't compile (must
137 have been sleepy when committing). "return" instead of "break"
139 (g_io_win32_fd_add_watch): Cannot check if the file descriptor is
140 readable by calling ReadFile to read zero bytes. ReadFile blocks
141 on NT even if trying to read nothing at all. So, don't check if
142 file descriptor is readable; assume this function isn't called
145 Sun Jul 30 10:44:16 2000 Tim Janik <timj@gtk.org>
147 * gmain.c (g_get_current_time): fix tor's recent changes which
148 got rid of a required variable in the non-windows path.
150 2000-07-30 Tor Lillqvist <tml@iki.fi>
152 Finally, a new and improved IO Channel and condition watch
153 implementation for Win32. Based on code provided by Craig Setera.
155 When watching file descriptors, for which there is no select()
156 like functionality on Win32 that would work on all Win32 platforms
157 for all types of file descriptors (including anonymous pipes), we
158 start a new thread that blocks while trying to read from the file
159 descriptor. When the read returns, a Win32 Event is signalled that
160 the polling routine eventually notices. Meanwhile, the data being
161 read is stored in a circular buffer, from where the IO channel's
162 read() method picks it up.
164 If the buffer fills up the reading thread has to wait for space
165 becoming available. For this another Win32 Event is used. The IO
166 Channel's read() method signals this when it has read some data
169 The separate reader thread(s), and the circular buffer(s) with
170 associated events mean lots of possibilities for fun parallellism
171 errors. But it seems to work OK, i.e. GIMP runs.
173 * gmain.c: Small changes to the Win32 polling function.
174 (g_main_win32_get_poll_func): New function. Perhaps it would be a
175 good idea to provide this on all platforms.
177 * giowin32.c: The bulk of the new implementation.
178 (g_io_channel_win32_wait_for_condition): New function. To be used
179 where on Unix one does a select() on the channel's fd, like
180 libgimp's gimp_extension_process(). Could be provided on all
183 * glib.h: Update documentation for IO Channels on Win32. Remove
184 the declarations for the as of now obsolete old functions related
185 to IO Channels for pipes with "wakeup" messages.
187 * glib.def: Some new functions.
189 * tests/gio-test.c: New file, to test GIOChannel and main loop.
192 * tests/makefile.mingw.in: Add it.
194 (Later the same night:)
196 * giowin32.c: Compile in the debugging code all the time, but only
197 output debug messages if told so. Add (unadvertised) function to
198 turn on/off debug messages for a channel.
200 (buffer_read): Don't loop. It is expected behaviour to return a
201 short read occasionally, for instance when reading from
202 pipes. It's the calling code that should loop if it *knows* how
203 much the writer has written.
205 * tests/gio-test.c: Correct the program's name in the output.
206 (recv_message): Loop calling g_io_channel_read() (in a new
207 function read_all()) until we have all the bytes we want (that we
208 know the writer has written/will write).
210 Thu Jul 27 05:15:11 2000 Tim Janik <timj@gtk.org>
212 * gstrfuncs.c (g_strlcpy, g_strlcat): completed tor's fix
213 to cover both #ifdef branches.
215 2000-07-26 Tor Lillqvist <tml@iki.fi>
217 * gstrfuncs.c (g_strlcpy, g_strlcat): Return 0 on error, not NULL.
219 * glib.def: Add g_strlcpy, g_strlcat.
221 * glibconfig.h.win32.in: Add gsize and gssize.
223 Wed Jul 26 12:59:31 2000 Tim Janik <timj@gtk.org>
225 * *.[hc]: applied patch from Andreas Persenius <ndap@swipnet.se> that
226 updates the license headers to the GNU Lesser General Public License,
227 as well as updating the copyright year to 2000.
229 Wed Jul 26 05:47:48 2000 Tim Janik <timj@gtk.org>
234 * glib.h: added g_strlcat() and g_strlcpy() wrappers, supplied by
235 David Wheeler <dwheeler@ida.org>:
237 * glib.h, gstrfuncs.c: added g_strlcpy and g_strlcat to support
238 safe manipulation of fixed-length string buffers.
239 These functions were originally developed by Todd Miller to simplify
240 development of security-related programs, and
241 are available on many (but not all) Unix-like systems,
242 including OpenBSD, FreeBSD, and Solaris. See
243 ftp://ftp.openbsd.org/pub/OpenBSD/src/lib/libc/string/strlcpy.3
244 and http://www.openbsd.org/security.html.
245 If there's a strlcpy/strlcat on the system, it's called, otherwise
246 an implementation is provided.
248 * testglib.c: Added tests for g_strlcpy, g_strlcat.
250 Wed Jul 26 05:03:24 2000 Tim Janik <timj@gtk.org>
252 * acglib.m4 (GLIB_SIZEOF): include <stdlib.h> and <stddef.h> if
253 STDC_HEADERS is defined.
256 * glibconfig.h: define gsize and gssize in terms of GLIB_SIZEOF_SIZE_T
258 * glib.h (g_return_if_reached): applied darin's fix for copy'n
259 paste error in the macro implementation.
261 Wed Jul 26 00:46:03 2000 Tim Janik <timj@gtk.org>
263 * glib.h: applied patch from Darin Adler <darin@eazel.com> which
264 supplies g_return_if_reached(), g_return_val_if_reached() and
267 2000-07-22 Tor Lillqvist <tml@iki.fi>
269 * build-dll: Fix resource handling, the resource file got left out
270 from the DLL after all... Remove the WIN32APIHEADERS, not needed
271 with current windres.
273 * glib.def: Add new functions.
275 2000-07-20 Sebastian Wilhelmi <wilhelmi@ira.uka.de>
277 * gutils.c, glib.h: Mark the functions g_basename and g_dirname
278 deprecated. They will issue an warning once, when compiled with
279 G_ENABLE_DEBUG, but continue to work as before. Instead the
280 functions g_path_get_basename and g_path_get_dirname should be
281 used, which BOTH return newly allocated memory, that has to freed
282 by g_free. The new g_path_get_basename now strips trailing slashes
283 from the path. This fixes #5097. For discussion see
284 http://mail.gnome.org/pipermail/gtk-devel-list/2000-April/003139.html
286 * gwin32.c, testglib.c, tests/dirname-test.c: Use the new
287 functions instead of the old ones.
289 * ghash.c, gscanner.c, glib.h: Mark the functions
290 g_hash_table_freeze, g_hash_table_thaw and thus
291 g_scanner_freeze_symbol_table and g_scanner_thaw_symbol_table
292 deprecated. They will issue an warning once, when compiled with
293 G_ENABLE_DEBUG. This fixes Bug #3883. For discussion see
294 http://mail.gnome.org/pipermail/gtk-devel-list/2000-April/003139.html
296 2000-07-19 Sebastian Wilhelmi <wilhelmi@ira.uka.de>
298 * configure.in, glib.h: glibconfig.h and glib.h now include files
299 outside of the extern "C" block. Makes some C++ compiler
300 happy. Reported by Denis Vakatov <vakatov@peony.nlm.nih.gov>.
302 Sat Jul 15 23:49:03 2000 Owen Taylor <otaylor@redhat.com>
304 * glib/glib.texi: Remove incomplete start of info file -
305 real docs are in RDP.
307 Sat Jul 15 22:44:22 2000 Owen Taylor <otaylor@redhat.com>
309 * configure.in: Add build/Makefile and
310 build/win32/Makefile to AC_OUTPUT() so things build
313 Sat Jul 15 09:11:46 2000 Tim Janik <timj@gtk.org>
315 * gstrfuncs.c (g_strncasecmp): fixed an off by 0 error (yeah,
316 the function went off when the while (n--) loop failed due to
317 n==0 ;), reported by Jean-Louis HAMEL <jlhamel@club-internet.fr>.
319 2000-07-15 Tor Lillqvist <tml@iki.fi>
321 * Makefile.am (SUBDIRS): Include the "build" module in GLib, too,
322 to make it more self-contained. If your CVS client doesn't
323 automatically get it, do a cvs get build in glib.
325 * */makefile.mingw.in: Include make.mingw from build in the glib
328 Fri Jul 14 16:26:35 2000 Owen Taylor <otaylor@redhat.com>
332 Fri Jul 14 12:22:49 2000 Owen Taylor <otaylor@redhat.com>
334 * configure.in Makefile.am glib-config.m4 glib.m4: Move
335 glib-config to glib-config-2.0 move glib.m4 to
338 * Makefile.am gobject/Makefile.am gmodule/Makefile.am
339 gthread/Makefile.am tests/Makefile.am: Change
340 library names to libglib-1.3.la, etc, so that we
341 can distinguish glib-1.2 and glib-2.0 on the linkline.
343 * Makefile.am gobject/Makefile.am gmodule/Makefile.am:
344 Move include files into /usr/include/glib-2.0.
346 Thu Jul 6 18:54:49 2000 Owen Taylor <otaylor@redhat.com>
348 * docs/Makefile.am (EXTRA_DIST): Remove info files
351 2000-07-14 Tor Lillqvist <tml@iki.fi>
353 * glib.def: Add g_error functions.
355 * makefile.mingw.in: Add gbacktrace.o.
357 * gbacktrace.c: No need to include <process.h>.
359 2000-07-12 Havoc Pennington <hp@redhat.com>
361 * glib.h: #include <gerror.h>
363 * Makefile.am (include_HEADERS): Add gerror.h
364 (libglib_la_SOURCES): Add gbacktrace.c
366 * gbacktrace.c: Move g_on_error_query() in here (moved on the
367 server, so history is preserved)
369 * gerror.h: GError interface
371 * gerror.c: GError implementation replaces stuff that's now in
374 Sun Jul 9 21:20:45 2000 Owen Taylor <otaylor@redhat.com>
376 * gunicode.h: Include stddef.h instead of stdlib.h
378 2000-07-08 Tor Lillqvist <tml@iki.fi>
380 * glib.h (GLIB_VAR): Rename the GUTILS_C_VAR macro to GLIB_VAR.
382 * gunicode.h: Mark the g_utf8_skip array with GLIB_VAR.
384 * glib.def: Add two missing entry points.
