1 Wed Jan 20 15:05:25 PST 1999 Manish Singh <yosh@gimp.org>
5 * Makefile.am: since we require automake 1.4 now, ditch the build
6 top-level first and just use SUBDIRS
8 * gmodule/Makefile.am: use noinst_LTLIBRARIES instead of overriding
11 Wed Jan 20 16:16:01 EST 1999 Jeff Garzik <jgarzik@pobox.com>
13 * gutils.c (g_get_any_init):
14 Revert previous patch, move error value set line down to proper
15 place. Caught by Elliot Lee <sopwith@redhat.com>.
17 Wed Jan 20 20:48:58 GMT 1999 Adam D. Moss <adam@gimp.org>
19 * gutils.c (g_get_any_init): All gtk apps were broken
20 here on my setup. I changed the error to a warning to at
21 least get things limping.
23 1999-01-20 Sebastian Wilhelmi <wilhelmi@ira.uka.de>
25 * gutils.c (g_get_any_init): Use getpwuid_r with the right
26 signature, if available.
28 * configure.in, acconfig.h: Test for existance of getpwuid_r and
31 Tue Jan 19 20:52:43 1999 Tor Lillqvist <tml@iki.fi>
33 * Makefile.am (EXTRA_DIST): Add new Win32 files.
35 * README.win32: Describe the conditional compilation macros.
37 * makefile.lcc: Removed.
39 1999-01-19 Sebastian Wilhelmi <wilhelmi@ira.uka.de>
41 * configure.in: Moved test for localtime_r and rand_r to a place
42 after the determination of the G_THREAD_LIBS and use them for
43 finding those functions, necessary on systems with a different C
44 library libc_r for threaded progs, like FreeBSD 2.2.x. Info from
45 Timur I. Bakeyev <timur@comtat.ru>.
47 Tue Jan 19 00:44:24 1999 Josh MacDonald <jmacd@axis.hip.berkeley.edu>
49 * gstrfuncs.c (g_strsignal): The "extern char* sys_siglist"
50 declaration breaks systems with different declarations, like
51 mine (FreeBSD). So, I added a configuration variable to decide
52 whether the declaration is neccesary. Change also appears on
53 line 275 of configure.in.
55 Sun Jan 17 17:13:54 CST 1999 Shawn T. Amundson <amundson@gtk.org>
57 * Released GLib 1.1.13
61 docs/glib-config: ver=1.1.13
63 Mon Jan 18 00:02:46 1999 Tim Janik <timj@gtk.org>
65 * NEWS: updates for 1.1.13.
67 Sun Jan 17 17:15:59 1999 Tim Janik <timj@gtk.org>
69 * gtree.c (g_tree_traverse): removed assertion about rtree->root being
70 NULL, so this function can be applied to unpopulated trees as well, fix
71 provided by Simon Kagedal <sika8225@csd.uu.se>.
73 Sun Jan 17 11:17:19 1999 Christopher Blizzard <blizzard@appliedtheory.com>
75 * glib.spec: add packager address
77 Sun Jan 17 11:10:05 1999 Christopher Blizzard <blizzard@appliedtheory.com>
79 * glib.spec: increment version number to 1.1.13
81 Sun Jan 17 16:14:03 1999 Tim Janik <timj@gtk.org>
83 * glib.h (g_error): if !defined (G_LOG_DOMAIN) define G_LOG_DOMAIN as
84 ((gchar*) 0) instead of ((void*) 0) to silence egcs C++ mode, reported
85 by Seth Alves <alves@hungry.com>.
87 Sun Jan 17 14:13:52 1999 Tim Janik <timj@gtk.org>
89 * glib.h: added a define G_HOOK_DEFERRED_DESTROY, to substitute a noop
90 GHookList.hook_destroy function.
92 * ghook.c (g_hook_destroy_link): don't really call hook_destroy if it
93 is G_HOOK_DEFERRED_DESTROY. for the case where we invoke hook->destroy()
94 we now clean up the hook completely afterwards, i.e. data, func and
95 destroy are immediately set to NULL and hook_free can't play with that
98 * gmain.c (g_source_add): set hook_destroy to G_HOOK_DEFERRED_DESTROY,
99 instead of using an ugly _noop() hack, this is to avoid an uneccessary
100 function invokation. set hook_free to g_source_destroy_func, this way
101 we always invoke the destroy notifiers for user_data and source_data
102 after execution of dispatch(). thus, g_source_destroy_func() will always
103 be called within the main_loop lock (this wasn't really assured
104 before), and can release and reaquire the look around destroy notifier
107 Sun Jan 17 11:01:40 1999 Tor Lillqvist <tml@iki.fi>
109 * config.h.win32 glibconfig.h.win32: Increment version number here
112 Sun Jan 17 05:12:17 1999 Tim Janik <timj@gtk.org>
114 * configure.in: incremented version number to 1.1.13, bin age 0,
117 * glib.h: added hook_destroy member to struct _GHookList.
119 * ghook.c (g_hook_destroy_link): if the hook_list defines a hook_destroy
120 function, use that to marshal hook destruction notifiers.
122 * gmain.c: removed g_source_free_func and added g_source_destroy_func,
123 which marshalls the user_data and source_data destructors of sources
124 outside of the main_loop lock.
125 removed GIdleData struct since its sole member callback can be passed
126 as source_data directly.
127 added a gboolean return value to all g_source_remove* functions,
128 indicating whether the source could be removed, because these functions
129 don't issue warnings upon unseccessful removals themselves. this way at
130 least the caller gets a chance to warn upon failing removals.
131 (g_main_iterate): set in_check_or_prepare around calls to check() or
133 (g_main_pending): simply return FALSE if called from within check() or
135 (g_main_iteration): issue a warning if called from within check() or
136 prepare() and bail out with FALSE.
137 (g_main_run): likewise.
138 (g_source_remove_by_funcs_user_data): new function to remove sources by
139 user data and function table.
140 (g_idle_remove_by_data): new function to really remove idles only, since
141 g_source_remove_by_user_data would remove timeouts or other sources as
144 1999-01-16 Tor Lillqvist <tml@iki.fi>
146 * Merge in current Win32 version:
148 * README.win32: More text.
150 * config.h.win32 glibconfig.h.win32: Update to match the
151 corresponding generated files on Unix.
153 * makefile.msc: Update with new source files, and gthread
154 library. Use the compiler flag -MD instead of using -D_DLL and
155 /nodefaultlib:libc msvcrt.lib in the link phase.
157 * glib.def: Update to include new functions, drop removed ones.
159 * glib.h: Add comments about main loop and polling on Win32. (In
160 general, it's only for the GIMP's use.) Add Win32 IO Channel
161 functions. Remove the obsoleted old IO Channel stuff (which was
164 * giowin32.c: New file.
166 * gmain.c: Include config.h, conditionalize <sys/time.h>
167 inclusion. Add g_poll implementation for Win32 (only for the
168 GIMP's needs for now, it's hard or even impossible to be as clean
169 and generic as on Unix). Implement g_get_current_time on Win32. If
170 threads aren't supported, don't try to wake up main thread's
171 loop. On Win32, use a semaphore and not a pipe to wake up the main
174 * gmessages.c: On Win32, allocate a console window if the standard
175 output handle is invalid before writing to stdout, and reopen stdout
176 to that console window.
178 * giochannel.c: Conditionalize unistd.h inclusion. Some indentation
181 * gstrfuncs.c: Include <signal.h>.
183 * gutils.c: On Win32, also check the HOMEDRIVE and HOMEPATH
184 environment variables.
186 Sun Jan 17 01:06:38 1999 Timur Bakeyev <mc@bat.ru>
188 * configure.in: Add support for FreeBSD 2.2.x macro _THREAD_SAFE.
190 Wed Jan 13 11:27:24 PST 1999 Manish Singh <yosh@gimp.org>
192 * glib.h: don't use inline for !__GNUC__ case with g_warning,
193 g_error, and g_message wrappers since some compilers throw
194 away arguments in this case
196 Tue Jan 12 21:59:14 PST 1999 Manish Singh <yosh@gimp.org>
200 * ltmain.sh: upgrade to libtool 1.2d (with fixes for irix6 and osf)
202 * testglib.c: removed unused cruft
204 Tue Jan 12 09:57:06 PST 1999 Manish Singh <yosh@gimp.org>
206 * gdate.c: #warning isn't portable, check for gcc
208 1999-01-12 Sebastian Wilhelmi <wilhelmi@ira.uka.de>
210 * gdate.c: Include config.h to get HAVE_LOCALTIME_R macro.
212 * configure.in: Moved function check back to the place, they've
215 1999-01-11 Jeff Garzik <jgarzik@pobox.com>
218 Do not use vasprintf.
220 * configure.in, glib.h, testglib.c, tests/.cvsignore,
221 tests/Makefile.am, tests/alloca-test.c:
224 1999-01-11 Sebastian Wilhelmi <wilhelmi@ira.uka.de>
226 * configure.in: Test for localtime_r only after including the
227 right MT enabling CFLAGS (i.e. -D_REENTRANT on most systems).
229 * configure.in: Find right thread system on DG/UX. Thanks to Marc
230 J. Fraioli <fraioli@dg-rtp.dg.com> for hint. Finally removed the
231 stuff for -fstack-check, that didnt work anyway.
233 * gdate.c (g_date_set_time): Emit warning, if no localtime_r
234 function is available on thread enabled systems. Define ptm only,
235 if really needed, and assert on it.
237 Sat Jan 9 15:08:44 1999 Jeff Garzik <jgarzik@pobox.com>
240 Make all aux functions static.
242 * tests/Makefile.am, tests/dirname-test.c, tests/type-test.c:
243 New tests dirname-test and type-test, from testglib.
245 Sat Jan 9 13:53:00 1999 Jeff Garzik <jgarzik@pobox.com>
248 Add checks for vasprintf, localtime_r.
250 * gdate.c (g_date_set_time):
251 Use localtime if localtime_r is not available.
253 * gstrfuncs.c (g_strdup_vprintf):
254 Use glibc vasprintf if possible; it's a bit faster than using
255 GLib routines, and makes output code a bit smaller.
258 Remove HAVE_VSNPRINTF and HAVE_VPRINTF. autoheader picks these
259 up automatically and puts them in config.h.in.
261 Thu Jan 7 15:14:08 1999 Owen Taylor <otaylor@redhat.com>
263 * gmain.c (g_source_free_func): Call the source-specific
264 free function when the hook is freed not when it
265 is destroyed; this fixes a bug where a timeout destroyed
266 from itself would access already freed data.
268 1999-01-07 Sebastian Wilhelmi <wilhelmi@ira.uka.de>
270 * gutils.c (g_get_any_init): Here we must replace getpwuid by
271 getpwuid_r, but as I do not know how for now, I just made a FIXME
274 * gdate.c (g_date_set_time): localtime --> localtime_r to make it
277 * configure.in: We do not need to check for broken solaris mutex
278 intitializer any longer. Provide a macro to show the used thread
279 implementation. Not nice, but this is needed until thread support
280 is completed here inside glib.
282 Tue Jan 5 12:23:52 PST 1999 Manish Singh <yosh@gimp.org>
284 * configure.in: ditch whitespace before GLIB_SYSDEFS #includes,
285 the # needs to be in column 1
287 * gthread/testgthread.c: cleanups
289 * tests/node-test.c: #include unistd.h for exit()
291 1999-01-04 Andrew T. Veliath <andrewtv@usa.net>
293 * glib.h (g_array_insert_val): fix macro to be shorthand of
296 Mon Jan 4 15:35:29 PST 1999 Manish Singh <yosh@gimp.org>
298 * acglib.m4: some echos interpret \n's and some don't. Deal with
301 Mon Jan 4 20:58:50 1999 Tim Janik <timj@gtk.org>
303 * gscanner.c: eliminated extraneous "register" qualifiers in variable
306 * gmain.c: #undef events and revents which may have been defined in
307 sys/poll.h for SVR3,4 compatibility on some AIX systems. fix been
308 provided by Philippe Defert <Philippe.Defert@cern.ch>.
310 Mon Jan 4 14:38:11 1999 Tim Janik <timj@gtk.org>
312 * glib.h: adjusted the g_strdup_a, g_strndup_a and g_strconcat3_a macros
313 so their arguments get only evaluated once. changed g_strconcat3_a to
314 have the same semantics as g_strconcat, i.e. if a certain argument is
315 NULL, the rest of the parameter list is skipped.
317 * tests/alloca-test.c: removed #if/#endif G_HAVE_ALLOCA, we need to be
318 able to compile on all systems. added test for g_strconcat() semantics.
320 * we need a alloca() replacement for !G_HAVE_ALLOCA rsn!
322 Mon Jan 4 02:58:13 CST 1999 Shawn T. Amundson <amundson@gtk.org>
324 * Released GLib 1.1.12
331 docs/glib-config.1: version=1.1.12
333 * tests/Makefile.am: fixed so it works properly with
336 Sun Jan 3 01:38:14 EST 1999 Jeff Garzik <jgarzik@pobox.com>
339 add array-test.c, tree-test.c
341 * tests/array-test.c, tests/tree-test.c:
342 New module, tests array family
344 * tests/hash-test.c, tests/list-test.c, tests/slist-test.c,
345 tests/string-test.c, tests/node-test.c:
346 Clean out cruft left over from testglib.
348 Sat Jan 2 22:42:25 EST 1999 Jeff Garzik <jgarzik@pobox.com>
350 * Makefile.am, configure.in, tests/*:
351 Added 'make check' tests, based on testglib code.
353 Sat Jan 2 19:52:45 EST 1999 Jeff Garzik <jgarzik@pobox.com>
355 * glib.h, testglib.c:
356 Added g_alloca, g_new_a, g_new0_a macros.
358 Sat Jan 2 16:45:44 EST 1999 Jeff Garzik <jgarzik@pobox.com>
360 * testglib.c: Added g_strdup_printf check.
362 Fri Jan 1 21:58:40 EST 1999 Jeff Garzik <jgarzik@pobox.com>
365 (g_strdup_a, g_strndup_a): Handle NULL strings like g_strdup.
366 s/g_strconcat_a/g_strconcat3_a/ to reflect fixed number of args
369 Added g_strdup, g_strconcat checks.
370 Added str==NULL checks for alloca string macros.
371 s/g_strconcat_a/g_strconcat3_a/
373 Fri Jan 1 18:30:41 PST 1999 Manish Singh <yosh@gimp.org>
375 * testglib.c: made the alloca tests follow the testglib style
377 * Makefile.am: minor cleanups, mostly cosmetic
379 Fri Jan 1 20:43:19 EST 1999 Jeff Garzik <jgarzik@pobox.com>
381 * glib.h: added g_strndup_a macro
384 Added tests for new alloca-based string routines.
385 Reformatted a couple strings.
387 Sat Jan 2 02:20:59 1999 Tim Janik <timj@gtk.org>
390 (g_hook_list_invoke):
391 (g_hook_list_invoke_check):
392 (g_hook_list_marshal_check):
393 (g_hook_list_marshal): avoid unneccessary extra hook referencing (the
394 explicit hook referencing became unneccessarry with my changes from
395 Mon Dec 21 21:48:29 1998).
397 * gmain.c (g_main_iterate): fixed reference counting leaks with
398 premature loop aborts.
400 Fri Jan 1 22:47:44 1999 Tim Janik <timj@gtk.org>
402 * gscanner.c (g_scanner_unexp_token): handle G_TOKEN_IDENTIFIER_NULL
403 as G_TOKEN_IDENTIFIER.
405 Fri Jan 1 17:09:19 EST 1999 Jeff Garzik <jgarzik@pobox.com>
407 * configure.in, glib.h:
408 Added two new alloca-based function macros, g_strdup_a and
409 g_strconcat_a. These are stack-based and much faster than
410 their g_malloc-based counterparts. Kudos to Ulrich Drepper
411 for help on this one.
413 Wed Dec 30 18:24:57 CST 1998 Shawn T. Amundson <amundson@gtk.org>
415 * Released GLib 1.1.11
421 docs/glib-config.1: version=1.1.11
423 1998-12-30 Raja R Harinath <harinath@cs.umn.edu>
425 * glib.h (G[U]INT64_FROM_{LE,BE}):
426 Define to G[U]INT64_TO_{LE,BE}, not G[U]INT32_TO_{LE,BE}.
428 Fri Dec 25 19:56:33 PST 1998 Manish Singh <yosh@gimp.org>
431 * configure.in: provide defaults for POLL sysdefs, simple enums
432 don't work with bitwise logic. Reported by Daniel Skarda
433 <0rfelyus@atrey.karlin.mff.cuni.cz>
435 Wed Dec 23 00:43:25 CST 1998 Shawn T. Amundson <amundson@gtk.org>
441 docs/glib-config.1: version=1.1.10
443 Wed Dec 23 04:18:11 1998 George Lebl <jirka@5z.com>
445 * gmain.c: (g_get_current_time) don't cast to timeval since
446 timeval is for some reason not always a struct of longs, weird
448 Tue Dec 22 10:32:11 1998 Tim Janik <timj@gtk.org>
450 * ghook.c (g_hook_first_valid): fixed buglet that could cause bogus
453 Mon Dec 21 21:48:29 1998 Tim Janik <timj@gtk.org>
456 * gmain.c: there was a reference count race for hooks during invocation
457 loops. since all (known) hook loop implementations, do currently start
458 out with g_hook_first_valid() and iterate with g_hook_next_valid(),
459 g_hook_first_valid() will now return a referenced hook, and
460 g_hook_next_valid() will "eat" that, and eventually transfer it to
461 the next hook. <sigh> unfortunately this requires g_hook_next_valid()
462 to take the hook_list as additional argument.
464 * gmain.c (g_main_iterate): adjusted callers of g_hook_next_valid().
466 Mon Dec 21 03:48:04 1998 Tim Janik <timj@gtk.org>
468 * gmain.c (g_main_iterate): default initialize source_timeout with -1
469 so we have a sane timeout value if (*prpare) doesn't set it.
471 Sat Dec 19 16:56:02 1998 Owen Taylor <otaylor@redhat.com>
473 * glib.h gmain.c (G_PRIORITY_LOW): Add #defines defining
476 Sat Dec 19 16:56:02 1998 Owen Taylor <otaylor@redhat.com>
478 * gmain.c (g_main_poll): Allocate space for pollfd's
479 _after_ adding poll wake-up-pipe record.
481 * gmain.c (g_main_add_poll): Changed name
482 of internal function g_main_add_poll_unlocked()
483 back from the non-sensical g_main_add_unlocking_poll().
485 Sat Dec 19 06:25:55 1998 Tim Janik <timj@gtk.org>
487 * glib.m4: fixed a minor bug that would let configures bail out
488 if the MODULES argument contained newlines.
490 * acglib.m4: new file to be included by configure.in. it holds
491 special GLIB_ autoconfiguration macros, eventually some of the
492 easier ones should be moved into glib.m4, e.g. GLIB_IF_VAR_EQ,
493 GLIB_STR_CONTAINS or GLIB_ADD_TO_VAR.
495 Sat Dec 19 04:27:17 1998 Tim Janik <timj@gtk.org>
497 * fixed up sys/poll.h and sys/types.h inclusions.
499 Sat Dec 19 03:10:50 1998 Tim Janik <timj@gtk.org>
501 * fixed up gthread includes, cleaned up glibconfig.h a little bit.
503 Fri Dec 18 12:51:39 1998 Owen Taylor <otaylor@redhat.com>
505 * gmain.c: Fix errors in computation of timeout
506 expiration times > 1sec.
508 1998-12-18 Sebastian Wilhelmi <wilhelmi@ira.uka.de>
510 * configure.in (have_threads): Changed the last pthread_cond_init
511 to pthread_attr_init.
513 Fri Dec 18 00:03:17 1998 Tim Janik <timj@gtk.org>
517 (g_main_is_running): new function to check whether a main loop has been
519 (g_main_new): added a gboolean argument to determine whether the loop
520 should be considered initially running or not. however, g_main_run ()
521 will still reset the main loops running state to TRUE upon initial
525 (g_main_iterate): documented this function's purpose in 5 steps.
526 for step 2), flag sources as G_SOURCE_READY even if !dispatch and
527 check G_SOURCE_READY prior to (*prepare), so we don't call (*prepare)
528 on them multiple times.
530 Thu Dec 17 23:43:47 1998 Tim Janik <timj@gtk.org>
532 * gmain.c (g_main_add_poll): reordered arguments, so GPollFD* comes
533 first, <sigh> (sorry Snorfle, i should have let you know in the first
535 (g_main_dispatch): stack G_HOOK_FLAG_IN_CALL flags. call source's
536 destructor when destroying a source.
538 1998-12-17 Sebastian Wilhelmi <wilhelmi@ira.uka.de>
540 * garray.c (g_ptr_array_remove_index): Fixed size in g_memmove,
541 reported by Alexander Larsson <alla@lysator.liu.se>.
543 * gmem.c: Fixed bug, that made compile fail for -DENABLE_MEM_PROFILE.
545 Wed Dec 16 23:04:26 CST 1998 Shawn T. Amundson <amundson@gtk.org>
547 * Released GLib 1.1.9
554 docs/glib-config.1: version=1.1.9
556 Wed Dec 16 22:32:13 CST 1998 Shawn T. Amundson <amundson@gtk.org>
559 giounix.c: s/g_main_poll_add/g_main_remove_add/g
560 s/g_main_poll_remove/g_main_remove_poll/g
561 s/g_main_poll_add_unlocking/g_main_add_unlocking_poll/g
564 Wed Dec 16 20:04:10 PST 1998 Manish Singh <yosh@gimp.org>
566 * configure.in: check for pthread_attr_init in all cases, du4
567 needs this since most of the other functions are just #defines
569 Thu Dec 17 04:10:49 1998 Tim Janik <timj@gtk.org>
571 * glib.h (G_LOCK_DECLARE_*): if !G_THREADS_ENABLED, eat the
572 trailing semicolon with a bogus function declaration, instead
573 of with a bogus variable declarations, so we avoid unused
576 Wed Dec 16 07:49:16 PST 1998 Shawn T. Amundson <amundson@gtk.org>
578 * Released GLib 1.1.8 (CVS tag is GLIB_1_1_8a)
580 1998-12-16 Sebastian Wilhelmi <wilhelmi@ira.uka.de>
582 * giounix.c (g_io_channel_unix_get_fd): Fixed small bug.
584 1998-12-16 Joel Becker <jlbec@ocala.cs.miami.edu>
586 * glib.h, gscanner.c: changed func_data to user_data in
587 g_scanner_scope_symbol_foreach and associated friends, because
588 AIX headers #define func_data.
590 1998-12-16 Sebastian Wilhelmi <wilhelmi@ira.uka.de>
592 * configure.in: Fixed stupid bug, that made
593 g_static_mutex_lock(*mutex) not work as expected.
595 * docs/glib-config.1: Updated to reflect the existence of gthread.
597 * gmain.c (g_main_poll_add_unlocking): Changed
598 g_main_poll_add_unlocked to g_main_poll_add_unlocking to match
599 semantic, (indeed, main_loop must be locked, when calling this
600 function). Removed the unlocking from the end of that function, as
601 that is not right. Made a 'HOLDS' comment above the function.
603 Wed Dec 16 03:16:58 1998 Tim Janik <timj@gtk.org>
605 * configure.in: version bump to 1.1.8, binary age 0, interface age 0.
607 * glib.h: changed g_lock() to G_LOCK(), g_unlock() to G_UNLOCK() and
608 g_trylock() to G_TRYLOCK(), since these are macros that expand to
609 nothing with --disable-threads.
610 changed G_LOCK_DEFINE() to G_LOCK_DECLARE() and introduced
611 G_LOCK_DECLARE_STATIC() to achive the results of static G_LOCK_DECLARE().
612 changed semantics of g_thread_supported to g_thread_supported() so it
613 can be used as a function like g_module_supported(). the actuall
614 definition is still a macro that expands into a variable for
615 performance reasons though.
616 various indentation and coding style cleanups.
618 * configure.in: added --enable-threads that defaults to yes.
620 * gmutex.c: changed tests g_thread_supported to g_thread_supported (),
621 changed variable settings of g_thread_supported
622 to g_threads_got_initialized.
636 changed s/g_lock/G_LOCK/, s/g_unlock/G_UNLOCK/,
637 s/static G_LOCK_DEFINE/G_LOCK_DECLARE_STATIC/.
639 Tue Dec 15 23:16:05 CST 1998 Shawn T. Amundson <amundson@gtk.org>
646 docs/glib-config.1: Incremented version
648 Wed Dec 16 22:29:48 1998 Joel Becker <jlbec@ocala.cs.miami.edu>
650 * configure.in: fixed the AIX thread checking.