386 Thu Jul 6 15:35:28 2000 Owen Taylor <otaylor@redhat.com>
390 * Makefile.am (EXTRA_DIST): Dist fixes.
392 * configure.in: Moderate the warnings just a little bit.
394 2000-07-05 Tor Lillqvist <tml@iki.fi>
396 * README.win32: Update.
398 Mon Jul 3 17:58:02 2000 Owen Taylor <otaylor@redhat.com>
400 * gutf8.c (g_utf8_get_charset_internal): Fix up
401 to correspond to configure.in checks.
403 Mon Jul 3 17:18:19 2000 Owen Taylor <otaylor@redhat.com>
405 * glib.h: Comment g_get_codeset() out of the header file
406 temporarily. (Very similar to g_get_charset(), need
409 2000-07-01 Tor Lillqvist <tml@iki.fi>
411 * glib.def: Add new entry points.
413 * makefile.{mingw,msc}.in: Add the new Unicode object files.
415 Thu Jun 29 15:57:28 2000 Owen Taylor <otaylor@redhat.com>
419 * Makefile.am: added snapcheck target to go along with snapshot
421 * gstring.c glib.h (g_string_hash): Add g_string_hash to
422 go along with g_string_equal.
424 Tue Jun 27 12:40:23 EDT 2000 David A. Wheeler <dwheeler@dwheeler.com>
426 * glib.h: Added g_string_equal for comparing GStrings;
427 changed g_str_equal so it returns gboolean (instead of gint).
429 * gstring.c: Modified GString implementation to support embedded
430 ASCII NUL ('\0') characters, and implemented g_string_equal.
432 * testglib.c tests/string-test.c: Added tests for g_string_equal
433 and tests for proper handling of embedded ASCII NUL characters.
435 Wed Jun 28 22:52:00 2000 Owen Taylor <otaylor@redhat.com>
437 * Makefile.am (libglib_la_SOURCES): Fix
438 gunichartable.h => gunichartables.h. (From Eric Limings)
440 Fri Jun 23 17:20:26 2000 Tim Janik <timj@gtk.org>
442 * glib.h: define gstring in terms of gchar*. this typedef reflects
443 the type name of the primitive G_TYPE_STRING in the gobject module.
445 Wed Jun 21 12:09:03 2000 Owen Taylor <otaylor@redhat.com>
447 * gunicode.h gutf8.c guniprop.c gunidecomp.[ch] gunichartables.h
448 Makefile.am glib.h: Initial pass at adding unicode support
449 functions. A few things still need to be implemented, a bit
450 of cleanup needs to be done, tests need to be added, and
451 the docs need to be finished, but this should allow replacing
452 most or all use of libunicode.
454 2000-06-06 Tor Lillqvist <tml@iki.fi>
456 * giowin32.c (g_io_channel_win32_pipe_readable): If we are
457 watching the same pipe for different conditions (with different
458 callbacks), check them all. Only call the callback for G_IO_IN
459 from here. (This bug popped up when a watch for G_IO_ERR|G_IO_HUP
460 was added to gimplib.)
462 2000-05-30 Tor Lillqvist <tml@iki.fi>
464 * gutils.c (g_locale_get_codeset): Implement on Win32.
466 * glib.def: Add g_get_codeset.
468 * tests/Makefile.am (EXTRA_DIST): makefile.cygwin* has been
469 renamed to makefile.mingw*.
471 Tue May 30 16:01:32 2000 Owen Taylor <otaylor@redhat.com>
473 * glib.h gutils.c: Move the g_locale_get_codeset() up in the
474 header file to correspond to to comments about memory
475 management. Rename to g_get_codeset() to avoid
476 polluting the g_locale_* namespace, which probably
477 would have g_locale_get_codeset (GLocale *locale).
480 Mon May 29 14:10:35 2000 Owen Taylor <otaylor@redhat.com>
482 * gutils.c (g_locale_get_codeset): Add function to get the
483 codeset name for the current locale.
485 * configure.in acconfig.h: Add check for nl_langinfo(CODESET);
487 Fri May 19 11:39:29 2000 Tim Janik <timj@gtk.org>
489 * gutils.c (g_snprintf):
490 (g_vsnprintf): added argument assertments.
492 * gstring.c (g_string_assign): added argument assertments.
493 (g_string_truncate): make len a guint.
495 Fri May 19 09:00:44 2000 Tim Janik <timj@gtk.org>
497 * gmem.c (g_free): fixed SIZEOF_LONG==4 assumption with
498 ENABLE_MEM_CHECK, from Art Haas <ahaas@neosoft.com>.
500 * gslist.c (g_slist_reverse): shut up compiler.
502 * gscanner.c (g_scanner_get_token_ll): removed inline assignment.
504 * garray.c: remove index>=0 checks for unsigned indices.
506 * gmain.c (g_idle_prepare): timeout assignment fix.
508 * gtree.c (g_tree_node_rotate_right): shut up compiler.
510 2000-05-13 Tor Lillqvist <tml@iki.fi>
513 * tests/makefile.mingw.in
514 * build-dll: Rename makefile.cygwin(.in) to
515 makefile.mingw(.in), which better describes what it is. Move the
516 build of gmodule, gthread and gobject DLLs to makefiles in those
517 directories. Move resource file handling and build number bump to
518 build-dll, where it sits much cleaner.
521 * Makefile.am (EXTRA_DIST): Update accordingly.
523 * glib.h: Add G_PI, G_PI_2, G_PI_4, G_E, G_LN2, G_LN10 and
524 G_SQRT2. M_PI etc aren't necessarily in <math.h> in strict ISO C
527 * glib.def: Add g_strcanon.
529 * gtree.c (g_tree_node_rotate_left): Remove unused variables.
531 * gwin32.c (g_win32_opendir): Remove unneeded statement.
533 Thu May 4 02:04:46 2000 Tim Janik <timj@gtk.org>
535 * configure.in (STRIP_DUMMY): some Make 3.79 $(strip ) versions are
536 broken and require an empty arg, give it to them.
538 Fri Apr 28 23:54:35 2000 Tim Janik <timj@gtk.org>
540 * setup things for a new sub-library libgobject:
542 * Makefile.am (SUBDIRS): added gobject
544 * glib-config.in: feature -lgobject.
546 * configure.in (AC_OUTPUT): generate gobject/Makefile.
548 * glib.m4 (AM_PATH_GLIB): feature gobject module.
550 * glib.spec.in: added %{prefix}/lib/libgobject-1.3.so.*
552 Fri Apr 28 21:41:49 2000 Tim Janik <timj@gtk.org>
554 * glib.h: added G_STRLOC macro.
555 G_STRUCT_OFFSET(): signedness corrections.
556 (G_CSET_DIGITS): list 0-9.
557 * gscanner.c (g_scanner_config_template): use G_CSET_DIGITS.
563 (g_strreverse): return the modified string instead of void, so
564 calls to these functions can be nested.
565 (g_strcanon): new function, canonicalizes string according to
566 a given character set.
568 Fri Apr 28 19:45:16 2000 Tim Janik <timj@gtk.org>
570 * gasyncqueue.c (g_async_queue_unref): get rid of an unused variable.
572 Wed May 10 19:52:44 2000 Owen Taylor <otaylor@redhat.com>
574 * glib.m4: Print found version when test succeeds.
576 2000-05-04 Tor Lillqvist <tml@iki.fi>
579 * tests/makefile.cygwin.in: Include the common makefile snippet
582 Maybe CVSROOT/modules should be changed so that the 'build' module
583 is included within the glib module (and gtk+, and gimp, and maybe
584 others later), in the same way as the 'macros' module is included
585 in lots of GNOME CVS modules?
587 2000-05-02 Tor Lillqvist <tml@iki.fi>
589 * glib.def: Add new functions.
591 * makefile.{cygwin,msc}.in (glib_OBJECTS): Add new object files.
593 * tests/makefile.{cygwin,msc}.in: Add threadpool-test.
595 2000-04-28 Sebastian Wilhelmi <wilhelmi@ira.uka.de>
597 * gasyncqueue.c: New File implementing an asynchronous queue to be
598 used for asynchronous inter-thread communication.
600 * gthreadpool.c: New File implementing a thread pool to be used
601 for distributing work among several threads.
603 * glib.h: Added the type and function declarations for these two
606 * tests/threadpool-test.c: New File implementing a test for the
607 thread pool. This also checks the asynchronous queue underlying
610 * tests/Makefile.am: Changed accordingly.
612 2000-04-26 Sebastian Wilhelmi <wilhelmi@ira.uka.de>
614 * configure.in: Look for both pthread_create and pthread_join in
615 the thread library. Some systems define one of them, but not both
616 in libc. Arghh. Now we really start a thread and join it later and
617 check, whether the thread to actually ran.
619 * glib.h, gcache.c, gtree.c: Changed the 'value' parameter of
620 g_cache_remove from gpointer to gconstpointer. Dito for the 'key'
621 parameter of g_tree_lookup and g_tree_remove and the 'data'
622 parameter of g_tree_search. This function now takes a function of
623 type GCompareFunc instead of GSearchFunc. This fixes Bug
624 #8267. Thanks to Juan Toledo <toledo@users.sourceforge.net> for
627 * glib.h: Removed declaration of GSearchFunc.
629 * gmem.c: s/GSearchFunc/GCompareFunc/.
631 2000-04-19 Tor Lillqvist <tml@iki.fi>
633 * glib.def: Update entry point list.
635 2000-04-19 Sebastian Wilhelmi <wilhelmi@ira.uka.de>
637 * glib.h (G_TRYLOCK): Made the debugging G_TRYLOCK call also work
638 for compilers with funny G_STMT_(START|END) macros.
640 * tests/thread-test.c: Implemented a check for that.
642 * gutils.c (g_getenv): Changed the win32 part of this function to
643 be thread safe and to make the returned environment string
644 persistent to match the UN*X behavior. This is again a response to
647 * glib.h (G_LOCK_NAME): Removed parentheses around the lock name,
648 as that seems to cause problems for some compilers and really
651 Wed Apr 19 08:32:32 2000 Tim Janik <timj@gtk.org>
653 * gscanner.c (g_scanner_new): make sure that
654 scanner->config->cset_skip_characters is "" instead of NULL, so we
655 don't segfault further on.