651 Just an error on the wildcard. BTW, -D_THREAD_SAFE
654 Wed Dec 16 02:02:48 1998 Tim Janik <timj@gtk.org>
656 * glib-config.in (lib_glib): fixed library and flag ordering for
659 Tue Dec 15 17:17:46 1998 Owen Taylor <otaylor@redhat.com>
661 * glib.h giounix.c giochannel.c: Use an "inheritance"
662 scheme for IO channel memory allocation.h
664 1998-12-15 Havoc Pennington <hp@pobox.com>
666 * gdate.c (g_date_prepare_to_parse): Solaris has a broken strftime
667 that produced garbage output for the test date I was using to
668 set up the parser. So use a different date that Solaris seems
671 1998-12-15 Sebastian Wilhelmi <wilhelmi@ira.uka.de>
673 * configure.in: Dont complain, if --without-threads or
674 --with-threads=none is supplied; Test for pthread_attr_init
675 instead of pthread_cond_init, if threads seems to be supported by
676 standard glib. (CFLAGS): Use G_THREAD_CFLAGS for compiling of glib
679 * glib.h, gmutex.c: Changed private to private_key to avoid
680 problems when compiling with under C++.
682 1998-12-15 Jeff Garzik <jgarzik@Pobox.com>
685 bugfixes, correctly support --without-threads and friends
687 * gthread/testgthread.c: corrected 64-bitness problem
689 Tue Dec 15 10:40:09 1998 Owen Taylor <otaylor@redhat.com>
691 * gnode.c glist.c gslist.c: Make sure all
692 calls to g_node_validate_allocator are within
693 current_allocator lock, so we have consistency
694 on that point. (Should not really matter,
695 but this way we match the comments)
697 * glist.c (g_list_free_1): Removed some lines
698 that should never have been committed. (For
701 1998-12-11 Sebastian Wilhelmi <wilhelmi@ira.uka.de>
703 * gthread/gthread-nspr.c, configure.in: Added new default thread
704 implementation on top of the mozilla nspr library.
706 * gmem.c, gmessaged.c, gthread/gthread.c, gthread/gthread-*.c:
707 Changed GStaticMutex and GStaticPrivate in gmem.c and gmessages.c
708 into GMutex and GPrivate resp. to make error reporting and use of
709 gmem possible in most (not all, though) gthread functions. Also
710 initialized the modules via new init functions.
712 * configure.in: Fixed syntax bug in definition of type
715 * gthread/testgthread.c: Updated to work with nspr, but see note
716 there for remaining problems.
718 1998-12-10 Sebastian Wilhelmi <wilhelmi@ira.uka.de>
720 * gmutex.c, glib.h: Now abort, if a mutex/cond/private is
721 allocated before the thread system is set up.
723 * gthread/gthread.c (g_thread_init): Removed g_thread_try_init(),
724 as it is not necessary. Changed the error message. Corrected logic
725 for g_thread_use_default_impl.
727 * gmutex.c (g_mutex_init): Keep the thread private data array
728 after calling g_thread_init().
730 1998-12-09 Sebastian Wilhelmi <wilhelmi@ira.uka.de>
732 * gthread/testgthread.c (new_thread): Now also working for posix
733 threads; (wait_thread): Now a better implementation, that does not
736 * Made the thread related code follow GNU coding standard.
738 * Made a comment (HOLDS:) above each function, that expects the
739 given locks to be held.
741 * Changed try_lock to trylock throughout.
743 * glib.c: Eventually removed the #if 0'ed code for old GStaticMutex.
745 * glib.c: Corrected g_trylock macro for G_DEBUG_LOCKS.
747 * gmain.c (g_main_poll_add_unlocked): first take a new poll record
748 form the poll_free_list.
750 * gmem.c, gstrfuncs.c, gutils.c: Made it MT safe.
752 * gthraed/*.c: Added copyright headers.
754 * gthread/gthread-solaris.c: do not use g_log for errors, as g_log
755 uses these module and endless recursions might happen, just use a
756 plain fprintf(stderr,...).
758 * gthread/gthread.c (g_thread_try_init): Call g_mutex_init().
760 * gthread/testgthread.c: updated test program.
762 Tue Dec 8 18:49:56 1998 Owen Taylor <otaylor@redhat.com>
764 * Start at adding thread-safety. (mostly work
765 of Sebastian Wilhelmi <wilhelmi@ira.uka.de>)
767 - configure.in now looks for a system thread implementation.
768 Currently support is included for POSIX threads
769 and Solaris threads. The default support is built
770 into a separate library -lgthread.
772 - The thread implementation can be modified by passing
773 a vector of functions g_thread_init().
775 - The default or supplied functions are used to
776 implement a small set of thread functions for
777 mutexes, condition variables, and thread-private
780 - GLib now uses these functions to provide thread
781 safety. (In the sense that all global static
782 data is locked... individual structures must still
783 be locked by the caller.)
785 Sat Dec 12 19:08:59 1998 Tim Janik <timj@gtk.org>
787 * configure.in: always define G_HAVE_INLINE if __cplusplus is
788 defined, reported by Wan-Teh Chang <wtc@netscape.com>.
790 Thu Dec 10 21:49:39 CST 1998 Shawn T. Amundson <amundson@gtk.org>
792 * Released GLib 1.1.7
799 docs/glib-config.1: Increased version to 1.1.7
801 Wed Dec 9 22:44:44 EST 1998 Joel Becker <jlbec@ocala.cs.miami.edu>
803 * Removed #define of G_COMPILED_WITH_DEBUGGING from glibconfig.h
804 It shouldn't be there, and it was causing a double #define.
805 (config.h defines it also)
807 Tue Dec 8 12:18:38 CST 1998 Shawn T. Amundson <amundson@gtk.org>
809 * Released GLib 1.1.6
813 Mon Dec 7 23:10:41 PST 1998 Manish Singh <yosh@gimp.org>
815 * gmain.c: fixes to #undef HAVE_POLL case
817 1998-12-02 Havoc Pennington <hp@pobox.com>
819 * gdate.c (g_date_set_month): If Julian is valid, we have to
820 update the dmy representation before setting the components
822 (g_date_set_day): Same.
823 (g_date_set_year): Same.
825 1998-12-02 Havoc Pennington <hp@pobox.com>
827 * testgdate.c, testgdateparser.c: Two new files. This is kind of
828 ugly code, but I want to go ahead and make the tests available.
829 It isn't contaminating any other code. :-) Since one of these
830 is interactive and the other takes a while to run, I've kept
831 them separate from testglib for now.
832 * Makefile.am: Build gdate test programs.
834 1998-12-02 Havoc Pennington <hp@pobox.com>
836 * gdate.c (g_date_valid): Rely on GDate::dmy and GDate::Julian
837 flags, rather than re-checking the validity of the actual
838 values. This should be the correct behavior, the old way was
841 * glib.h, gdate.c: Changed MDY to DMY throughout.
843 Sat Nov 28 12:53:47 1998 Owen Taylor <otaylor@redhat.com>
845 * Makefile.am configure.in acconfig.h giochannel.c
846 glib.h glist.c gmain.c gutils.c:
848 - Revised GIOChannel to provide a generic virtual-function
850 - Added unix fd-based GIOChannel's
851 - Added generic main-loop abstraction
852 - Added timeouts and idle functions using main-loop abstraction.
854 1998-12-02 Sebastian Wilhelmi <wilhelmi@ira.uka.de>
857 * gdate.c: changed `gpointer struct_tm_p' parameter of
858 g_date_to_struct_tm back to `struct tm *tm' and forward declared
859 `struct tm' in glib.h; yes, this is nice, we still need not
862 Tue Dec 1 23:01:44 CST 1998 Shawn T. Amundson <amundson@gtk.org>
868 docs/glib-config.1: Incremented versions to 1.1.6.
870 Wed Dec 2 02:10:59 1998 Tim Janik <timj@gtk.org>
872 * gdate.c: s/time_t/GTime/ and s/g_print/g_message/
875 * glib.h: removed #include <time.h>, changed time_t paramter of
876 g_date_set_time() to time_t, changed struct tm parameter of
877 g_date_to_struct_tm to `gpointer struct_tm_p'. yes, this is not
878 nice, but including time.h actually breaks a bunch of code.
880 * incremented GLib version to 1.1.6.
882 1998-11-30 Havoc Pennington <hp@pobox.com>
884 * gdate.c: New file, implements calendrical calculations.
886 * glib.h: Added declarations for GDate module.
888 Mon Nov 30 07:12:10 1998 Tim Janik <timj@gtk.org>
891 * ghook.c: added g_hook_list_marshal_check() to eventually destroy
892 hooks after they got marshalled.
894 Sun Nov 29 17:31:43 EST 1998 Jeff Garzik <jgarzik@pobox.com>
896 * configure.in, Makefile.am, gmodule/Makefile.am:
897 Put -DFOO stuff into Makefile.am INCLUDES.
899 Sun Nov 29 14:17:09 PST 1998 Manish Singh <yosh@gimp.org>
901 * configure.in: put debug -DFOO stuff into CPPFLAGS, not CFLAGS,
902 so CFLAGS can be overridden at make time properly
904 Sat Nov 28 01:23:25 1998 Tim Janik <timj@gtk.org>
906 * ghash.c: implemented incremental freezing facility.
908 Thu Nov 26 01:36:20 1998 Tim Janik <timj@gtk.org>
911 * ghash.c: reverted the g_hash_table_set_key_freefunc() addition,
912 since it's to specialized and needs to be resolved in a generic
915 Tue Nov 24 18:57:59 PST 1998 Manish Singh <yosh@gimp.org>
917 * applied glib-tml-981120-0, change log appended below.
919 * glibconfig.h.win32: passthrough 64-bit constants unchanged, VC++
922 Fri Nov 20 22:26:43 1998 Tor Lillqvist <tml@iki.fi>
924 * glib.h: Moved MSC pragmas from glib.h to glibconfig.h.win32.
925 peer_offset field in WIN32 part of GIOChannel removed, need_wakeups
926 added. Added "extern" to __declspec(dllimport).
928 * gutils.c: Initialise need_wakeups.
930 * glibconfig.h.win32: Pragmas moved here. Define G_GINT64_CONSTANT.
932 * gmodule/gmodule.def: Added g_module_build_path.
934 * gscanner.c: (g_scanner_cur_value) Move initailisation of v
937 * glib.def: Added g_(s)list_sort.
939 * makefile.msc: A few more comments.
941 Tue Nov 24 14:05:47 EST 1998 Michael K. Johnson <johnsonm@redhat.com>
943 * glib.h: added GFreeFunc and g_hash_table_set_key_freefunc()
945 * ghash.c: added g_hash_table_set_key_freefunc() implementation.
946 Modified the prototypes of the functions g_hash_node_destroy() and
947 g_hash_nodes_destroy(), and changed the functions that call them
948 to match the new definitions.
949 This changes no external interfaces, and should create no binary
950 or source incompatibilities. It does add a member to the
951 GHashTable structure.
953 Tue Nov 24 09:40:00 1998 Tim Janik <timj@gtk.org>
955 * glib.h: removed the GListAllocator type and its g_*_allocator_*()
956 function variants (which weren't working anyways) in favour of a
957 generic GAllocator type. new functions:
958 g_allocator_new, g_allocator_free, g_slist_push_allocator,
959 g_slist_pop_allocator, g_list_push_allocator, g_list_pop_allocator,
960 g_node_push_allocator and g_node_pop_allocator.
962 * gstring.c: removed bogus slist allocator code.
963 * gtree.c: maintain own list of free tree nodes and don't waste
964 GSLists for that, removed bogus slist allocator code.
965 * glist.c: use GAllocators for node allocation.
966 * gslist.c: use GAllocators for node allocation.
967 * gnode.c: use GAllocators for node allocation.
969 * gdataset.c: cleanups wrt automatic initialization.
971 Mon Nov 23 10:03:58 1998 Owen Taylor <otaylor@gtk.org>
973 * glib.h garray.[ch]: added g_array_insert_vals() to
974 insert elements at an arbitrary index, and
975 g_array_insert_val() macro.
977 Sun Nov 22 17:07:03 1998 Tim Janik <timj@gtk.org>
980 * gslist.c: new function g_slist_copy() to duplicate a list with all its
982 * glist.c: new function g_list_copy.
984 Sat Nov 21 14:57:39 CST 1998 Shawn T. Amundson <amundson@gtk.org>
990 glib.spec: Updated to version 1.1.5
992 * Released GLib 1.1.5
994 Wed Nov 18 1998 Elliot Lee <sopwith@redhat.com>
996 * gdataset.c: There was a code path into g_data_set_internal
997 through a function (g_datalist_id_set_full). Neither function
998 checked for g_dataset_init having been run, but
999 g_data_set_internal used the memchunk. I added
1000 a check into g_data_set_internal so it will do the
1001 initialization if needed. (There's probably a better way.)
1003 Mon Nov 16 07:48:06 1998 Tim Janik <timj@gtk.org>
1005 * glib.h (g_bit_nth_msf): fixed off-by-one error, so we don't waste
1006 a loop iteration if (-1) was passed, reported by Andreas Bombe
1007 <andreas.bombe@munich.netsurf.de>.
1009 Fri Nov 13 15:17:34 1998 Owen Taylor <otaylor@redhat.com>
1011 * glist.c gslist.c glib.h: Added g_list_sort() and
1012 g_slist_sort() to merge sort GLists and GSLists.
1013 Submitted by Sven Over <sven.over@ob.kamp.net>
1016 * testglib.c: Test the new sort functions.
1018 Wed Nov 11 23:23:22 EST 1998 Jeff Garzik <jgarzik@pobox.com>
1020 * Makefile.am : INCLUDES is the right way to add to CFLAGS, not
1021 DEFS. Also there are bugs with '+=' in makefiles.
1022 Got rid of DEFS line by moving G_LOG_DOMAIN setting into INCLUDES.
1023 Removed redundant -I from INCLUDES.
1025 Wed Nov 11 18:11:24 EST 1998 Gregory McLean <gregm@comstar.net>
1027 * docs/*.sgml : Batch of new documentation that should be easier
1028 to maintain and extend. Plus generate whatever sort of doc file
1029 you would like. I didn't change the Makefile stuff as I'm not sure
1030 what default doc type people want. Oh and this is all DocBook format.
1033 Tue Nov 10 17:12:45 PST 1998 Manish Singh <yosh@gimp.org>
1035 * configure.in: use __extension__ for long long on gcc >= 2.8 and
1036 egcs, and provide a G_GINT64_CONSTANT wrapper so -ansi -pedantic
1039 * glib.h: make the endian x86 asm __const__ so the compiler can do
1040 better optimizations. Also remove the cc clobber, these shouldn't
1041 be changing condition codes. Ditch some redundant casts. Add an
1042 optimization for 64-bit endian conversions in x86. Use constant
1043 wrapper for the generic method.
1045 * testglib.c: use constant wrappers for 64-bit constants
1047 1998-11-04 Phil Schwan <pschwan@cmu.edu>
1049 * configure.in: Added 'strncasecmp' to the list of functions to be
1051 * glib.h: Added a prototype for 'g_strncasecmp'
1052 * gstrfuncs.c: (g_strncasecmp) new function modeled closely after
1055 Wed Nov 4 15:06:44 PST 1998 Manish Singh <yosh@gimp.org>
1057 * config.h.win32 (new file)
1058 * glibconfig.h.win32
1060 * makefile.msc: applied glib-tml-981104, win32 sync
1062 1998-11-03 Sebastian Wilhelmi <wilhelmi@ira.uka.de>
1066 (g_array_remove_index): new function for removing an entry from an
1067 array while preserving the order
1068 (g_array_remove_index_fast): new function for removing an entry
1069 from an array. the order might be distorted
1070 (g_ptr_array_remove_index_fast, g_ptr_array_remove_fast): new
1071 functions; working similiar to the above. (they have the semantic
1072 of the old g_ptr_array_remove[_index] functions)
1073 (g_ptr_array_remove_index, g_ptr_array_remove): new semantic. now
1074 the order of the elements in the array is not changed
1075 (g_byte_array_remove_index, g_byte_array_remove_index_fast): new
1076 functions; byte_array wrapper for g_array_remove_index[_fast]
1078 Sun Nov 1 23:00:18 PST 1998 Manish Singh <yosh@gimp.org>
1081 * configure.in: endian macros defined using the glibconfig.h
1084 * ghook.c: casts for GHookFunc and GHookCheckFunc to avoid warnings
1086 Sat Oct 31 20:14:22 PST 1998 Manish Singh <yosh@gimp.org>
1088 * applied glib-tml-981101-1 patch from Tor Lillqvist (ChangeLog
1089 entry appended below)
1093 * glib.h: use G_HAVE_GINT64, since HAVE_GINT64 is gone
1096 * gscanner.c: #include <config.h> in here too, for HAVE_UNISTD_H
1098 1998-11-01 Tor Lillqvist <tml@iki.fi>
1100 * glib.def: Added the new functions.
1103 (g_getenv): Better implementation on Win32. No loop necessary.
1104 (g_get_any_init): Use P_tmpdir if defined as tmp directory. Don't
1105 set home dir always to NULL on Win32. Don't use tmp directory as
1106 last resort home directory, but return NULL if no home
1107 directory is known (application must check).
1109 * makefile.msc: Better touch command, just COPYing a single file
1110 sets the timestamp from that file, which isn't what touching should
1113 1998-10-31 Raja R Harinath <harinath@cs.umn.edu>
1115 Land glib-rrh-19981025-0.patch.
1116 * Makefile.am (glibconfig.h): New rule.
1117 (stamp-gc-h): New rule. Generate `glibconfig.h'.
1118 * configure.in (AM_CONFIG_HEADER): It is now `config.h'.
1119 (HAVE_BROKEN_WCTYPE): On Solaris, look for iswalnum in -lw before
1120 concluding "broken wctype".
1121 (glibconfig.h): Use AC_OUTPUT_COMMANDS to put generation code into
1123 * glib.h: Remove a lot of tests and defines. All these have been
1124 moved to `configure.in (glibconfig.h)'.
1125 * gerror.c: Include <config.h>.
1127 * gstrfuncs.c: Likewise.
1128 * gutils.c: Likewise.
1130 Sat Oct 31 05:08:26 1998 Tim Janik <timj@gtk.org>
1132 * glib.h: removed old G_ENUM(), G_FLAGS(), G_NV() and G_SV() macros.
1133 added macros G_STRUCT_OFFSET(), G_STRUCT_MEMBER_P() and
1134 G_STRUCT_MEMBER() for handling structure fields through their offsets.
1135 (struct _GHookList): added a hook_free function member, that can be used
1136 to free additional fields in derived hook structures.
1137 g_hook_free(): if hook_list->hook_free != NULL, call this function prior
1138 to freeing the hook. (this functionality should have been there in the
1139 first place, it just got forgotten as an implementation detail).
1141 Wed Oct 28 00:49:32 PST 1998 Manish Singh <yosh@gimp.org>
1143 * glib.h: G_BREAKPOINT for alpha from Robert Wilhelm
1144 <robert@physiol.med.tu-muenchen.de>
1146 Tue Oct 27 07:25:53 1998 Tim Janik <timj@gtk.org>
1150 (g_scanner_sync_file_offset): new function for file input.
1151 rewind the filedescriptor to the current buffer position and blow
1152 the file read ahead buffer. usefull for third party uses of our
1153 filedescriptor, which hooks onto the current scanning position.
1154 (this became neccessary with the implementation of buffered
1156 (g_scanner_input_file):
1157 (g_scanner_input_text): automatically blow the read ahead buffer.
1158 (g_scanner_get_char): blow the read ahead buffer when the end of
1159 input is reached, i.e. a '\000' char is read.
1161 Tue Oct 27 00:27:31 PST 1998 Manish Singh <yosh@gimp.org>
1163 * glib.h: get rid of duplicate #g_htonl and friends in the
1166 Mon Oct 26 22:42:01 PST 1998 Manish Singh <yosh@gimp.org>
1168 * glibconfig.h.win32
1170 * makefile.msc: win32 patch from Tor Lillqvist (glib-tml-981027-0),
1171 fixes some minor win32 buglets
1173 Mon Oct 26 22:12:03 PST 1998 Manish Singh <yosh@gimp.org>
1175 * glib.h: reimplemented the endian stuff, using inline asm
1176 for x86. #define g_htonl and friends.
1178 * testglib.c: new tests for the endian stuff
1180 * configure.in: care for AIX in gmodule linker flags test (from
1181 Joel Becker <jlbec@raleigh.ibm.com>). Check $host_os for linux
1182 instead of existance of /usr/include/linux
1184 * gutils.c: buh-bye evil warning. Thou hast been #ifdef'd out
1187 Tue Oct 27 03:00:50 1998 Tim Janik <timj@gtk.org>
1189 * glib.h: removed dummy structure definitions for struct _GCache,
1190 _GTree, _GTimer, _GMemChunk, _GListAllocator and _GStringChunk.
1192 * gutils.c: implement glib's inline functions _after_ all include
1193 statements have been processed.
1194 removed Tor's MAXPATHLEN check since there already was one supplied
1195 further down in this file.
1196 (LibMain): special cased the #ifdef __LCC__ case for NATIVE_WIN32,
1197 since lcc maybe used on other platforms as well. why in hell is this
1199 (g_get_any_init): for windows, if the user name is supplied, use it as
1201 in general, if there is no homedir specified, use the tmpdir that
1204 * gtimer.c (g_timer_elapsed): changed a g_assert() statement to
1207 * applied glib-tml-981020-0.patch for WIN32 portability, added some
1208 comments and g_return_if_fail() statements, minor indentation fixes.
1209 ChangeLog entry from Tor Lillqvist is appended.
1211 * glib.h (struct dirent): use lower case structure members.
1216 s/COMPILING_GLIB/GLIB_COMPILATION/
1218 1998-10-20: Tor Lillqvist <tml@iki.fi>
1220 * README.win32 glib.def gmodule.def
1221 * glibconfig.h.win32 gmodule/gmoduleconf.h.win32:
1222 New files for the Windows port. The .def files list exported
1223 symbols for the Microsoft linker and compatibles.
1226 Added checks for some platform-dependent headers: pwd.h sys/param.h
1227 sys/select.h sys/time.h sys/times.h unistd.h, and the function lstat.
1230 Conditionalized inclusion of system-dependent headers. Changes
1231 for Windows: no gdb to do a stack trace. Just call abort().
1234 Changes for Windows:
1236 Added macros G_DIR_SEPARATOR, G_DIR_SEPARATOR_S for
1237 platform-dependent file name syntax elements. Added macros
1238 G_SEARCHPATH_SEPARATOR, G_SEARCHPATH_SEPARATOR_S for
1239 platform-dependent search path syntax conventions.
1241 Added pragmas for Microsoft C to make it more pedantic.
1243 Marked GLib's global variables for export from DLL.
1245 Added the function g_strescape that escapes backslashes.
1247 Added functions g_path_is_absolute and g_path_skip_root to
1248 handle platform-dependent file name syntax.
1250 Added the function g_getenv that expands environment variables
1251 that contain references to other environment variables, as is
1252 typical on Windows NT.
1254 Added the GIOChannel structure which is used to encapsulate the
1255 IPC mechanism used by the GIMP's plug-ins, and possibly other
1256 things later. On Unix a GIOChannel encapsulates just a file
1257 descriptor. On Windows it contains a file handle from _pipe() and a
1258 few other things related to the implementation of gdk_input_add
1259 and GIMP plug-in communication. Subject to change.
1261 Removed duplicate declarations of the version variables.
1263 For the Microsoft compiler, declare own implementation of
1264 ftruncate and the <dirent.h> functions.
1267 Define a symbolic name for the profiling table size.
1270 Conditionalized inclusion of unistd.h. On Windows, output using
1274 Conditionalized inclusion of unistd.h. Added changes for
1275 Microsoft C. Added CR to the skipped character set. Added small
1276 workaround for MSC compiler bug in g_scanner_cur_value.
1279 Added the function g_strescape, which escapes the backslash
1280 character. Needed especially when printing Windows filenames.
1283 Conditionalized inclusion of unistd.h and sys/time.h. Added
1284 implementations for Windows.
1287 Conditionalized inclusion of platform-dependent headers. Use
1288 the platform-independent file name syntax macros.
1289 Conditionalize code on platform-dependent features. Added the
1290 functions g_path_is_absolute g_path_skip_root and g_getenv.
1291 Added the GIOChannel-related functions. Added
1292 compiler-dependent Unix compatibility functions for Windows.
1294 * makefile.lcc makefile.msc:
1295 New files. Compiler-specific makefiles for LCC-Win32 and
1296 Microsoft C. Only Microsoft C is actually supported currently.
1299 Added pathname check cases for Windows. Added workaround for
1300 bug in the Microsoft runtime library. Improved some tests a bit.
1302 Sun Oct 25 01:24:01 CST 1998 Shawn T. Amundson <amundson@gtk.org>
1304 * Released GLib 1.1.4
1308 * docs/glib-config.1:
1309 * docs/glib.texi: Moved docs from gtk and created glib-config.1
1310 from gtk-config.1 in gtk
1314 * sanity_check: added in docs directory, change rule 'release'
1315 to 'snapshot' and created a new 'release' rule for doing the
1316 distribution. Added in a 'sanity' rule.