657 2000-04-18 Sebastian Wilhelmi <wilhelmi@ira.uka.de>
659 * glib.h, glist.h, gslist.h: Changed the 'data' parameters from
660 gpointer to gconstpointer for the functions
661 g_(list|slist)_(remove|find|find_custom|index), as they do not
662 change this parameter. This fixes bug #4836.
664 * glib.h: Changed comment for g_getenv to reflect, that the
665 returned memory must not be freed. Fixes bug #8983.
667 2000-04-17 Sebastian Wilhelmi <wilhelmi@ira.uka.de>
669 * configure.in, acconfig.h: Add configure test for garbage
670 collector friendliness for GLib. If enabled, ENABLE_GC_FRIENDLY
673 * garray.c, ghash.c, glist.c, gmain.c, gmem.c, gnode.c, gqueue.c,
674 gslist.c, gtree.c: If ENABLE_GC_FRIENDLY is defined, NULLify all
675 memory released by the user, but cached by GLib. This lets a
676 garbage collector have a more correct view of the actually used
679 * garray.c, glib.h: Added g_(array|ptr_array|byte_array)_sized_new
680 functions, that reserve a certain amount of memeory for the array
681 at creation time to avoid reallocation. Fixes bug #6707 from
682 Charles Kerr <ckerr@osserver1.nssl.noaa.gov>.
684 * glib.h, gqueue.c, tests/queue-test.c (main): Renamed
685 g_queue_create to g_queue_new in conformance to all other GLib
688 2000-04-07 Sebastian Wilhelmi <wilhelmi@ira.uka.de>
690 * grand.c (g_rand_new): Fixed bug. Thanks to Marko Kreen
691 <marko@l-t.ee> for reporting that.
693 2000-03-26 Tor Lillqvist <tml@iki.fi>
695 * README.win32: Tell about using the mingw-based gcc, which is
696 much easier than modifying the cygwin gcc to product mingw code
697 for the msvcrt runtime.
699 * makefile.cygwin.in (WIN32APIHEADERS): Kludge to make it work
700 with a "pure" mingw gcc, too.
702 2000-03-24 Sebastian Wilhelmi <wilhelmi@ira.uka.de>
704 * garray.c: Made GArray behave correctly. Now zero_terminated
705 really means, that the element array->data[array->len] exists and
706 is zeroed, and clear means that any unassigned elements obtained
707 through g_array_set_size (the only way to get unassigned elements
708 AFAICT) are zeroed. Added some macros to make the code more
709 obvoius. Also made GPtrArray zero elements after
710 g_ptr_array_set_size. This is done in a portbale way (assignment
711 of NULL instead of just memsetting it to zero), though that might
712 be more portability than we actually want.
714 * Makefile.am, gthread/Makefile.am, gmodule/Makefile.am,
715 tests/Makefile.am: Added various win32 related *.in files to
716 EXTRA_DIST to let 'make distcheck' procude all the corresponding
717 files, which it silently fails to do currently.
719 2000-03-23 Sebastian Wilhelmi <wilhelmi@ira.uka.de>
721 * configure.in: After finding the right thread library (containing
722 e.g. pthread_create) we now search for the right realtime library
723 (containing e.g. sched_get_priority_max). Makes the output of the
724 thread related libraries correct.
726 * gtimer.c (g_usleep): The current implementation of g_usleep
727 (simply calling select) doesn't work reliable for multi-threaded
728 programs on some platforms (bad omen for the main loop....), so I
729 changed the implementation for thread-using programs to wait for a
730 GCond for the specified amount of time (NB: sleep and usleep are
731 not MT-safe in general, because they often use signals).
733 Wed Mar 22 16:49:57 2000 Owen Taylor <otaylor@redhat.com>
735 * gmem.c (g_mem_chunk_area_compare): Fix indentation.
737 2000-03-22 Elliot Lee <sopwith@redhat.com>
739 * gmem.c (g_mem_chunk_area_compare): Fix 64-bitness bug in
740 comparing two pointers more than 4G apart.
742 2000-03-22 Tor Lillqvist <tml@iki.fi>
744 * gutils.c: Move Win32-only includes after inclusion of glib.h, so
745 that G_OS_WIN32 is defined.
747 * glibconfig.h.win32.in: Add GSystemThread.
749 2000-03-22 Sebastian Wilhelmi <wilhelmi@ira.uka.de>
751 * gutils.c (g_get_current_dir): Ok, Marcus Brinkmann
752 <Marcus.Brinkmann@ruhr-uni-bochum.de> convinced me, that 128 KB
753 path length might not be enough for the HURD. So I changed the
754 loop to at least avoid an integer overflow, which could happen at
757 * configure.in: Test for sched_yield as the native yield function
758 first. Corrected typo g_thread_sleep -> g_usleep. Corrected
759 message for the pthread_create test. Negative Priorities are
760 allowed (and used on Solaris), so consider
761 sched_get_priority_min failed only if it returns -1, not <0. Check
762 for sched_get_priority_min also in -lrt, if not found in -lpthread
763 alone and add -lrt to G_THREAD_LIBS then. Remove special case
764 handling of priorities for older solaris versions and posix
765 threads. Thanks to Wan-Teh Chang <wtc@netscape.com> for suggesting
766 some of those changes.
768 * config.guess, config.sub, ltconfig, ltmain.sh:
769 Upgrade to libtool 1.3.4.
771 2000-03-21 Sebastian Wilhelmi <wilhelmi@ira.uka.de>
773 * glib.h, configure.in, gutils.h: always define G_GNUC_EXTENSION,
774 even when not needed by GLib. That's actually also the way, the
775 GLib reference manual describes that macro. Therefore I had to
776 remove the lonesome #include <glibconfig.h> in gutils.c, which
777 doesn't seem to be needed there however. This change should make
778 Ben Gertzfield <che@debian.org> happy.
780 * gutils.c: Furthermore two warnings in gutils.c were voided,
781 which crept in due to my last change.
783 * gutils.c (g_get_current_dir): Allocate only up to 128KB for a
784 pathname. While this is an arbitrary value just like 2048, it
785 seems to be enough (after all, even 4GB is an arbitrary value).
787 2000-03-20 Sebastian Wilhelmi <wilhelmi@ira.uka.de>
789 * gmain.c (g_main_poll): Warn in case of an error during the call
790 to poll(2). Closes Bug#7564 as reported by David Helder
793 * gutils.c (g_get_current_dir): Make g_get_current_dir work on
794 systems with unlimited pathname length like the HURD (It worked
795 there before, but only for pathes shorter than 2048). Closes
796 Bug#4525 as reported by Marcus Brinkmann
797 <Marcus.Brinkmann@ruhr-uni-bochum.de>.
799 2000-03-17 Sebastian Wilhelmi <wilhelmi@ira.uka.de>
801 * giounix.c (g_io_unix_write, g_io_unix_read): Interpret EINTR as
804 2000-03-17 Sebastian Wilhelmi <wilhelmi@ira.uka.de>
806 * configure.in: Added the missing POSIX_NO_YIELD and
807 POSIX_NO_PRIORITIES warning messages.
809 * configure.in: Use AC_TRY_RUN instead of AC_TRY_LINK, to test for
810 real thread support. On solaris pthread_create can be linked to
811 even in -lc, but it doesn't work then.
813 * configure.in: Don't use priorities for threads, when the
814 minimal/maximal priorities couldn't be determined at configure
817 * configure.in, gthread.c: Always define GSystemThread in
818 glibconfig.h to represent a system thread.
820 * configure.in: Do not use native recursive threads, when
821 possibe. We use some features, that they do not expose (namely the
824 * glib.h, gthread.c: Redefined GStaticRecMutex. The functions are
825 now implemented in a different way, which should be way
826 faster. Alsothere are now functions g_static_rec_mutex_unlock_full
827 and g_static_rec_mutex_lock_full to leave/enter a recursive mutex
830 * gthread.c (g_thread_self): Do not test the system_thread to be
831 non-zero to speed things up.
833 * gthread.c (g_mutex_init): Therefore set the system_thread of the
836 * tests/thread-test.c: Rerun all tests once again, but this time
837 we fool the system into thinking, that the available thread system
838 is not native, but userprovided.
840 2000-03-13 Sebastian Wilhelmi <wilhelmi@ira.uka.de>
842 * gqueue.c (g_queue_push_tail_link, g_queue_push_head_link): We
843 want the next and prev pointer of the inserted link to be NULL.
845 2000-03-06 Sebastian Wilhelmi <wilhelmi@ira.uka.de>
847 * configure.in: Another small change to the pthread_.. search
848 pattern. Should work *now* for AIX.
850 2000-03-04 Tor Lillqvist <tml@iki.fi>
852 * gwin32.c (g_win32_error_message): New function that returns the
853 message string for a Win32 error code.
855 * glib.h: Declare it.
857 * glib.def: Export it, plus g_node_copy.
859 2000-03-03 Sebastian Wilhelmi <wilhelmi@ira.uka.de>
861 * configure.in: Make the search for pthread_attr_... prototypes
862 find names at the start of a line also, like it is on AIX. Thanks
863 to Valdis Kletnieks <Valdis.Kletnieks@vt.edu> for the info.
865 Wed Mar 1 10:39:39 2000 Tim Janik <timj@gtk.org>
867 * gslist.c (g_slist_reverse): minor optimization.
869 * testglib.c (g_node_test): added a couple of tests for
873 * gnode.c (g_node_copy): new function to copy subtrees,
874 supplied by dbsears@ix.netcom.com.
875 changed iterator to walk the children list backwards, so
876 we get down from O(n^2) to O(n).
878 * gnode.c (g_node_first_sibling): applied patch from
879 dbsears@ix.netcom.com to optimize access if node->parent
882 * gutils.c (g_get_any_init): backed out HAVE_PW_GECOS check around
883 assignment of g_real_name, sicne HAVE_PW_GECOS is never defined and
884 thus breaks the original code.