1320 * INSTALL: Updated for the release
1322 Sun Oct 25 07:30:10 1998 Tim Janik <timj@gtk.org>
1325 (g_scanner_peek_next_char):
1326 (g_scanner_get_char): no more characters are available if read() returns
1327 anything less than 1, not only on 0 return.
1330 * gstrfuncs.c: final API cleanup of string arrays to suit sopwith's
1331 argument order requirements.
1332 (g_strsplit): renamed from g_str_array_split().
1333 (g_strjoinv): renamed from g_str_array_joinv(), string array is
1334 now passed as last parameter. removed `const' qualifier from string
1335 array pointer (again).
1336 (g_strjoin): new function from sopwith to concatenate strings with
1337 an additional seperator.
1338 (g_strfreev): renamed from g_str_array_free.
1340 Sat Oct 24 22:23:04 PDT 1998 Manish Singh <yosh@gimp.org>
1342 * gnode.c: small fix to g_node_children_foreach to make it work right.
1343 From Paco Moya <paco@cadnotebk17.eecs.berkeley.edu>
1345 Wed Oct 21 19:22:58 1998 Tim Janik <timj@gtk.org>
1347 * configure.in (G_MODULE_HAVE_DLERROR): reverted part of the changes
1348 from Sebastian Wilhelmi (1998-10-20). don't specify a specific loading
1349 behaviour (i.e. RTLD_NOW) when we are not sure whether it is supplied.
1350 changed the return value from (!f2 && f1) to (!f2 || f1) so we default
1351 to not needing an underscore. reverted the extra plugin.c building,
1352 because .lo files are not ELF libraries, this can't work. This needs
1353 further investigation on Solaris...
1355 Wed Oct 21 17:03:05 1998 Tim Janik <timj@gtk.org>
1358 * gstrfuncs.c: API cleanups of the new g_str* functions for consistency
1359 with the existing naming scheme for g_str* functions. grouped the g_str*
1360 functions that will return a newly allocated string seperatedly. all of
1361 the in_place arguments were skipped, the caller is supposed to pass a
1362 g_strdup()ed string if he wants to retrive a new copy. indentation and
1363 coding style fixups. added some g_return_if_fail() statements.
1364 string array functions are prefixed with g_str_array_.
1365 (g_strdelimit): return the modified string like all other g_str*
1366 functions, that operate in place.
1367 (g_strchug): renamed from g_str_chug(), removed in_place argument.
1368 (g_strchomp): renamed from g_str_chomp(), removed in_place argument.
1369 (g_strstrip): renamed from g_str_strip(), removed in_place argument.
1370 (g_str_array_join): renamed from g_strconcatv(), since it actually
1371 operates on a string array and has totaly different semantics from
1372 g_strconcat(). check for separator != NULL, don't segfault if the first
1373 string is NULL. removed the `const' from the string array that's passed,
1374 so users can operate on gchar** string arrays.
1375 (g_str_array_split): renamed from g_str_split() because we actually
1376 produce a string array. reimplemented this function for efficiency.
1377 removed macro aliases g_str_joinv and g_str_join.
1379 1998-10-20 Elliot Lee <sopwith@redhat.com>
1381 * glib.h, gstrfuncs.c: Move string join/split/free routines from
1382 libgnome/gnome-string, rename, and add g_str_chug.
1384 1998-10-20 Sebastian Wilhelmi <wilhelmi@ira.uka.de>
1386 * configure.in: made the check for needed _ in module's func-names
1387 work on solaris. (G_MODULE_NEED_USCORE must be set outside
1388 AC_CACHE_VAL, dlopen(0,0) refused to work on solaris).
1390 Tue Oct 20 03:32:58 PDT 1998 Manish Singh <yosh@gimp.org>
1393 * acconfig.h: added test for endianness
1395 * glib.h: #define endian macros for system and some conversions
1398 Thu Oct 8 06:47:27 1998 Tim Janik <timj@gtk.org>
1401 * gdataset.c: removed functions g_dataset_id_set_destroy and
1402 g_datalist_id_set_destroy and macros g_dataset_set_destroy and
1403 g_datalist_set_destroy.
1404 added new functions g_dataset_id_remove_no_notify and
1405 g_datalist_id_remove_no_notify plus associated macros
1406 g_dataset_remove_no_notify and g_datalist_remove_no_notify, which
1407 will remove a certain data portion without invokation of its destroy
1408 notifier, this should only be used in very controled circumstances.
1410 Wed Oct 7 05:31:24 1998 Tim Janik <timj@gtk.org>
1413 removed the #pragma } statement after extern "C" {. use
1414 (c-set-offset 'inextern-lang 0) to fixup emacs cc-mode indentation.
1417 * ghook.c: API and code cleanups. changed the G_HOOK_ACTIVE and
1418 G_HOOK_IN_CALL enum vals to G_HOOK_FLAG_ACTIVE and G_HOOK_FLAG_IN_CALL.
1419 changed the G_HOOK_IS_ACTIVE() and G_HOOK_IS_IN_CALL() macros to
1420 G_HOOK_ACTIVE() and G_HOOK_IN_CALL().
1421 fixed the g_hook_find* functions, so they iterate over non-active
1423 (g_hook_first_valid):
1424 (g_hook_next_valid): added an extra argument gboolean may_be_in_call,
1425 which indicates whether G_HOOK_IN_CALL() hooks are considered valid
1426 or not. these two functions are meant as iterators for the hook list
1427 invocation, so they need to a) provide functionality to implement
1428 may_recurse bahaviour and b) only walk active hooks.
1430 Tue Oct 6 14:29:47 1998 Tim Janik <timj@gtk.org>
1432 * gmem.c (g_malloc0): fixed memory offsett when ENABLE_MEM_CHECK is
1433 defined and ENABLE_MEM_PROFILE is not (patch from Martin Pool
1434 <mbp@wistful.humbug.org.au>).
1436 Sat Oct 3 01:18:10 1998 Tim Janik <timj@gtk.org>
1440 (g_hook_list_invoke):
1441 (g_hook_list_invoke):
1442 (g_hook_list_marshal): properly handle the first valid hook to call,
1443 it maybe IN_CALL already.
1444 (g_hook_list_marshal): minor name change.
1446 Fri Oct 2 23:21:21 1998 Tim Janik <timj@gtk.org>
1449 * ghook.c: renamed GHook.id to GHook.hook_id so we don't get troubles
1450 with ObjC which reserves "id" as a keyword (Raja R Harinath
1451 <harinath@cs.umn.edu>).
1453 Wed Sep 30 10:53:03 1998 Tim Janik <timj@gtk.org>
1455 * Makefile.am: added ghook.c.
1458 * ghook.c: generic callback maintenance functions.
1460 * glib.h: define G_GNUC_UNUSED.
1462 Fri Sep 25 00:04:37 1998 Tim Janik <timj@gtk.org>
1464 * configure.in: version bump to 1.1.4, binary age 0, interface age 0.
1466 * glib.h (struct _GScanner): dumped peeked_char and text_len in favour
1467 of *text_end and *buffer for buffered read()s.
1469 * gscanner.c: changed peeking and retrival of next character so we
1470 have buffered reads. fixed minor bug with number parsing error
1471 reporting. made some static!!! variables local ones (why did we use
1472 static temporary variables in the lowlevel tokenization code anyways?).
1474 Mon Sep 21 23:23:10 CDT 1998 Shawn T. Amundson <amundson@gtk.org>
1476 * Released GLib 1.1.3
1478 Mon Sep 21 07:43:13 1998 Tim Janik <timj@gtk.org>
1480 * glib.h: added g_renew() which works as g_new() for g_realloc().
1482 Mon Sep 21 02:22:12 1998 Tim Janik <timj@gtk.org>
1484 * NEWS file update for upcoming release of GLib + GModule
1485 version 1.1.3, binary age 0, interface age 0. (GModule uses
1486 the same version numbers as GLib.)
1488 * glib.h: swap the inclusion of of float.h and limits.h to work
1489 around a egcs 1.1 oddity on Solaris 2.5.1 (fix provided by
1490 Per Abrahamsen <abraham@dina.kvl.dk>).
1493 * gscanner.c: renamed the GValue union to GTokenValue, this should
1494 not affect source compatibility in most cases.
1496 * ghash.c: added some g_return_if_fail() statements. make
1497 g_hash_table_lookup_node() an inline function so we save an extra
1498 function invokation on lookups.
1500 Sun Sep 20 18:21:46 1998 Owen Taylor <otaylor@redhat.com>
1502 * ltmain.sh: Patch to libtool-1.2b to make --disable-static
1505 Sun Sep 20 02:09:44 1998 Josh MacDonald <jmacd@axis.hip.berkeley.edu>
1507 * glib.h: New function g_hash_table_foreach_remove is similar to
1508 g_hash_table_foreach, but the callback's return value indicates
1509 whether to remove the element (if TRUE) or not (if FALSE).
1510 Returns the number of elements deleted.
1512 Fri Sep 18 11:31:50 PDT 1998 Manish Singh <yosh@gimp.org>
1515 * gstrfuncs.c: added g_memdup implementation
1517 Fri Sep 18 18:46:14 1998 Tim Janik <timj@gtk.org>
1520 * gdataset.c: make the datalists a safe type (not using a generic
1521 gpointer) by expecting a GData* argument in the g_datalist functions.
1522 provide g_dataset_foreach() and g_datalist_foreach() functions that
1523 allow a GDataForeachFunc function to walk the data lists.
1524 (g_dataset_destroy_internal): made this function truely reentrant (i.e.
1525 can be called from within destroy notifiers as well).
1526 the *_foreach functions are _not_ reentrant (unless all the other
1527 dataset and datalist functions).
1529 Fri Sep 18 03:41:20 1998 Tim Janik <timj@gtk.org>
1531 * gtree.c (g_tree_new): check for key_compare_func != NULL (reported
1534 Thu Sep 17 18:55:46 PDT 1998 Manish Singh <yosh@gimp.org>
1536 * config.h.in: removed from repository
1540 * mkinstalldirs: updated to latest automake version
1542 Thu Sep 17 06:36:25 1998 Tim Janik <timj@gtk.org>
1545 * gdataset.c: implemented g_datalist_* along the lines of g_dataset,
1546 but operates on an opaque gpointer *datalist; pointer, e.g. for the
1547 implementation of GtkObject named data.
1548 we cache a certain portion of the already freed data entries now, to
1549 gain a slight performance improve with data reallocation.
1551 Tue Sep 15 14:57:30 1998 Owen Taylor <otaylor@redhat.com>
1553 * Makefile.am glib-config.in l*: Update to libtool-1.2b,
1554 change library versioning scheme to drop LT_RELEASE
1555 from the -l line, while keeping it in the soname.
1557 Fri Sep 11 02:11:46 1998 Tim Janik <timj@gtk.org>
1559 * glib.h: explicitely include the prototypes for inline functions
1560 to cure gcc warnings for -Wmissing-prototypes.
1562 Wed Sep 9 02:52:04 PDT 1998 Manish Singh <yosh@gimp.org>
1564 * configure.in: added -posix check for NeXTStep
1566 Tue Sep 8 05:04:06 1998 Tim Janik <timj@gtk.org>
1568 * glib.h: abandon the use of ATEXIT(), we keep the fallback
1569 macros for backwards compatibility. people ought to use g_atexit().
1571 * gutils.c (g_atexit): new function to take over the implementation
1572 of ATEXIT. this function is guarranteed to succeed, similar to
1575 Mon Sep 7 20:07:38 PDT 1998 Manish Singh <yosh@gimp.org>
1577 * configure.in: comment the -std1 check and save LIBS properly
1579 Mon Sep 7 07:53:21 1998 Tim Janik <timj@gtk.org>
1581 * configure.in: check for all three inline keywords individually.
1583 * glib.h: inlining hassle. for compilers that don't allow the `inline'
1584 keyword, mostly because of strict ANSI C compliance or dumbness, we try
1585 to fall back to either `__inline__' or `__inline'.
1586 we define G_CAN_INLINE, if the compiler seems to be actually *capable*
1587 to do function inlining, in which case inline function bodys do make
1588 sense. we also define G_INLINE_FUNC to properly export the function
1589 prototypes if no inlinig can be performed. we special case most of the
1590 stuff, so inline functions can have a normal implementation by defining
1591 G_INLINE_FUNC to extern and G_CAN_INLINE to 1.
1593 * ltconfig: (compiler PIC flag test): special case linux for non
1594 aout systems to honour lcc's position independant code (cases
1595 "linux*aout)" and "linux*)" got added). (this needs to go into
1596 libtool which does an advanced test, checking for __LCC__).
1598 * autogen.sh: take $CC=lcc into account by invoking automake with
1599 --include-deps so lcc isn't scared by gcc's auto-dependancy
1600 generation code. care about $ACLOCAL_FLAGS. optionally feature
1603 * minor fixups in other places to cure some of lcc's warnings.
1605 Sun Sep 6 19:08:53 PDT 1998 Manish Singh <yosh@gimp.org>
1607 * configure.in: added -std1 check for ANSI compliance (from gtk)
1609 Sun Sep 6 12:31:50 PDT 1998 Manish Singh <yosh@gimp.org>
1611 * glib.h: provide proper ATEXIT behavior on NeXTStep by !atexit
1613 Sat Sep 5 18:03:36 1998 Tom Tromey <tromey@cygnus.com>
1615 * gutils.c (GLIB_INLINE): Define.
1616 * glib.h (GLIB_INLINE): New define.
1617 (g_bit_nth_msf): Use it. Also, add prototype.
1618 (g_bit_storage): Likewise.
1619 (g_bit_storage): Likewise.
1621 Sat Sep 5 04:40:02 1998 Tim Janik <timj@gtk.org>
1624 (g_chunk_new0): use g_mem_chunk_alloc0() to allocate the memchunk,
1625 so the correct size of the memchunk is allocated with 0's and not
1626 a memory portion of the size of the desired type.
1628 * gmem.c: new function g_mem_chunk_alloc0() which will initialize
1629 a memory area allocated with g_mem_chunk_alloc() with 0's.
1631 Wed Sep 2 19:13:28 1998 Owen Taylor <otaylor@redhat.com>
1633 * garray.c glib.h (g_array_[ap/pre]pend_vals): make
1634 data argument const.
1636 Wed Aug 26 06:32:40 1998 Tim Janik <timj@gtk.org>
1639 * gstrfuncs.c: new function g_strnfill() to return a new string
1640 of specified length, filled with a specific character.
1642 Tue Sep 1 23:46:31 1998 Josh MacDonald <jmacd@axis.hip.berkeley.edu>
1644 * testglib.c (main): Update the array tests. Reduce the number of
1645 iterations for the prepend test from 10000 to 100 since it is
1646 O(n^2) and was taking longer than I would like to wait.
1648 * garray.c: I've worked on the GArray interface, mostly. It was
1649 seriously broken before and I hate to do it, but this is going to
1650 break some code. It is important to do this now, because more and
1651 more people are starting to use glib and the interface was both
1652 broken and inconsistent. First, rename the _truncate functions of
1653 both the GArray and GPtrArray classes to _set_size, since this
1654 function can also be used to extend the arrays. GArray now
1655 accepts two more initialization arguments: clear and element_size.
1656 Instead of providing the type to each access function, the array
1657 now stores the element size. Clear, if set, causes the library to
1658 zero element's memory as the array expands. The major broken-ness
1659 here was that array->len was in bytes, not elements. Now, since
1660 the array knows its element size, array->len is correct and I have
1661 removed the g_array_length macro. The only macro which now
1662 accepts the type as an argument is g_array_index, which casts the
1663 element to the right type--this interface does not change. The
1664 append and prepend functions simply need the types removed.
1665 g_ptr_array_remove_index now returns the removed element.
1667 * gprimes.c (g_spaced_primes_closest): Move this function out of
1668 ghash.c and rename it from g_hash_closest_prime. Fix the primes
1669 so that they are actually prime (they weren't all -- isn't that
1672 Mon Aug 24 02:08:56 1998 Tim Janik <timj@gtk.org>
1677 (g_vsprintf): removed this function which was not publically
1678 exported in glib.h. to export it, it should have been named
1679 differently in the first place, since its semantics differ from
1680 vsprintf(). apart from that, it was a possible cause for
1681 problems since it worked on a previously allocated memory area and
1682 was used in a lot places of glib. exporting it would have been a
1683 guararant for problems with threaded programs.
1684 (g_printf_string_upper_bound): exported this function to return
1685 a string size, guarranteed to be big enough to hold the fully
1686 expanded format+args string. added 'q', 'L' and 'll' flag handling.
1687 in fact, the newly allocated area is in most cases much bigger than
1689 (g_strdup_vprintf()): new function returning a newly allocated string
1690 containing the contents of *format and associated args (size is
1691 calculated with g_printf_string_upper_bound()).
1692 (g_strdup_printf): new function which wraps g_strdup_vprintf().
1694 * configure.in: check for va_copy() or __va_copy() alternatively.
1695 check whether va_lists can be copyied by value.
1697 * glib.h: provide a definition for G_VA_COPY.
1703 pass va_lists by value, not by reference, since this causes problems
1704 on platforms that implement va_list as as arrays. internaly, use
1705 G_VA_COPY (new_arg, org_arg); va_end (new_arg); to produce a second
1706 va_list variable, if multiple passes are required. changed all
1711 renamed g_debug() to g_on_error_query(), cleaned up a bit.
1712 renamed g_stack_trace() to g_on_error_stack_trace() since both
1713 functions cluttered different namespaces.
1714 there is an appropriate comment in glib.h now that explains the
1715 unix and gdb specific dependencies of both functions.
1716 removed g_attach_process().
1717 g_on_error_stack_trace() should probably be handled with caution,
1718 i've seem several different linux versions (2.0.x) become unstable
1719 after invokation of this function.
1721 1998-08-18: Elliot Lee <sopwith@redhat.com>
1723 * In gmem.c, add the ability to exclude memory chunks from the
1724 memory profiling information.
1726 Tue Aug 18 18:23:09 PDT 1998 Manish Singh <yosh@gimp.org>
1729 * gstrfuncs.c: added g_strndup
1731 Tue Aug 18 04:40:17 1998 Tim Janik <timj@gtk.org>
1734 * gmessages.c: new function g_log_set_always_fatal() to set an
1735 additional fatal_mask for log levels that are considered to be fatal
1736 globally (required by gtk). since this mask is not domain-associated,
1737 it is restricted to the log levels, introduced by glib itself.
1741 * gtree.c (g_tree_node_check):
1742 don't use g_print() calls for informational/debugging output,
1743 but log all this stuff through g_log() with G_LOG_LEVEL_INFO.
1744 libraries shouldn't use printf(), g_print() or g_printerr() at all.
1746 Tue Aug 18 02:46:44 1998 Tim Janik <timj@gtk.org>
1749 * glib.h (__STRICT_ANSI__): if __STRICT_ANSI__ is defined, make
1750 `inline' a noop, since strict ANSI rules don't permit `inline'.
1752 Mon Aug 17 15:21:42 1998 Tim Janik <timj@gtk.org>
1754 * grel.c: made private functions static.
1756 Sun Aug 16 23:23:46 CDT 1998 Shawn T. Amundson <amundson@gtk.org>
1758 * gmodule/Makefile.am: added gmodule-dl.c and gmodule-dld.c to
1760 * glib.spec: version = 1.1.3
1762 Mon Aug 17 01:46:14 1998 Tim Janik <timj@gtk.org>
1764 * glib.m4: feature an extra MODULES parameter, so glib-config can
1765 be invoked with the "gmodule" argument.
1767 * glib.h: changed the log level to G_LOG_LEVEL_CRITICAL for all
1768 g_return*_if_fail statements, and made them issue a message
1769 like "assertion `%s' failed".
1771 * gmessages.c (g_logv): ugh, don't pass log_domain as NULL to
1774 Sun Aug 16 20:28:27 1998 Tim Janik <timj@gtk.org>
1776 * version bump to 1.1.3, binary age 0, interface age 0.
1778 * glib.h: be nice to platforms that don't have gint64 and don't
1779 issue #warning on every compilation. since glib doesn't require
1780 gint64 itself, packages that need gint64 should test for this
1784 * gutils.c: added a new function g_vsnprintf().
1786 Sun Aug 16 Elliot Lee
1788 glib.h: #error out if we don't recognize the SIZEOF_VOID_P
1789 #warning if no gint64
1791 Fri Aug 14 16:41:53 1998 Tim Janik <timj@gtk.org>
1793 * glib.h: added static inline functions for bit mask tests:
1794 g_bit_nth_lsf, g_bit_nth_msf and g_bit_storage.
1796 Fri Aug 13 14:23:37 1998 Tim Janik <timj@gtk.org>
1800 revised the message handling system, which is now based on a new
1801 mechanism g_log*. most of the assertment macros got adapted to
1802 feature the new g_log() call with an additional specification of
1803 the log level in a preprocessor macro G_LOG_DOMAIN. if G_LOG_DOMAIN
1804 is undefined upon the includion of glib.h, it'll be defined with a
1805 value of (NULL) and thus preserves the original bahaviour for
1806 warning and error messages. the message handler setting functions
1807 for g_warning, g_error and g_message are only provided for backwards
1808 compatibility and might get removed somewhen.
1810 * Makefile.am: feature the G_LOG_DOMAIN macro to set the log domain
1811 to "GLib" upon compilation. we currently have to add this definition
1812 to the DEFS variable.
1813 * testglib.c: we need an ugly #undef G_LOG_DOMAIN at the start
1814 of this file currently, since automake doesn't support per target
1817 * glib.h: changed some gints to gbooleans, made a few const corrections,
1818 removed some superfluous G_STMT_START{}G_STMT_END wrappers, added some
1819 in other required places.
1823 (g_node_insert_before):
1825 (g_node_append_data):
1826 (g_node_prepend_data):
1827 (g_node_insert_data_before):
1828 (g_node_insert_data):
1830 return (node), so these macros/functions can be usefully chained with
1833 Mon Aug 10 17:56:11 PDT 1998 Manish Singh <yosh@gimp.org>
1835 * glib.h: it's GTime now, and it's back!
1837 Mon Aug 10 02:17:19 1998 Tim Janik <timj@gtk.org>
1839 * Makefile.am: minor hack to cause SUBDIRS (gmodule) to be build
1840 last. we do this by making all-recursive-am depend on all-am.
1842 Sun Aug 9 15:56:11 1998 Tim Janik <timj@gtk.org>
1844 * configure.in: added GModule checks. generate files in gmodule/.
1845 * glib-config.in: support library specifications `glib' and `gmodule'.
1846 * Makefile.am: feature the gmodule/ subdir.
1848 Wed Aug 5 10:04:29 PDT 1998 Shawn T. Amundson <amundson@gtk.org>
1850 * Released GLib 1.1.2
1852 Wed Aug 05 01:15:36 1998 George Lebl <jirka@5z.com>
1854 * testglib.c: fix 64-bitness in g_prints, sizeof doesn't
1855 seem to return int so I cast it for printing, probably
1858 Tue Aug 4 19:54:06 PDT 1998 Shawn T. Amundson <amundson@gkt.org>
1860 * Released GLib 1.1.1
1862 Tue Aug 4 15:17:54 1998 Tim Janik <timj@gtk.org>
1864 * configure.in: version bump to 1.1.1, binary age 1, interface age 0.
1867 * INSTALL: updates and fixes.
1868 * COPYING: include the GNU LGPL, rather than shipping an empty file.
1869 * AUTHORS: listed original authors here, and added people who made
1870 significant improvements to glib.
1873 * gutils.c: implement g_get_current_dir() which returns a newly
1874 allocated string, instead of a g_getcwd() variant that operates
1876 export glib_interface_age and glib_binary_age.
1877 as a convenience, macro definitions have been added for
1878 g_node_insert_data, g_node_insert_data_before, g_node_append_data and
1879 g_node_prepend_data.
1881 * testglib.c: minor cleanups, print current dir.
1883 Mon Aug 3 16:02:26 1998 Tim Janik <timj@gtk.org>
1886 * gnode.c: change order of gpointer data; field in struct _GNode to
1887 be partly binary compatible with GList and GSList.
1889 1998-08-03 Sebastian Wilhelmi <wilhelmi@ira.uka.de>
1891 * garray.c (g_ptr_array_remove_index): bugfix: index check for
1892 array has been wrong.
1894 Fri Jul 31 22:17:05 1998 Tim Janik <timj@gtk.org>
1896 * testglib.c (g_node_test): added a GNode test.
1898 Fri Jul 31 09:08:16 1998 Tim Janik <timj@gtk.org>
1900 * Makefile.am: compile gnode.c.
1903 * gnode.c: added implementation of n-way trees.
1905 * gtree.c (g_tree_traverse): added a warning to the switch() statement
1906 which says that G_LEVEL_ORDER is not implemented.