886 * merged changes from 1.2.7.
888 Fri Jan 28 11:37:41 2000 Owen Taylor <otaylor@redhat.com>
890 Bug #4156 - Changes vaguely modelled after Scott Gifford's patch
892 * gtimer.c (g_timer_elapsed): Never report negative times -
895 * gmain.c (g_timeout_prepare): Guard against unexpected
896 clock shifts by never setting a timeout of more than
897 data->interval msecs.
899 2000-02-27 Tor Lillqvist <tml@iki.fi>
901 * glib.def: Add new functions.
903 2000-02-23 Tor Lillqvist <tml@iki.fi>
905 * README.win32: Add a missing step to the setup instructions for
906 gcc-2.95.2. Thanks to Arnaud Charlet.
908 * glib.def: Add missing entry point.
910 2000-02-18 Sebastian Wilhelmi <wilhelmi@ira.uka.de>
912 * configure.in: Changed GCC version test to also accept major
913 versions > 2. Thanks to Ben Gertzfield <che@debian.org> for
916 Thu Feb 17 12:53:44 2000 Tim Janik <timj@gtk.org>
918 * gstring.c: changed g_str_hash() to a 31 bit version based on
919 a submission by Karl Nelson and hand optimized ad absurdum by
922 * gstring.c: applied patch from havoc for new gstring functions,
923 added some more sanity checks, coding style fixups.
925 2000-02-13 Havoc Pennington <hp@pobox.com>
927 * tests/string-test.c (main): Add tests for the new GString
930 * testglib.c (main): Add tests for the new GString features
932 * gstring.c (g_string_insert_len): New function; insert
933 a given length of string at a given position.
934 (g_string_append): reimplement in terms of g_string_insert_len
935 (g_string_append_len): new function
936 (g_string_insert_c): accept -1 for "pos" arg to mean "append"
937 (g_string_append_c): reimplement in terms of g_string_insert_c
938 (g_string_prepend): reimplement in terms of g_string_insert_len
939 (g_string_prepend_len): new function
940 (g_string_prepend_c): reimplement in terms of g_string_insert_c
941 (g_string_insert): reimplement in terms of g_string_insert_len
943 * glib.h: Declare g_string_insert_len, g_string_append_len,
946 Sun Feb 13 08:16:47 2000 Tim Janik <timj@gtk.org>
948 * configure.in: wtf??? someone destroyed the configure.in, reverting to
949 an older version from Feb 4 which apears to still work.
951 2000-02-07 Tor Lillqvist <tml@iki.fi>
953 * gmodule.rc.in gthread.rc.in: Move to corresponding subdirectories.
956 * gmodule/Makefile.am
957 * gthread/Makefile.am: Change accordingly.
959 * makefile.cygwin: Corresponding changes, some cleanup.
961 2000-02-05 Tor Lillqvist <tml@iki.fi>
963 * glib.rc.in gmodule.rc.in gthread.rc.in: New files, for putting
964 version info in the DLLs on Win32.
966 * Makefile.am: Generate corresponding *.rc files and distribute
969 * makefile.cygwin.in: Add rules to automatically bump a "build
970 number" in the version info in the rc files each time the DLL is
971 built. But do this only for the person who releases binaries. If
972 others build the DLLs, the build number is set to zero.
974 Fri Feb 4 19:36:05 2000 Tim Janik <timj@gtk.org>
977 * gdataset.c: return stolen data from g_datalist_id_remove_no_notify()
978 and g_dataset_id_remove_no_notify() to avoid second lookup for common
981 2000-02-01 Tor Lillqvist <tml@iki.fi>
984 * gstrfuncs.c (g_filename_to_utf8, g_filename_from_utf8): New
985 functions for conversion between UTF-8 and the encoding expected
986 by C runtime functions like open() and stat(), and returned by
989 Implement them on Win32 where we use the system "ANSI" codepage,
990 which might be single-byte or double-byte. On Unix, just skip the
991 issue for now and provide dummy implementations that return a copy
996 * glib.def: Minor updates.
998 Wed Jan 26 05:24:38 2000 Tim Janik <timj@gtk.org>
1001 * gmain.c: s/current_time/dispatch_time/ for the dispatch() handlers.
1002 refetch the current time after invocation of poll() to cover up for
1003 the time spent in that function call.
1005 Fri Jan 21 10:18:24 2000 Owen Taylor <otaylor@redhat.com>
1007 * glib.h (G_N_ELEMENTS): Added G_N_ELEMENTS macro to determine
1008 the number of elements in an array.
1010 Sun Jan 9 13:28:36 2000 Tim Janik <timj@gtk.org>
1012 * gstrfuncs.c (g_strtod): correctly fetch the current locale,
1015 1999-12-16 Tor Lillqvist <tml@iki.fi>
1017 * gmodule/gmodule-win32.c: Use FormatMessage to translate system
1018 error codes into textual messages.
1020 1999-11-25 Sebastian Wilhelmi <wilhelmi@ira.uka.de>
1022 * glib.h (G_TRYLOCK): This of course should return TRUE in a
1023 program with a thread-disabled GLib.
1025 1999-11-18 Tor Lillqvist <tml@iki.fi>
1027 * glib.def: g_strjoin was missing.
1029 1999-11-16 Sebastian Wilhelmi <wilhelmi@ira.uka.de>
1031 * acconfig.h, config.h.win32.in, configure.in: Renamed
1032 GLIB_SIZEOF_PTHREAD_T to GLIB_SIZEOF_SYSTEM_THREAD to reflect
1035 * configure.in: Cope with systems, that have a pthread_t type,
1036 that is not a pointer. Hint from Karl Nelson
1037 <kenelson@ece.ucdavis.edu>. Define GLIB_SIZEOF_SYSTEM_THREAD to 4
1038 for Solaris. Cope with systems, that have no default mutex
1039 initialize, like obviously most DCE systems.
1041 * glib.h, gthread.c: Changed the prototype of thread_create and
1042 thread_self to return the system thread into provided memory
1043 instead of a return value. This is necessary, as HPUX has a
1044 pthread_t, that is bigger than the biggest integral type there.
1046 * gthread.c: system_thread is no longer a pointer, but an memory
1047 area of size GLIB_SIZEOF_SYSTEM_THREAD. Changed the
1048 zeroinitialization and the tests for zeroness accordingly.
1050 1999-11-09 Sebastian Wilhelmi <wilhelmi@ira.uka.de>
1052 * configure.in: Create docs/glib-config.1 from
1053 docs/glib-config.1.in. Makes 'make distcheck' happy (and me too).
1055 * glib-config.1: Removed from CVS, as it is a generated file.
1057 1999-11-08 Sebastian Wilhelmi <wilhelmi@ira.uka.de>
1059 * configure.in: Make the test for getpwuid_r work on newer AIX
1060 versions, too. Still works on Solaris and Linux. Patch from Craig
1061 Rodrigues <rodrigc@mediaone.net>.
1063 1999-11-08 Tor Lillqvist <tml@iki.fi>
1065 * gwin32.c (g_win32_getlocale): Look at env vars LC_ALL, LC_CTYPE
1066 and LANG first. Some refinements to the sublanguage logic.
1068 1999-11-04 Tor Lillqvist <tml@iki.fi>
1070 * makefile.{cygwin,msc}.in: Add gwin32 object. Add rule to make .i
1071 (preprocessed source) files.
1073 1999-11-01 Tor Lillqvist <tml@iki.fi>
1076 * glib.def: Rename Win32-only functions from gwin_* to g_win32_*
1077 to match the GLib naming conventions.
1082 * Makefile.am: Move the Win32-only functions to the new
1085 1999-10-31 Tor Lillqvist <tml@iki.fi>
1087 * gutils.c (gwin_getlocale): New Win32-specific function, returns
1088 a Unixish current locale string (en, zh_TW etc).
1090 * glib.h: Declare it.
1092 * glib.def: Export it.
1094 * testglib.c: Test it.
1096 * gmessages.c (Win32: ensure_stdout_valid): Some improvements,
1097 make sure we don't call AllocConsole several times, which I think
1100 Sun Oct 31 18:55:01 1999 ape@spacetec.no (Asbjorn Pettersen)
1102 * gcache.c (g_cache_remove): Test if node is NULL.
1103 If not tested, GIMP's script-fu will crash.
1105 Sun Oct 17 18:11:40 1999 Tim Janik <timj@gtk.org>
1107 * gdataset.c (g_data_set_internal): remove g_dataset_global_lock around
1108 destroy() notification here as well.
1110 1999-10-15 Sebastian Wilhelmi <wilhelmi@ira.uka.de>
1112 * gdataset.c (g_datalist_clear_i): Avoid Freezing, when g_datalist
1113 is called recursivly. Reported by Ola Andersson <rand@ling.umu.se>.
1115 Tue Oct 12 14:17:12 1999 Tim Janik <timj@gtk.org>
1117 * glib.h: removed useless g_string(x) macro that cluttered the namespace
1118 and was just a poor wrapper around the cpp '#' symbol, use #x if you
1119 need to work around this.
1120 added new macro G_STRINGIFY(arg) that will convert arg to a string,
1121 no matter whether it contains macros or not.
1123 1999-10-12 Tor Lillqvist <tml@iki.fi>
1125 * config.h.win32.in: Define the new GLIB_SIZEOF_* constants here,
1128 * glib.h: Small Win32 comments improvement.
1130 Tue Oct 12 12:16:12 1999 Tim Janik <timj@gtk.org>
1132 * gmessages.c (g_printf_string_upper_bound): completly new implementation
1133 for printf string upper bounds calculation.
1134 we handle all glibc 2.1 format specifiers now, except for positional
1135 parameters (%nn$...) and wide char strings, plus some obscure upper
1136 case variants of the standard conversions. this fixes a lot of
1137 bugs in the old code, i.e.
1138 - NULL format strings
1139 - floats with exponents >+24
1141 - precision specifications in general
1142 - negative field widths
1143 - %p for SIZEOF_VOID_P > 4 platforms
1144 we now issue warnigns in places where the old code would have
1145 caused buffer overruns anyways. warnings are suppressed when invoked
1146 from glogv(), to avoid infinite recursions if someone passes a log
1147 message that comes with really obscure format specifications.