1908 Mon Jul 27 00:17:30 CDT 1998 Shawn T. Amundson <amundson@gtk.org>
1910 * Released GLib 1.1.0
1912 Mon Jul 27 01:02:27 1998 Tim Janik <timj@gtk.org>
1914 * glib.h: #if 0'ed out the GTime definition, until it is definitively
1915 needed. #if 0'ed out the g_getcwd() version, because it is the wrong
1918 Sat Jul 25 16:09:00 1998 Mark Crichton <crichton@expert.cc.purdue.edu>
1920 * glib.h: gtime changed to g_time. gtime is used in
1921 /usr/include/time.h in NetBSD, causing multiple headaches.
1922 If this isn't the right way of fixing it.... ;)
1924 Thu Jul 23 00:29:14 1998 Tim Janik <timj@gtk.org>
1927 * gscanner.c: new functions to make a scanner scope sensitive wrt
1929 g_scanner_scope_foreach_symbol, g_scanner_scope_lookup_symbol,
1930 g_scanner_scope_remove_symbol, g_scanner_scope_add_symbol and
1931 g_scanner_set_scope.
1932 g_scanner_add_symbol, g_scanner_remove_symbol and
1933 g_scanner_foreach_symbol are now aliases for scope 0.
1935 Mon Jul 20 23:05:34 1998 George Lebl <jirka@5z.com>
1937 * glib.h: typo fixed for alphas for gint64
1939 Tue Jul 14 09:05:18 1998 Tim Janik <timj@gtk.org>
1942 * gutils.c: new fuction g_dirname() which returns a newlly
1945 Fri Jul 10 06:33:43 1998 Tim Janik <timj@gtk.org>
1948 * gutils.h: added a bunch of utility/wrapper functions:
1949 g_basename(), g_getcwd(), g_get_user_name(), g_get_real_name(),
1950 g_get_home_dir(), g_get_tmp_dir(), g_get_prgname() and g_set_prgname().
1952 * gutils.c: removed all g_str* functions.
1953 * gstrfuncs.c: moved the bunch g_str* functions from gutils.c in this
1954 place. this file shall never include <unistd.h> to avoid clashes for
1955 some of the g_str* functions on some OSes.
1957 Fri Jul 10 00:29:03 EEST 1998 Lauri Alanko <nether@gimp.org>
1960 * ghash.c: Renamed g_hash_table_lookup_full to
1961 g_hash_table_lookup_extended to conform with naming conventions.
1963 Tue Jul 7 03:18:58 EEST 1998 Lauri Alanko <nether@gimp.org>
1966 * ghash.c: Generic cleanup, added a function:
1967 (g_hash_table_lookup_full): Return whether the lookup succeeded,
1968 and also retrieve the key and value. This allows one to
1969 distinguish between failed lookup and finding a NULL, and also
1970 allows one to free a key in the hash.
1972 Mon Jul 6 10:12:05 PDT 1998 Manish Singh <yosh@gimp.org>
1974 * ltconfig: fix for properly detecting shared lib support on
1975 SunPro cc (taken from libtool 1.2.a)
1977 Sat Jul 4 13:38:52 PDT 1998 Manish Singh <yosh@gimp.org>
1979 * glib.h: added g_array_length
1981 Tue Jun 30 11:58:25 1998 Tim Janik <timj@gtk.org>
1983 * gscanner.c (g_scanner_unexp_token): take symbol_2_token into
1984 account. react on valid/invalid string pairs.
1986 Sat Jun 27 21:55:46 PDT 1998 Manish Singh <yosh@gimp.org>
1988 * glib.m4: ftp.glib.org -> ftp.gtk.org, since glib.org isn't
1991 Fri Jun 19 03:11:02 1998 Tim Janik <timj@gtk.org>
1993 * gdataset.c: removed g_dataset_try_key, g_dataset_force_id and
1994 g_dataset_retrive_key in favour of GQuarks.
1995 a GQuark is an numeric id wich is associated with a certain string.
1996 (g_quark_try_string): try to get the quark associated with this string,
1997 if the lookup failed return 0.
1998 (g_quark_from_string): get the associated quark for a string, if there
1999 isn't currently a GQuark associated with this string, then allocate a
2000 new quark and return that.
2001 (g_quark_from_static_string): like the above function, but the string
2002 isn't strdup()ed to save memory.
2003 (g_quark_to_string): get the string that is associated with a certain
2006 * gdataset.c (g_dataset_id_set_data_full): invoke the destroy function
2007 _after_ the new data has been setup.
2009 Thu Jun 18 02:35:21 1998 Owen Taylor <otaylor@gtk.org>
2011 * glib.h: Changed messages for g_return_[val]_if_fail to
2012 be somewhat more clear: assertion "blah" failed.
2014 1998-06-18 Federico Mena Quintero <federico@nuclecu.unam.mx>
2016 * testglib.c (main): Use GINT_TO_POINTER casts to remove compiler
2019 * grel.c: #include <string.h>
2021 Fri Jun 12 15:39:06 1998 Tim Janik <timj@gtk.org>
2023 * glib.h (GScanner): cleanups of the structure fields (binary
2026 Fri Jun 12 00:39:28 1998 Josh MacDonald <jmacd@icw.EECS.Berkeley.EDU>
2028 * glib.h: add new hash and equal functions g_int_*. complement
2029 g_direct_hash with g_direct_equal.
2031 * grel.c: new file, GRelations implement tuples of N-N mappings.
2032 A comment in glib.h briefly describes the interface.
2034 * ghash.c: new function, g_hash_table_size
2036 * glib.h: new typedefs, gsize, gssize, gtime.
2038 * garray.c: new functions implementing a simplified GArray. This
2039 GPtrArray is an array of gpointers and has functions to add and
2040 remove elements, much like java.lang.Vector.
2042 * garray.c: new functions for the single-byte special case of
2043 GArray. The functions g_byte_array* operate on arrays of bytes.
2044 Internally, a GArray is used.
2046 * testglib.c: tests for g_ptr_array, g_byte_array, and g_relation...
2048 1998-06-11 Federico Mena Quintero <federico@nuclecu.unam.mx>
2050 * gdataset.c: #include <string.h>
2052 Thu Jun 11 04:15:31 1998 Tim Janik <timj@gtk.org>
2055 * gdataset.c: new function g_dataset_retrive_key. adjusted prealloc
2056 sizes, to take up less space on initial allocation.
2058 1998-06-10 Raja R Harinath <harinath@cs.umn.edu>
2060 * acinclude.m4: New file. Contains `libtool.m4' from libtool-1.2,
2061 the version from which glib's libtool forked. Needed for people
2062 who use post-1.2 alphas of libtool.
2063 * configure.in (enable_mem_check, enable_mem_profile): Replace
2064 `echo -n' with AC_MSG_CHECKING.
2065 (fd_set): Explain test for `fd_set' better.
2067 Wed Jun 10 19:29:51 1998 Owen Taylor <otaylor@gtk.org>
2069 * Makefile.am glib.m4 configure.in:
2071 Moved out from GTK+; added AM_PATH_GLIB macro.
2073 Wed Jun 10 12:56:07 1998 Owen Taylor <otaylor@gtk.org>
2075 * glib.h: renamed g_const_pointer => gconstpointer
2077 Tue Jun 9 17:47:33 1998 Owen Taylor <otaylor@gtk.org>
2079 * glib.h: Remove #error - HP/UX.
2081 Sat May 23 19:00:01 1998 Owen Taylor <otaylor@gtk.org>
2083 gtk-rrh-980412-0.patch (Raja R Harinath <harinath@cs.umn.edu>)
2084 gtk-jbuhler-980516-0 (Jeremy Buhler <jbuhler@cs.washington.edu>) ]
2086 * glib.h ghash.c gstring.c gdataset.c gutils.c:
2087 - Added new typedef g_const_pointer; expunged all incorrect
2088 uses of 'const gpointer'.
2089 - Fixed up warnings that that created,
2090 - Changed GHashFunc and GCompareFunc to take g_const_pointer
2091 arguments. (Necessary, but will cause warnings in existing
2093 - Added other new const in harmless positions.
2095 Mon Jun 8 01:06:47 1998 Tim Janik <timj@gtk.org>
2097 * glib.h: added enum-helper macros for code generation.
2098 added G_BREAKPOINT().
2100 Sat Jun 6 14:09:22 PDT 1998 Manish Singh <yosh@gimp.org>
2102 * gmem.c: commented out MEM_PROFILE and MEM_CHECK, causing weird
2105 Wed Jun 3 06:19:42 1998 Tim Janik <timj@gtk.org>
2107 * glib.h (g_chunk_new0): convenience macro, for allocating small chunks
2108 like g_chunk_new() with additional 0 initialization.
2110 Mon Jun 1 04:43:27 1998 Tim Janik <timj@gtk.org>
2112 * ghash.c (g_hash_table_insert): wrote a comment describing why
2113 a hash node's key should not also get replaced when overriding
2116 Tue May 26 18:30:06 1998 Tim Janik <timj@gtk.org>
2118 * glib.h (g_string_sized_new): new function to controll the preallocated
2121 * glib.h (g_strreversed): new function to reverse a string.
2123 Mon May 18 22:14:39 1998 Owen Taylor <otaylor@gtk.org>
2124 (Yasuhiro SHIRASAKI <joke@awa.tohoku.ac.jp> : gtk-joke-980517-0.patch)
2126 * gutils.c: Restored a missing prototype for g_vsprintf.
2128 Wed May 20 05:02:26 1998 Tim Janik <timj@gtk.org>
2130 * glib.h: conditionally define NULL, FALSE and TRUE.
2131 (g_mem_chunk_create): new convenience macro as a short hand for
2133 (g_chunk_free): new convenience macro to be consistent with g_chunk_new.
2135 Tue, 19 May 1998 09:00:02 +0200 Paolo Molaro <lupus@debian.org>
2137 * gcompletion.c: generic functions for com<TAB>pletion...
2139 Sun May 17 10:48:27 1998 Tim Janik <timj@gtk.org>
2141 * gscanner.c (g_scanner_unexp_token): provide usefull default
2142 specifications for identifier_spec and symbol_spec.
2144 * glib.h: new functions g_slist_nth_data and g_list_nth_data to return
2145 the data of the nth element in the list.
2147 Fri May 15 22:31:49 1998 Tim Janik <timj@gtk.org>
2149 * gscanner.c (g_scanner_unexp_token): removed spurious va_end(args)
2150 that for some reason didn't produce a compiler warning on my machine
2151 (is va_end defined to nothing for i386?).
2153 Fri May 15 12:32:08 1998 rodo <doulik@karlin.mff.cuni.cz>
2155 * gscanner.c: replaced some snprintf with g_snprintf
2157 Fri May 15 00:56:59 1998 Tim Janik <timj@gtk.org>
2159 * glib.h: further support for gcc function attributes: G_GNUC_FORMAT,
2160 G_GNUC_NORETURN and G_GNUC_CONST.
2162 * gscanner.c (g_scanner_stat_mode): changed stat() to lstat().
2163 (g_scanner_msg_handler): "\n" at end of line!
2164 (g_scanner_foreach_symbol): new function to iterate over the symbol
2165 table (GScanner does value-wrapping).
2167 Thu May 14 04:14:12 1998 Tim Janik <timj@gtk.org>
2169 * glib.h: typedef gint gboolean;
2170 this is needed to provide portability with big-endian platforms (e.g.
2171 try sizeof(bool) for c++ on big-endians - it's 4).
2172 this is also needed to maintain some_union.d_gint==some_union.d_gboolean.
2173 plus, gint to gboolean casts and vice versa need to be possible without
2176 Tue May 12 19:22:58 1998 Owen Taylor <otaylor@gtk.org>
2178 * glib/glib.h: Added macros G[U]INT_TO_POINTER() and
2179 GPOINTER_TO_[U]INT for storing small integers integers
2182 * glib/testglib.c: Print sizeof() results
2183 as g_print("%ld", (glong)sizeof(foo)), to deal with
2184 size_t being long on Alpha's.
2186 Tue May 12 16:54:15 1998 Owen Taylor <otaylor@gtk.org>
2187 (James A <jamesa@demon.net> : gtk-jamesa-980511-4.patch)
2189 * glib.h gstring.c gmessages.c: Added some missing
2192 * gutils.c (g_strsignal.c): Added missing return statements.
2194 Mon May 11 21:11:54 1998 Owen Taylor <otaylor@gtk.org>
2196 * gutils.c gmessages.c: Moved g_error, g_warning, g_message and
2197 g_print from gutils.c to new file gmessages.c, to avoid having to
2198 include <unistd.h> in gutils.c which was causing problems for the
2199 g_strsignal implementation on FreeBSD boxes.
2201 Mon May 11 09:53:43 1998 Tim Janik <timj@gtk.org>
2203 * configure.in: preserve automake CFLAGS.
2205 * Makefile.am: fully rename the created library to libglib-1.1.la.