1149 Tue Oct 12 11:49:00 1999 Tim Janik <timj@gtk.org>
1151 * gstrfuncs.c: nuked old g_printf_string_upper_bound() version.
1153 Tue Oct 12 03:34:40 1999 Tim Janik <timj@gtk.org>
1155 * glib.h: added GFloatIEEE754 and GDoubleIEEE754 unions to access sign,
1156 mantissa and exponent of IEEE floats and doubles (required by the new
1157 version of g_printf_string_upper_bound). the unions are endian specific,
1158 we handle G_LITTLE_ENDIAN and G_BIG_ENDIAN as of currently. ieee floats
1159 and doubles are supported (used for storage) by at least intel, ppc and
1161 http://twister.ou.edu/workshop.docs/common-tools/numerical_comp_guide/ncg_math.doc.html
1163 Mon Oct 11 18:01:49 1999 Tim Janik <timj@gtk.org>
1165 * configure.in: added additional checks to figure sizes of size_t,
1166 ptrdiff_t and intmax_t (required by g_printf_string_upper_bound).
1168 Wed Oct 6 12:44:23 PDT 1999 Manish Singh <yosh@gimp.org>
1170 * configure.in: blah. use G_WITH_CYGWIN instead of G_HAVE_CYGWIN
1172 1999-10-05 Tor Lillqvist <tml@iki.fi>
1174 * glib.h: (Win32) Drop the mapping of POSIX function names to the
1175 underscored versions, it's unnecessary after all. With MSVC we get
1176 them from oldnames.lib, with gcc-2.95 and mingw32 from
1177 -lmoldname-msvc. Add comment about what headers to include for
1180 * glibconfig.h.win32.in: Don't define WIN32 and NATIVE_WIN32.
1182 * gerror.c (g_on_error_query): (Win32) Slightly increased verbosity.
1184 * build-dll: Don't strip.
1186 * tests/string-test.c tests/dirname-test.c: Use G_OS_WIN32.
1188 * glib.def: Add g_thread_use_default_impl.
1190 Sun Oct 3 19:46:55 PDT 1999 Manish Singh <yosh@gimp.org>
1192 * configure.in: use G_HAVE_CYGWIN instead of G_OS_FEATURE_CYGWIN
1194 Sun Oct 3 19:25:42 PDT 1999 Manish Singh <yosh@gimp.org>
1198 * glibconfig.h.win32: G_OS_FOO #defines. I *think* I got the cygwin
1199 and beos stuff right, but I haven't tested it. The respective
1200 porters should fix any screwups
1210 * testglib.c: use G_OS stuff
1212 Wed Sep 22 01:53:18 1999 Tim Janik <timj@gtk.org>
1214 * glib.h (NULL): define NULL as (0L) if __cplusplus is defined, to
1215 avoid "ANSI C++ forbids implicit conversion from `void *' in argument
1216 passing" errors upon NULL usage in C++ programs (gcc-2.95 is on crack
1217 for erroring out on this, instead of just issueing a warning).
1219 * glib.h (g_trash_stack_pop): use uncasted NULL again.
1221 Fri Sep 17 10:24:45 1999 Tim Janik <timj@gtk.org>
1223 * gmem.c (g_mem_chunk_compute_size) (g_mem_chunk_new): applied patch
1224 from Soeren Sandmann <sandmann@daimi.au.dk>, to force mem chunk's area
1225 sizes to be a multitiple of atom_size, and to eliminate the MAX_MEM_AREA
1226 restriction of 65536 bytes. we also catch cases where users pass an area
1227 size < atom size with a return_if_fail statement now (which is ok,
1228 because previously this lead to memory corruption anyways).
1230 Thu Sep 16 13:19:54 1999 Tim Janik <timj@gtk.org>
1232 * glib.h (g_trash_stack_pop): add explicit (GTrashStack*) cast for NULL
1233 pointer to cure ANSI C++ error.
1235 Mon Sep 13 23:25:59 1999 Tim Janik <timj@gtk.org>
1237 * gmessages.c (g_logv): in case we have to abort the program,
1238 debugging is enabled and we are not called recursively, try
1239 to abort with raise (SIGTRAP) first, so developers may ignore
1240 certain failure conditions during debugging stage.
1242 Thu Aug 26 15:09:36 1999 Tim Janik <timj@gtk.org>
1245 * gmodule/Makefile.am:
1246 * gthread/Makefile.am: added --export-dynamic so we can load dynmic
1247 modules, (required, according to the libtool 1.3.3 docu).
1249 1999-07-23 Sebastian Wilhelmi <wilhelmi@ira.uka.de>
1251 * grand.c (g_rand_new): Use /dev/urandom, as it doesn't block,
1252 which /dev/random might do. Do not XOR the time, when getting the
1253 seed form /dev/urandom, as this is good itself. Prevent the
1254 initial seed from being zero, which causes the PRNG to produce
1255 only zeros. Hints from Colin Plumb <colin@pgp.com>.
1257 1999-08-17 Tor Lillqvist <tml@iki.fi>
1259 * glib.h (g_trash_stack_push): Add a cast.
1262 * glist.c: Make the inline functions static inline, and add
1263 separate extern wrappers. Not all compilers produce callable entry
1264 points for inline functions, even if gcc does.
1266 Sun Aug 15 02:47:14 1999 Tim Janik <timj@gtk.org>
1268 * glib.h (g_trash_stack_pop): eliminate memset() call, since string.h
1269 has not neccessarily been included prior to glib.h.
1271 Mon Aug 2 21:03:10 1999 Tim Janik <timj@gtk.org>
1273 * configure.in: added --enable-msg-prefix option.
1275 * gmessages.c (g_log_default_handler): feature "prg_name (pid:%u): "
1276 if --enable-msg-prefix was selected (use "(process:%u): " if
1277 g_get_prgname () returns NULL, along the lines of g_on_error_query).
1279 1999-08-03 Tor Lillqvist <tml@iki.fi>
1283 * tests/strfunc-test.c: Rename g_strccpy to g_strcompress and
1284 g_strecpy to g_strescape per Tim Janik's suggestion. Dropped the
1285 destination parameter, always g_malloc a new string. Fix bug in
1286 g_strcompress, octal digits were gobbled up without limit, should
1289 Sources that use g_strescape must have ifdefs to be compilable
1290 both with GLib 1.2 and 1.3.
1292 Sat Jul 31 17:52:03 PDT 1999 Manish Singh <yosh@gimp.org>
1295 * gstrfuncs.c: the #define for g_strescape interfered with the
1296 compilation of the function, so just remove the function and
1297 note that it's deprecated in the header
1299 1999-08-01 Tor Lillqvist <tml@iki.fi>
1301 * gstrfuncs.c (g_strccpy, g_strecpy): New functions.
1303 * glib.h: Declare and document them. Define the deprecated
1304 g_strescape as a macro that calls g_strecpy.
1306 * tests/strfunc-test.c (main): Test them.
1308 * makefile.{cygwin,msc}.in
1309 * tests/makefile.{cygwin,msc}.in: Remove gstack and its test
1312 * glib.def: Additions and removals.
1314 * README.win32: Improve gcc build instructions.
1316 * build-dll: Also build import library for MSVC.
1318 Sat Jul 24 20:11:35 1999 Tim Janik <timj@gtk.org>
1320 * merged GLib 1.3.0 with glib-1.2.3 from Fri Jul 16 22:18:36.
1321 * incorporated proposed cleanups from gtk-devel-list.
1323 * bumped version number to GLib-1.3.1
1329 removed string tokenisation (we got g_strsplit() and g_strjoin()
1330 already) and readline functions.
1331 s/g_list_delete/g_list_delete_link.
1332 implemented g_slist_delete_link.
1333 removed notion of g_ATEXIT() macro in glib.h, this is an *internal*
1334 macro, g_atexit() is provided for public consumption.
1335 added GTrashStack inline utility functions.
1336 reimplement double eneded queues.
1337 removed GStack implementation, people can use a queue or a (singly)
1338 linked list for this task.
1339 deprecated g_strescape(), we need the SunOS variants here.
1341 * gdate.c: added DEBUG_MSG() macro to wrap old messages.
1345 * upgrade to libtool 1.3.3.
1347 1999-07-21 Tor Lillqvist <tml@iki.fi>
1349 Win32: With the latest gcc (2.95, pre-release), we can have binary
1350 compatibility with MSVC by using the switch -fnative-struct. No
1351 longer build DLLs with .gcc in the name when using gcc.
1353 * README.win32: Renew gcc build instructions.
1355 * build-dll: Comments change, handle also .a files.
1357 * makefile.cygwin.in
1358 * tests/makefile.cygwin.in: Remove .gcc from DLL name.
1360 1999-07-13 Tor Lillqvist <tml@iki.fi>
1362 * README.win32: Correct URL for mingw runtime sources.
1364 * build-dll: Combine commands with &&.
1366 * glib.h: Map also rmdir() and hypot() for MSVCRT library.
1368 * makefile.cygwin.in
1369 * tests/makefile.cygwin.in: New DLL naming style. GCC-compiled DLLs are
1370 now called *.gcc.dll, to avoid binary incompatibilities with
1371 MSVC-compiled versions.
1373 * makefile.msc.in: Cosmetics.
1375 1999-07-07 Sebastian Wilhelmi <wilhelmi@ira.uka.de>
1377 * configure.in: Test for pthread_join rather than for
1378 pthread_create to determine the right thread-lib. Makes it work on
1379 mips-sgi-irix6.5. Hitn from to Jari Vuoksenranta
1380 <javu@piano.ux.phys.jyu.fi>.
1382 1999-07-02 Tor Lillqvist <tml@iki.fi>
1384 * README.win32: Note about need to fix another bug in the mingw32
1387 * makefile.msc.in: Debugging turned on via an nmake variable,
1388 no need to edit the makefile.
1390 1999-07-01 Sebastian Wilhelmi <wilhelmi@ira.uka.de>
1392 * configure.in, acconfig.h, gutils.c: Added a g_memmove
1393 replacement for platforms without memmove, where bcopy can't
1394 handle overlapping copies and the corresponding checks, which is
1395 taken form the PERL Configure routine.
1397 * glib.h: Updated the commentary about g_memmove to be right and
1400 * configure.in: Removed test for rand_r, as it isn't used anymore.
1402 1999-06-30 Sebastian Wilhelmi <wilhelmi@ira.uka.de>
1404 * glib.h, grand.c: Finally removed the g_random_normal and
1405 g_rand_normal functions.
1407 1999-06-28 Tor Lillqvist <tml@iki.fi>
1409 * glib.def: Add missing export of g_strncasecmp.
1411 1999-06-21 Jose Mercado <jmercado@mit.edu>
1413 * glib.spec.in: Changed version number (1.1->1.3) in files section
1414 to allow rpm to build packages again.
1416 1999-06-21 Tor Lillqvist <tml@iki.fi>
1418 * README.win32: Update the pthreads snapshot version we want.
1419 Advice how to hand-expand the makefile.*.in files.
1421 * config.h.win32.in: Define values needed by Sebastian Wilhelmi's
1424 * glib.def: Add new functions.
1426 * glibconfig.h.win32.in: Update the pthreads snapshot version.
1429 * gthread.c: Include config.h, guard inclusion of unistd.h. When
1430 using gcc on Win32, g_thread_functions_for_glib_use must be marked
1431 for export here, too.
1433 * gtimer.c: Implement g_usleep on native Win32 using Sleep (which
1434 only has millisecond granularity, though).
1436 * makefile.cygwin.in
1437 * makefile.msc.in: Update pthreads snapshot version. File
1438 name changes. Remove testgthread.
1440 * tests/makefile.cygwin.in
1441 * tests/makefile.msc.in: Add thread-test. Link with gthread lib.
1443 1999-06-18 Jeff Garzik <jgarzik@pobox.com>
1445 * tests/Makefile.am: Re-order tests in alpha order.
1447 1999-06-18 Sebastian Wilhelmi <wilhelmi@ira.uka.de>
1449 * configure.in: Changed test for pthread_attr_setstacksize from
1450 AC_TRY_COMPILE to AC_TRY_LINK.
1452 1999-06-17 Sebastian Wilhelmi <wilhelmi@ira.uka.de>
1454 * configure.in, acglib.m4, acconfig.h, glib.h, gthread.c:
1455 Completed the thread support in GLib. Thread creation,
1456 prioritizing threads, yielding, joining threads as well as
1457 reader/writer locks and recursive mutexes are now in place. Please
1458 test heavily on your platform. It is so far tested on
1459 Linux/i386/pthreads, Solaris/Sparc/pthreads and
1460 Solaris/Sparc/solaristhreads.
1462 * gtimer.c, glib.h: Implement g_usleep (gulong microseconds) for
1463 thread safe sleeping. (sleep() is not MT-safe at all!)
1465 * gutils.c: Avoid compiler warning.
1467 * tests/Makefile.am, tests/thread-test.c: New program to test some
1468 aspects of the thread implementation.
1470 * gthread.c, Makefile.am: Renamed from gmutex.c to reflect the
1473 * configure.in: Purged all appearances of nspr.
1475 Wed Jun 2 11:42:46 PDT 1999 Manish Singh <yosh@gimp.org>
1481 * ltmain.sh: upgrade to libtool 1.3.2 (BeOS changes merged)
1483 1999-05-29 Tor Lillqvist <tml@iki.fi>
1485 * gstrfuncs.c (g_strescape): Backslashify also '"' characters.
1487 * glib.h: Document g_strescape.
1489 1999-05-12 Tor Lillqvist <tml@iki.fi>
1491 * glib.h (Win32): Map fileno to _fileno for mingw32. Map fstat to
1494 * README.win32: Advice also to remove -lmoldname in the
1495 patch to the egcs-1.1.2 spec file.
1497 Wed May 12 00:23:55 CDT 1999 Shawn T. Amundson <amundson@gtk.org>
1499 * gmodule/Makefile.am: Another small fix.
1501 1999-05-08 Tor Lillqvist <tml@iki.fi>
1503 * Makefile.am tests/Makefile.am: Correct rules for making the
1504 win32-related files that are made from corresponding .in files.
1505 Is there a cleaner way than explicitly writing rules that invoke
1508 Fri Jul 16 22:18:36 PDT 1999 Manish Singh <yosh@gimp.org>
1511 * ltmain.sh: upgrade to libtool 1.3.3
1513 1999-06-18 Sebastian Wilhelmi <wilhelmi@ira.uka.de>
1515 * gdate.c, gstrfuncs.c, gstring.c: Fixed the use of the
1516 is..... and to..... macros, which take unsigned chars, not chars!
1517 Thanks to Morten Welinder <terra@diku.dk> for pointing this out.
1519 Thu Jun 3 16:30:31 PDT 1999 Manish Singh <yosh@gimp.org>
1521 * gerror.c (g_on_error_query): check isatty() before querying so
1522 we don't loop endlessly
1524 Sat May 29 11:16:29 PDT 1999 Manish Singh <yosh@gimp.org>
1530 * ltmain.sh: upgrade to libtool 1.3.2
1532 1999-05-26 Sebastian Wilhelmi <wilhelmi@ira.uka.de>
1534 * gmain.c: provide a poll prototype for SunOS, as they do not do
1535 it self. Hint from Christian Parg <cparg@fs-design.de>.
1537 Tue May 25 12:23:07 1999 Owen Taylor <otaylor@redhat.com>
1539 * gstrfuncs.c (g_strchug): Use g_memmove() not memmove().
1540 (Reported by Charles Levert <charles@comm.polymtl.ca>)
1542 Mon May 10 22:03:52 CDT 1999 Shawn T. Amundson <amundson@gtk.org>
1544 * Released GLib 1.2.3
1546 1999-05-08 Tor Lillqvist <tml@iki.fi>
1548 * Makefile.am tests/Makefile.am: Correct rules for making the
1549 win32-related files that are made from corresponding .in files.
1550 Is there a cleaner way than explicitly writing rules that invoke
1553 Sat May 1 10:18:01 PDT 1999 Manish Singh <yosh@gimp.org>
1559 * ltmain.sh: upgrade to libtool 1.3
1561 1999-04-30 Tor Lillqvist <tml@iki.fi>
1563 * Makefile.am: Don't distribute glibconfig.h.win32.in, but
1564 glibconfig.h.win32. Generate it when making a dist. Also generate
1565 makefile.msc and config.h.win32 from corresponding .in files when
1568 * configure.in: Also substitute @GLIB_INTERFACE_AGE@ and
1569 @GLIB_BINARY_AGE@ (needed in config.h.win32).
1571 * glibconfig.h.win32.in: Use static mutex structure and initial
1572 value corresponding to the 1999-04-07 snapshot of pthreads-win32.
1574 * tests/Makefile.am: Distribute makefile.msc. Generate it when
1577 * tests/makefile.msc.in: New file.
1579 * tests/node-test.c: Include <stdlib.h> for exit().
1581 Thu Apr 29 02:16:36 1999 Tim Janik <timj@gtk.org>
1583 * gstrfuncs.c: minor code cleanups.
1585 Tue Apr 27 13:11:29 1999 Owen Taylor <otaylor@redhat.com>
1587 * gmain.c (g_main_poll): Mask out ERR HUP and NVAL from
1588 the events field so we don't give IRIX fits.
1590 Tue Apr 20 08:42:22 1999 Tim Janik <timj@gtk.org>
1592 * gscanner.c (g_scanner_unexp_token): behave conservative with
1593 G_TOKEN_IDENTIFIER_NULL and always assume scanner->value.v_string
1594 to be "null" in that case.
1596 1999-04-20 Havoc Pennington <hp@pobox.com>
1598 * gutils.c (g_vsnprintf): When using the vsnprintf()
1599 implementation, '\0'-terminate the resulting string
1600 and return its length rather than -1.
1602 Mon Apr 19 13:42:21 1999 Owen Taylor <otaylor@redhat.com>
1604 * gmain.c (g_main_iterate): Added missing
1605 #ifdef G_THREADS_ENABLED. (I never liked G_THREADS_ENABLED in
1608 1999-04-18 Havoc Pennington <hp@pobox.com>
1610 * gutils.c (g_snprintf): When using the vsnprintf()
1611 implementation, '\0'-terminate the resulting string
1612 and return its length rather than -1.
1614 Fri Apr 16 06:52:07 1999 Tim Janik <timj@gtk.org>
1616 * gscanner.c (g_scanner_unexp_token): feature G_TOKEN_EOF as a valid
1617 expected token as well, so we get "- expected end of file" instead of
1618 "- expected (unknown) token <0>".
1620 Tue Apr 13 16:16:14 CDT 1999 Shawn T. Amundson <amundson@gtk.org>
1622 * Released GLib 1.2.2
1624 1999-04-12 Elliot Lee <sopwith@mh69.mh.cuc.edu>
1626 * g_strchug(): s/strcpy/memmove/
1628 1999-04-12 Sebastian Wilhelmi <wilhelmi@ira.uka.de>
1630 * configure.in: Adjusted the test for an unimplemented
1631 getpwuid_r. Info from Michael Pruett <mikep@ugcs.caltech.edu>.
1633 Sun Apr 11 15:07:34 1999 Tim Janik <timj@gtk.org>
1635 * configure.in: bumped versin number to GLib 1.2.2, interface 2,
1640 Sun Apr 11 14:37:06 1999 Tim Janik <timj@gtk.org>
1642 * gstrfuncs.c (g_strcasecmp): always check for s1, s2 != NULL.
1644 Sat Apr 10 19:30:50 1999 Tim Janik <timj@gtk.org>
1646 * glib.h: removed braces around inline strings for the G_GNUC_FUNCTION
1647 and G_GNUC_PRETTY_FUNCTION macros, so the macros can be used for compile
1648 time string concatenation.
1650 Thu Apr 8 19:53:19 1999 Owen Taylor <otaylor@redhat.com>
1652 * gmain.c (g_main_iterate): Check for two threads
1653 calling g_main_iterate at once.