2206 this means we need to change certain portions of the Makefile.am on
2207 major/minor version bumps.
2209 * ltmain.sh: the -release option is not required anymore.
2211 * glib.h: provide G_GNUC_FUNCTION and G_GNUC_PRETTY_FUNCTION to
2212 avoid conditionals. unconditionally define NULL, TRUE, FALSE, MAX,
2213 MIN, ABS and CLAMP, these macros might be screwed from other headers.
2215 Mon May 11 01:44:10 1998 Tim Janik <timj@gtk.org>
2217 * gdataset.c: new file, gdatasets implement the object data
2218 mechanism from GtkObject. a generic data pointer is associated with
2219 a certain location and a key id.
2221 Sat May 9 20:08:12 1998 Owen Taylor <otaylor@gtk.org>
2223 * glib/gmem.c: Experimentally restore GMemChunk
2224 to its primeval state - where mem areas are
2225 freed incrementally instead of searching the tree
2226 every time a mem area is completely empty. Also,
2227 always keep one mem chunk around. (Reduced calls
2228 to malloc() a lot, but doesn't really improve
2229 performance significiantly)
2231 Thu May 7 08:17:28 1998 Tim Janik <timj@gtk.org>
2233 * glib.h (G_GNUC_PRINTF):
2234 (G_GNUC_SCANF): macros to facilitate the printf/scanf format argument
2237 * gstring.c: const corrections, string!=NULL checks at function entry.
2238 (g_string_down): new function for tolower(3) conversion.
2239 (g_string_up): new function for toupper(3) conversion.
2241 * gutils.c: const corrections.
2242 (g_strdown): g_string_down() counterpart.
2243 (g_strup): g_string_up() counterpart.
2245 * gscanner.c (g_scanner_unexp_token):
2247 (g_scanner_warn): new functions to let a scanner put out warnings
2248 or errors, especially to react on unexpected tokens.
2251 (g_slist_index): find out about about the position of a
2252 certain data pointer.
2253 (g_slist_position): find out about about the position of a
2257 (g_list_index): find out about about the position of a
2258 certain data pointer.
2260 Thu May 7 05:14:19 1998 Tim Janik <timj@gtk.org>
2262 * ltmain.sh: added a new commandline flag -postfix similar to -release,
2263 but will immediately change the library name.
2265 * Makefile.am: specify -postfix and -version-info
2267 * configure.in: version bump to 1.1.0. added GLIB_INTERFACE_AGE and
2268 GLIB_BINARY_AGE. calculate LT_* variables for libtool.
2270 Fri May 1 16:36:08 1998 Owen Taylor <otaylor@gtk.org>
2272 * gutils.c: (g_strcasecmp). Check for isupper before
2273 taking tolower, and account for macroized tolower.
2275 * gutils.c (g_error): Check for recursion.
2277 1998-04-27 Elliot Lee <sopwith@cuc.ml.org>
2279 * glist.c (g_list_position): New function to find the position of
2280 a link in a list - should be the inverse of g_list_nth(), but
2281 haven't tested it so poof.
2283 Thu Apr 30 21:41:30 1998 Owen Taylor <otaylor@gtk.org>
2285 * gstring.c : Check arguments more carefully,
2286 (gtk-draco-980423-1.patch; ramsey@rhrk.uni-kl.de)
2288 Tue Apr 7 19:36:48 1998 Owen Taylor <owt1@cornell.edu>
2290 * gutils.c (g_direct_compare): Removed, because that's what
2291 a NULL comparison function means. And it wasn't 64 bit safe.
2293 Mon Apr 6 18:43:25 1998 Tim Janik <timj@gtk.org>
2295 * gscanner.c (g_scanner_get_token_ll): fixed a bug that caused floats
2296 of the format ".xxx" to be parsed as "xxx".
2298 Fri Apr 3 20:36:35 1998 Owen Taylor <owt1@cornell.edu>
2300 * gutils.c (g_parse_debug_string): Make debug string
2301 parsine case-insensitive
2303 Fri Apr 3 17:03:18 PST 1998 Manish Singh <yosh@gimp.org>
2305 * gstring.c: corrected possible overrun when inserting into
2306 GStrings (thanks Elrond)
2308 Fri Apr 3 18:05:45 1998 Owen Taylor <owt1@cornell.edu>
2310 * testglib.c: Removed literal german from strings
2311 to appease SGI compiler.
2313 Thu Mar 26 20:47:21 1998 Owen Taylor <owt1@cornell.edu>
2315 * configure.in glib glibconfig.h.in: Add test for atexit/on_exit -
2316 use on_exit if atexit not found in definition of ATEXIT.
2318 Wed Mar 25 15:23:37 1998 Owen Taylor <owt1@cornell.edu>
2320 * Makefile.am: Switched glibconfig.h rule from HEADERS
2321 to DATA, so that it is not added to DISTFILES
2323 Wed Mar 18 22:27:08 PST 1998 Manish Singh <yosh@gimp.org>
2325 * garray.c: g_rarray_truncate length done correctly
2327 Sun Mar 15 07:13:34 1998 Tim Janik <timj@gimp.org>
2329 * gutils.c: changed *_handler variables to be named glib_*_handler,
2330 so you can easily access them from gdb.
2332 Sat Mar 14 17:47:43 1998 Owen Taylor <owt1@cornell.edu>
2334 * Makefile.am: Don't refer to current directory as $(top_builddir)
2335 to avoid confusing non-gmakes
2337 Sat Mar 14 01:37:35 1998 Owen Taylor <owt1@cornell.edu>
2339 * Makefile.am (configincludedir): Moved glibconfig.h to
2340 $(pkglibdir)/include
2342 Tue Mar 10 02:03:12 1998 Tim Janik <timj@gimp.org>
2344 * gscanner.c (g_scanner_destroy_symbol_table_entry): new function to
2345 free symbol table entries upon destruction
2346 (gtk-gronlund-980309-0.patch.gz).
2348 Mon Mar 9 15:02:21 1998 Tim Janik <timj@gimp.org>
2350 * glib.h: changed *_length functions to return guint.
2351 changed *_nth functions to take guint as argument.
2353 * glist.c: adapted g_list_length and g_list_length.
2355 * gslist.c: adapted g_slist_length and g_slist_length.
2357 Mon Mar 2 17:51:18 1998 Owen Taylor <owt1@cornell.edu>
2359 * glib.h gutils.c : changed g_strcasecmp
2360 to take gchar* not guchar*
2362 * testglib.c: Remove trailing ; after functions
2364 Sun Mar 1 19:04:40 1998 Owen Taylor <owt1@cornell.edu>
2366 * glib.h gstring.c: Added g_string_insert[_c]()
2367 and g_string_erase().
2369 From: Stefan Wille <1wille@vsys1.informatik.uni-hamburg.de>
2371 Mon Feb 16 23:05:06 1998 Owen Taylor <owt1@cornell.edu>
2373 * glist.c (g_list_insert_sorted): Changed function
2374 so elements are always inserted, even if they compare
2377 Thu Feb 12 22:48:11 1998 Owen Taylor <owt1@cornell.edu>
2379 * gstring.c glib.h: removed deprecated g_string_equal
2382 Tue Feb 10 13:04:36 1998 Owen Taylor <owt1@cornell.edu>
2384 * configure.in: Add check to see if the C library's
2385 iswalnum can actually be used. (Not true for
2388 Sat Feb 7 11:48:09 1998 Owen Taylor <owt1@cornell.edu>
2390 * gstring.c gutils.c: added some additional consts in
2391 appropriate places to remove a warning
2393 Sat Feb 7 11:15:54 1998 Owen Taylor <owt1@cornell.edu>
2395 * gutils.c: include <ctype.h> for tolower()
2397 Fri Jan 30 23:57:17 PST 1998 Manish Singh <yosh@gimp.org>
2399 * added and autoconfigured in a new utility function
2402 Wed Jan 28 23:53:27 PST 1998 Manish Singh <yosh@gimp.org>
2406 * testglib.c: the sort functions compared backwards. Fixed
2407 * glib.h: list iterator macros now check for NULL pointers
2409 Tue Jan 27 09:46:57 PST 1998 Manish Singh <yosh@gimp.org>
2411 * gstring.c: g_string_prepend and g_string_prepend_c had
2412 interchanged src and dest parameters for g_memmove. Fixed.
2414 Tue Jan 27 01:38:52 PST 1998 Manish Singh <yosh@gimp.org>
2416 * gslist.c: fixed a really, really lame error. g_slist_insert
2417 didn't hook the data in! Reworked the routine to reflect the
2418 functionality of g_list
2420 Wed Jan 21 01:13:25 1998 Tim Janik <timj@psynet.net>
2422 * Applied patch from (Raja R Harinath <harinath@cs.umn.edu>)
2423 to add function g_snprintf.
2424 * configure.in (AC_CHECK_FUNCS): Check for vsnprintf.
2425 * glib.h: Add prototype for g_snprintf.
2426 * glibconfig.h.in: Add HAVE_VSNPRINTF.
2427 * gutils.c (g_snprintf): new function.
2429 Sat Jan 17 23:52:40 1998 Owen Taylor <owt1@cornell.edu>
2431 * gstring.{c,h} gscanner.c:
2432 renamed g_string_equal => g_str_equal
2433 renamed g_string_hash => g_str_hash
2434 And const corrected. Old functions left in for now.
2436 Fri Jan 9 20:03:46 1998 Tim Janik <timj@psynet.net>
2438 * gutils.c (g_strerror): changed message for EAGAIN from
2439 "no more processes" to "try again" since EAGAIN is used with
2440 functions else than fork().
2442 * gscanner.c (g_scanner_get_token_ll): use strtol() instead of
2443 strtoul() to avoid conflicts with solaris.
2445 * merged the glib portions from Jan 2 to Jan 7 out of gtk+/ChangeLog
2448 Wed Jan 7 02:14:30 PST 1998 Manish Singh <yosh@gimp.org>
2453 * testglib.c: Added g_[s]list_insert_sorted function
2454 and appropriate tests in testglib
2456 Sat Jan 3 20:23:25 1998 Owen Taylor <owt1@cornell.edu>
2458 * glib.h: Changed guint32 -> guint for bitfields.
2459 (Bitfields must be int or unsigned int?)
2461 Fri Jan 2 23:52 PST 1998 Jay Painter <jpaint@serv.net>
2465 * glib.h: reverted glibconfig.h and glib.h files back to the
2466 way they were before my ugly hack.
2468 * gscanner.c: removed inlines from clist and gscanner
2470 Tue Dec 23 02:49:51 1997 Tim Janik <timj@psynet.net>
2472 * gscanner.c: new file for GScanner: Flexible lexical scanner for
2474 * glib_pre2.h: added GScanner includes. added g_strconcat and g_strtod.
2475 gutils.c (g_strconcat): new function for string concatenation of NULL
2476 terminated parameter list.
2477 (g_strtod): new function to perform best string to double conversion
2478 with or without consideration of the current locale.
2480 Mon Dec 15 19:33:58 1997 Tim Janik <timj@psynet.net>
2482 * glist.c: minor optimizations:
2483 (g_list_append): `if' optimized for common code path, commented out
2484 unneccessary `assert', saved one variable assignment.
2485 (g_list_prepend): saved two (conditioned) variable assignment.
2486 (g_list_insert): saved one (conditioned) variable assignment,
2487 saved one variable assignment.
2488 (g_list_remove): `if' optimized for common code path, saved two
2489 variable assignments by using `g_list_free_1' (which is even
2490 faster) instead of `g_list_free'.
2491 (g_list_reverse): saved allocation of one variable, saved one
2492 variable assignment.
2494 Wed Dec 10 23:27:20 1997 Tim Janik <timj@psynet.net>
2498 * glib.h: this file now gets concatenated by makeglib_h from
2499 glib_pre1.h and glib_pre2.h to merge in glibconfig.h wich got
2500 created by configure (done by Jay Painter).
2502 * glib_pre2.h: the g_assert*() and g_return_*_fail() macros
2503 are wrapped by G_STMT_START and G_STMT_END now, to avoid conflicts
2504 when used within if (...) g_macro(); else ... conditionals.
2506 Tue Dec 17 13:14:07 1996 Peter Mattis <pmattis@charnley.HIP.Berkeley.EDU>
2508 * glib.h: Changed 'g_return_if_fail' and 'g_return_val_if_fail' to
2509 not call 'g_string' but to simply stringify the
2510 expression. Calling 'g_string' causes the expression to be
2511 expanded which is undesired.
2513 Sun Dec 1 01:30:48 1996 Peter Mattis <pmattis@charnley.HIP.Berkeley.EDU>