1655 * gmain.c: If the set of poll file descriptors changes
1656 during a call to poll(), abort that call, and start
1657 a new poll. My test program still segfaults
1658 obscurely on glibc 2.0 (in read()!!!), but now it works on
1659 glibc 2.1, so I'll blame something else for the other segfault.
1661 1999-03-31 Sebastian Wilhelmi <wilhelmi@ira.uka.de>
1663 * configure.in: Fixed slight bug, that made configure hang on some
1664 systems. Please do not merge this into 1.3 branch. It's taken care
1665 of differently there. Info from J. Rhett Aultman
1666 <cuplan@alley.gator.net>
1668 Wed Mar 24 21:23:47 CST 1999 Shawn T. Amundson <amundson@gtk.org>
1670 * Released GLib 1.2.1
1675 sanity_check: updated
1677 * glibconfig.h.win32.in:
1679 docs/glib-config.1.in:
1680 docs/Makefile.am: Added files used to generate new files.
1682 * glibconfig.h.win32:
1683 docs/glib-config.1: Removed, now generated.
1685 * configure.in: Added to output now-generated files.
1687 Tue Mar 23 13:43:39 PST 1999 Manish Singh <yosh@gimp.org>
1689 * giounix.c: add user_data param to check and prepare functions
1691 Mon Mar 22 03:54:43 1999 Tim Janik <timj@gtk.org>
1694 * gmain.c: add user_data to the GSource ->check and ->prepare functions,
1695 so it can be used to e.g. pass a GPollFd.
1696 (g_main_poll): only add poll records with an events mask != 0 to the
1697 fd_array. don't even bother calling poll_func() if fds=timeout=0.
1698 added debugging printouts around poll_func() invokation that can be
1699 enabled with #define G_MAIN_POLL_DEBUG.
1701 Fri Mar 19 16:29:50 PST 1999 Manish Singh <yosh@gimp.org>
1707 * ltmain.sh: upgrade to libtool 1.2f
1709 * autogen.sh: libtool is not required to autogen glib
1711 * acconfig.h: remove WITH_SYMBOL_UNDERSCORE (not explictly
1714 1999-03-18 Sebastian Wilhelmi <wilhelmi@ira.uka.de>
1716 * gmem.c: Fixed another stupid fault of mine: Did
1717 s/g_static_/g_private_/g
1719 Wed Mar 17 03:17:42 1999 Tim Janik <timj@gtk.org>
1721 * configure.in bumped versin number to GLib 1.2.1, interface 1,
1726 * glib.h: added GLIB_CHECK_VERSION() macro similar to
1727 GTK_CHECK_VERSION().
1729 Sun Mar 14 17:50:35 1999 Tim Janik <timj@gtk.org>
1731 * gmem.c (g_mem_chunk_*): changed a bunch of g_assert() statements
1732 to g_return_if_fail().
1734 (g_mem_chunk_print):
1735 (g_mem_chunk_info): removed some extraneous "\n"s at the end of the log
1738 * gtimer.c (g_timer_*): changed a bunch of g_assert() statements
1739 to g_return_if_fail().
1740 * grel.c (g_*): changed a bunch of g_assert() statements to
1741 g_return_if_fail() and added some extra ones to check relation != NULL.
1743 1999-03-12 Sebastian Wilhelmi <wilhelmi@ira.uka.de>
1745 * configure.in: Also accept _Pctime_r instead of ctime_r, while
1746 seraching for the right `_REENTRANT' flag. This is for Digital
1747 UNIX 4.0d. Thanks to Sascha Brawer <sb@adasys.ch>.
1749 Tue Mar 9 23:25:50 1999 Tim Janik <timj@gtk.org>
1751 * configure.in: check for working realloc (NULL,).
1752 * gmem.c (g_realloc): use malloc() for initial allocation on systems
1753 where realloc(NULL,) will not work (this is the case on SunOS, reported
1756 Mon Mar 8 07:42:08 1999 Tim Janik <timj@gtk.org>
1758 * ghook.c (g_hook_unref): when !hook_list->is_setup, wrap the
1759 flag around the call to g_hook_free() to avoid spurious
1760 warnings (happens during destruction phase).
1762 1999-03-03 Sebastian Wilhelmi <wilhelmi@ira.uka.de>
1764 * glibconfig.h.win32, config.h.win32: Moved G_THREADS_IMPL_POSIX
1765 from config.h.win32 to glibconfig.h.win32
1767 * acconfig.h, configure.in, config.h.win32: Added test for DCE
1768 versions of mutex_trylock and cond_timedwait. The win32 versions
1769 are posix, aren't they?
1771 1999-03-02 Sebastian Wilhelmi <wilhelmi@ira.uka.de>
1773 * gmem.c: Fixed a stupid cut'n'paste error of mine. Thanks to
1774 Friedrich Dominicus <Friedrich.Dominicus@inka.de>
1776 1999-03-01 Sebastian Wilhelmi <wilhelmi@ira.uka.de>
1778 * gutils.c (g_get_any_init): Fixed yet another bloody
1779 implementation of getpwuid_r on AIX. Thanks to Olaf Dietsche
1780 <olaf.dietsche+list.gtk@netcologne.de>. I would like a configure
1781 test better than that, but have no idea, how to do that easily.
1783 Sun Feb 21 22:11:51 CST 1999 Shawn T. Amundson <amundson@gtk.org>
1785 * Released GLib 1.2.0
1789 Wed Feb 24 00:08:42 CST 1999 Shawn T. Amundson <amundson@gtk.org>
1791 * *.[ch]: inserted additional note to look for ChangeLog and
1792 AUTHORS file for a log of modifications.
1794 Sun Feb 21 14:01:00 1999 Dr Mike <drmike@redhat.com>
1796 * Made specfile generated, tweaked slightly
1798 Sat May 8 06:00:17 CDT 1999 Shawn T. Amundson <amundson@gtk.org>
1802 gthread/Makefile.am: Better testing reveals better
1803 methods. Fixes for BeOS.
1805 Sat May 8 01:52:29 CDT 1999 Shawn T. Amundson <amundson@gtk.org>
1816 gmodule/gmoduleconf.h.in
1817 gmodule/gmodule-beos.c
1818 gthread/Makefile.am: Port to BeOS by myself and Richard Offer.
1820 1999-05-06 Tor Lillqvist <tml@iki.fi>
1822 * makefile.msc.in makefile.cygwin.in glibconfig.h.win32.in
1823 config.h.win32.in tests/makefile.msc.in tests/makefile.cygwin.in:
1824 New files, used to generate corresponding non-.in files when
1825 making a dist. This is just so the version numbers will be kept in
1826 synch automatically.
1828 * configure.in: Also substitute @GLIB_MAJOR_VERSION@,
1829 @GLIB_MINOR_VERSION@, and @GLIB_INTERFACE_AGE@.
1831 * Makefile.am tests/Makefile.am: Also distribute makefile.cygwin.
1833 * gerror.c (g_on_error_query): On Win32, put up a MessageBox and
1836 * glib.def: Add a couple of functions.
1838 Sat May 1 10:26:20 PDT 1999 Manish Singh <yosh@gimp.org>
1844 * ltmain.sh: upgrade to libtool 1.3
1846 1999-04-25 Tor Lillqvist <tml@iki.fi>
1848 * README.win32: More editing.
1849 * build-dll: Use gcc, not ld to link.
1850 * glib.h: On native Win32 use _unlink().
1851 * gscanner.c: Use corrent NATIVE_WIN32 feature test macro,
1853 * gstring.c: Include <io.h> on Win32 for _read prototype.
1854 * gutils.c: Remove old IO channel code (was in #if 0).
1855 * makefile.cygwin: Don't need to link with kernel32 and msvcrt
1856 explicitly, they are included anyway.
1858 1999-04-24 Tor Lillqvist <tml@iki.fi>
1860 Support added for building using a GNU toolchain on Win32,
1861 i.e. gcc -mno-cygwin on cygwin (a.k.a. mingw32, using egcs-1.1.2).
1863 * README.win32: Updated.
1864 * build-dll makefile.cygwin tests/makefile.cygwin: New files.
1865 * glib.h glib.def glibconfig.h.win32: Slight updates.
1866 * gmain.c: No need to include <fcntl.h> and <io.h> on Win32.
1867 * gmain.c gutils.c testglib.c tests/string-test.c: Test for
1868 NATIVE_WIN32, not _MSC_VER.
1869 * gmutex.c: Must declare g_thread_functions_for_glib_use as
1870 exported (using the GUTILS_C_VAR macro).
1871 * gutils.c gmodule/libgplugin_[ab].c: LibMain not needed.
1872 * gmodule/gmoduleconf.h.win32: Need underscore with gcc.
1873 * gthread/gthread.c: With gcc on Win32, must use memcpy to assign
1874 value of g_thread_functions_for_glib_use (?).
1875 * makefile.msc tests/makefile.msc: Cosmetics.
1877 Fri Apr 23 14:29:25 BST 1999 Tony Gale <gale@gtk.org>
1879 * glib.h: Fix typo in g_string_ncasecmp macro (by me).
1880 Add b_string_strncasecmp macro.
1882 1999-04-22 Sebastian Wilhelmi <wilhelmi@ira.uka.de>
1884 * gutils.c (g_get_any_init): use sysconf (_SC_GETPW_R_SIZE_MAX) as
1885 the new initinal bufsize for getpwuid_r on systems, that support
1886 this. Hint from Holger Duerer <H.Duerer@zait.uni-bremen.de>.
1888 Sat Apr 17 20:55:13 BST 1999 Tony Gale <gale@gtk.org>
1890 * glib.h, gstring.c: Add new g_string functions for reading
1891 from file/socket descriptors, and tokenising strings.
1893 Added various g_string macros.
1895 Tue Apr 13 23:28:32 1999 Tor Lillqvist <tml@iki.fi>
1897 * README.win32: Mention the tests directory.
1899 * glib.def: Add the functions from grand.c.
1901 * glibconfig.h.win32: Add unsigned max values, and the format
1904 * makefile.msc: Add grand.
1906 * tests/{date-test,node-test}.c: Include <stdlib.h> for exit().
1908 * tests/makefile.msc: New file.
1910 1999-04-12 Sebastian Wilhelmi <wilhelmi@ira.uka.de>
1912 * glib.h: Moved struct declaration up. Style fixes.
1914 * grand.c: Style fixes. Only try to open /dev/random once.
1916 * tests/rand-test.c (main): New tests; Slight bug fix.
1918 1999-04-09 Sebastian Wilhelmi <wilhelmi@ira.uka.de>
1920 * grand.c, tests/rand-test.c: New files to implement the Mersenne
1921 Twister Pseudo Random Number Generator.
1923 * glib.h, AUTHORS, Makefile.am, tests/Makefile.am: Changed
1926 Thu Apr 8 21:12:30 CDT 1999 Shawn T. Amundson <amundson@gtk.org>
1928 * Released GLib 1.3.0
1930 1999-03-30 Sebastian Wilhelmi <wilhelmi@ira.uka.de>
1932 * configure.in: Added a check for the right format to printf and
1933 scanf long longs. It is %qi instead of %lli on FreeBSD for
1936 1999-03-28 Raja R Harinath <harinath@cs.umn.edu>
1938 * Makefile.am (glibconfig.h): Make sure `glibconfig.h' exists
1939 after the rule is fired.
1940 (install-exec-local): Install glibconfig.h only if the contents
1941 are different from the currently installed glibconfig.h.
1943 1999-03-26 Raja R Harinath <harinath@cs.umn.edu>
1945 * Makefile.am (configexecincludedir): Rename from
1946 configincludedir so that glibconfig.h will be installed
1947 as part of `make install-exec'.
1949 Thu Mar 25 22:45:47 1999 Tor Lillqvist <tml@iki.fi>
1951 * config.h.win32: Update version numbers.
1953 * glibconfig.h.win32: Update version numbers and pthreads-win32-
1954 related magic values.
1956 * README.win32: Some improvements.
1958 * makefile.msc: Add gqueue and gstack. Correct version number.
1960 Fri Mar 19 16:29:50 PST 1999 Manish Singh <yosh@gimp.org>
1966 * ltmain.sh: upgrade to libtool 1.2f
1968 * autogen.sh: libtool is not required to autogen glib
1970 * acconfig.h: remove WITH_SYMBOL_UNDERSCORE (not explictly
1973 1999-03-18 Jeff Garzik <jgarzik@pobox.com>
1975 * glib.def: Add new g_list, g_stack, g_queue functions.
1977 1999-03-18 Sebastian Wilhelmi <wilhelmi@ira.uka.de>
1979 * configure.in: added new AC_SUBST(GTHREAD_COMPILE_IMPL_DEFINES)
1980 to hold various defines to get the right thread implementation on
1981 different platforms. Also look in -ldce for pthread_create. Should
1982 make it work on HP-UX 10.x. Information from "D. Emilio Grimaldo
1983 Tunon" <emilio_tunon@nl.compuware.com>.
1985 1999-03-17 Jeff Garzik <jgarzik@pobox.com>
1987 * gstack.c, gqueue.c:
1988 Add copyright, clean up code a bit.
1990 1999-03-17 Sebastian Wilhelmi <wilhelmi@ira.uka.de>
1992 * configure.in: Added missing values for G_MAXU(SHORT|INT|LONG) on
1993 platforms with only /usr/include/values.h.
1995 * acconfig.h: Removed unnecessary macros.
1997 * glibconfig.h.win32, config.h.win32: Moved G_THREADS_IMPL_POSIX
1998 from config.h.win32 to glibconfig.h.win32. Taken from glib 1.2
2001 * configure.in: Also accept _Pctime_r instead of ctime_r, while
2002 seraching for the right `_REENTRANT' flag. This is for Digital
2003 UNIX 4.0d. Taken from glib 1.2 branch.
2005 Wed Mar 17 03:14:56 1999 Tim Janik <timj@gtk.org>
2007 * glib.h: added GLIB_CHECK_VERSION() macro similar to
2008 GTK_CHECK_VERSION().
2010 Wed Mar 17 01:46:28 1999 Tim Janik <timj@gtk.org>
2012 * merges from glib-1-2:
2014 Sun Mar 14 17:50:35 1999 Tim Janik <timj@gtk.org>
2016 * gmem.c (g_mem_chunk_*): changed a bunch of g_assert() statements
2017 to g_return_if_fail().
2019 (g_mem_chunk_print):
2020 (g_mem_chunk_info): removed some extraneous "\n"s at the end of the log
2023 * gtimer.c (g_timer_*): changed a bunch of g_assert() statements
2024 to g_return_if_fail().
2025 * grel.c (g_*): changed a bunch of g_assert() statements to
2026 g_return_if_fail() and added some extra ones to check relation != NULL.
2028 Tue Mar 9 23:25:50 1999 Tim Janik <timj@gtk.org>
2030 * configure.in: check for working realloc (NULL,).
2031 * gmem.c (g_realloc): use malloc() for initial allocation on systems
2032 where realloc(NULL,) will not work (this is the case on SunOS, reported
2035 Mon Mar 8 07:42:08 1999 Tim Janik <timj@gtk.org>
2037 * ghook.c (g_hook_unref): when !hook_list->is_setup, wrap the
2038 flag around the call to g_hook_free() to avoid spurious
2039 warnings (happens during destruction phase).
2041 1999-03-02 Sebastian Wilhelmi <wilhelmi@ira.uka.de>
2043 * gmem.c: Fixed a stupid cut'n'paste error of mine. Thanks to
2044 Friedrich Dominicus <Friedrich.Dominicus@inka.de>
2046 1999-03-16 Timur Bakeyev <mc@bat.ru>
2048 * configure.in: Fix problem with pthread_create in libc, as running
2049 "gcc test.c -l " is not legal.
2051 1999-03-16 Sebastian Wilhelmi <wilhelmi@ira.uka.de>
2053 * tests/type-test.c: Added a test for the
2054 G_(U)?INT(16|32|64)_FORMAT and G_(MIN|MAX|MAXU)(SHORT|INT|LONG)
2057 * configure.in: Removed G_(U)?INT8_FORMAT again, as it can't be
2060 * configure.in: Added the macros G_MAXU(SHORT|INT|LONG). I do not
2061 know how to handle these on platforms with /usr/include/values.h,
2062 but without /usr/include/limits.h. Please someone add this.
2065 1999-03-15 Sebastian Wilhelmi <wilhelmi@ira.uka.de>
2067 * configure.in: Added the macros G_(U)?INT(8|16|32|64)_FORMAT to
2068 use for printf and (much more important) scanf format strings for
2069 the corresponding GLib types.
2071 * glib.h Added G_(U)?(SHORT|INT|LONG)_FORMAT for consistency. It
2072 however makes no sense to also provide G_(FLOAT|DOUBLE)_FORMAT, as
2073 they are different for printf (f for both) and scanf (f for float,
2074 lf for double). Defining G_INT_FORMAT makes sense however, as we
2075 might want to define gint to something different than int someday
2076 in the future. Idea from Sascha Brawer <sb@adasys.ch>.
2078 1999-03-14 Jeff Garzik <jgarzik@pobox.com>
2081 Commented out debugging output.
2083 * tests/Makefile.am, tests/date-test.c:
2084 Added test of the GDate module, based closely on testgdate.c.
2086 * tests/Makefile.am:
2087 Bugfix - compile tests with @GLIB_DEBUG_FLAGS@.
2089 1999-03-14 Raja R Harinath <harinath@cs.umn.edu>
2091 * configure.in (glibconfig.h): Remove widechar tests and defines.
2092 (fd_set): Change the grep for `fd_mask' to search for `fd_set'.
2093 * gerror.c (fd_mask): Remove conditional typedef. It is not used
2094 elsewhere in the file.
2095 * gmain.c (fd_mask): Likewise.
2097 1999-03-12 Sebastian Wilhelmi <wilhelmi@ira.uka.de>
2099 * configure.in: Test for posix threads first, then for dce threads.
2101 1999-03-11 Sebastian Wilhelmi <wilhelmi@ira.uka.de>
2103 * configure.in: Revamped the thread configure stuff. Now dce
2104 threads (old posix draft) are recogniced. This is necessary,
2105 because dce threads are in fact working quite differently from
2106 posix threads. Also changed the conditions for checking for MT
2107 safe functions a bit, because G_THREADS_IMPL_NONE still have to
2108 compile thread safe. NOTE: Please do not commit my change to
2109 glib-1-2/{acconfig.h,configure.in,config.h.win32} from 1999-03-03,
2110 as the current change will take care of that too.
2112 Tue Mar 9 14:37:32 1999 Jeff Garzik <jgarzik@pobox.com>
2114 * Makefile.am, glib.h, gstack.c, gqueue.c,
2115 tests/Makefile.am, tests/queue-test.c, tests/stack-test.c:
2116 Added stack, queue ADTs and related tests.
2119 New g_list_delete() function.
2121 Sat Mar 6 11:03:08 1999 Asbjorn Pettersen <ape@lrdpf.spacetec.no>
2123 * gutils.c (g_get_any_init): add OS/2 changes.
2124 change '\\' in HOME to '/'.
2126 1999-03-03 Josh MacDonald <jmacd@spin.dsl.pacbell.net>
2128 * glib.def: g_spaced_primes_closest was omitted here, so I
2129 couldn't build Xdelta on Windows.
2131 1999-03-01 Sebastian Wilhelmi <wilhelmi@ira.uka.de>
2133 * gutils.c (g_get_any_init): Fixed yet another bloody
2134 implementation of getpwuid_r on AIX. Thanks to Olaf Dietsche
2135 <olaf.dietsche+list.gtk@netcologne.de>. I would like a configure
2136 test better than that, but have no idea, how to do that easily.
2138 Sat Feb 27 01:18:47 1999 Tim Janik <timj@gtk.org>
2140 * ChangeLog: moved old ChangeLog to ChangeLog.pre-1-2, and started
2143 * configure.in: set glib version to 1.3.0.