platform/upstream/glib.git
12 years agoGDBus: make use of reliable async cancellation
Ryan Lortie [Tue, 13 Mar 2012 20:41:57 +0000 (16:41 -0400)]
GDBus: make use of reliable async cancellation

Call g_simple_async_result_set_check_cancellable() after all
GSimpleAsyncResult creation in order to take advantage of the new
reliable cancellation feature.

The guarantee of reliable cancellation fixes a bug in dbusmenu (which
was already assuming that cancellation was reliable).  See this bug:
https://bugs.launchpad.net/ubuntu/+source/libdbusmenu/+bug/953562

https://bugzilla.gnome.org/show_bug.cgi?id=672013

Signed-off-by: David Zeuthen <davidz@redhat.com>
12 years agoFix "install" paths in VS property sheets
Chun-wei Fan [Thu, 15 Mar 2012 08:39:52 +0000 (16:39 +0800)]
Fix "install" paths in VS property sheets

...and cleanup the VS2010 property sheets a little bit

12 years ago[l10n] Updated German translation
Mario Blättermann [Wed, 14 Mar 2012 21:27:33 +0000 (22:27 +0100)]
[l10n] Updated German translation

12 years agoGSimpleAsyncResult: support reliable cancellation
Ryan Lortie [Tue, 13 Mar 2012 20:09:01 +0000 (16:09 -0400)]
GSimpleAsyncResult: support reliable cancellation

Add a function g_simple_async_result_set_check_cancellable() to provide
a GCancellable that is checked for being cancelled during the call to
g_simple_async_result_propagate_error().

This gives asynchronous operation implementations an easy way to
provide reliable cancellation of those operations -- even in the case
that a positive result has occured and is pending dispatch at the time
the operation is cancelled.

https://bugzilla.gnome.org/show_bug.cgi?id=672013

12 years agoInvalidate single_va_closure when adding an emission hook
Mike Gorse [Tue, 13 Mar 2012 21:02:30 +0000 (16:02 -0500)]
Invalidate single_va_closure when adding an emission hook

If the optimization is used for only having one closure handling a
signal emission, then hooks will not be run, so it should be disabled
when an emission hook is added.

https://bugzilla.gnome.org/show_bug.cgi?id=671918

12 years agog_log: mask log level before checking prefix flags
Ryan Lortie [Wed, 14 Mar 2012 00:10:39 +0000 (20:10 -0400)]
g_log: mask log level before checking prefix flags

We check if the log level is in the "prefixed" list by checking it
against the g_log_msg_prefix bitfield.

Unfortunately we were failing to mask by G_LOG_LEVEL_MASK first, so if
the FATAL bit was set (for example) then it would never match.  This was
the case for g_error().

https://bugzilla.gnome.org/show_bug.cgi?id=672026

12 years agoglib-init: restore default for G_MESSAGES_PREFIXED
Ryan Lortie [Wed, 14 Mar 2012 00:08:27 +0000 (20:08 -0400)]
glib-init: restore default for G_MESSAGES_PREFIXED

In the glib-init shuffle this ended up being set to no messages prefixed
by default.  Fix that so that we have prefixes back.

https://bugzilla.gnome.org/show_bug.cgi?id=672026

12 years agoglib-init: fix obvious buffer size mismatch
Ryan Lortie [Wed, 14 Mar 2012 00:08:01 +0000 (20:08 -0400)]
glib-init: fix obvious buffer size mismatch

12 years agoUpdated British English translation
Bruce Cowan [Tue, 13 Mar 2012 13:48:15 +0000 (13:48 +0000)]
Updated British English translation

12 years agoUpdated Danish translation
Kenneth Nielsen [Tue, 13 Mar 2012 08:11:21 +0000 (09:11 +0100)]
Updated Danish translation

12 years agoGSocketMsgFlags: annotate as a flags
Ryan Lortie [Mon, 12 Mar 2012 20:43:15 +0000 (16:43 -0400)]
GSocketMsgFlags: annotate as a flags

This was being incorrectly guessed as an enum.  Fix that.

https://bugzilla.gnome.org/show_bug.cgi?id=671942

12 years agoUpdated Norwegian bokmål translation
Kjartan Maraas [Mon, 12 Mar 2012 17:18:55 +0000 (18:18 +0100)]
Updated Norwegian bokmål translation

12 years agoFinnish translation update from http://l10n.laxstrom.name/wiki/Gnome_3.4 translation...
Timo Jyrinki [Mon, 12 Mar 2012 09:02:04 +0000 (11:02 +0200)]
Finnish translation update from l10n.laxstrom.name/wiki/Gnome_3.4 translation sprint

12 years agoUpdated Esperanto translation
Kristjan SCHMIDT [Sun, 11 Mar 2012 11:02:28 +0000 (12:02 +0100)]
Updated Esperanto translation

12 years agoUpdated Polish translation
Piotr Drąg [Fri, 9 Mar 2012 21:18:45 +0000 (22:18 +0100)]
Updated Polish translation

12 years agologging test: unset G_MESSAGES_DEBUG
Ryan Lortie [Fri, 9 Mar 2012 15:29:32 +0000 (10:29 -0500)]
logging test: unset G_MESSAGES_DEBUG

Having G_MESSAGES_DEBUG=all in the environment will fail this testcase.

12 years agodeprecated threads: fix race in GStaticRecMutex
Mark Janossy [Fri, 9 Mar 2012 14:54:23 +0000 (09:54 -0500)]
deprecated threads: fix race in GStaticRecMutex

The very last access to the 'depth' field of GStaticRecMutex in
g_static_rec_mutex_unlock_full() was being performed after dropping the
implementation mutex for the last time.

This allowed the lock to be dropped an additional time if it was
acquired in another thread right at that instant (which is somewhat
likely, since another thread could have just been woken up by the lock
being released).

https://bugzilla.gnome.org/show_bug.cgi?id=670846

12 years agoFix crash in signal emission optimization
Alexander Larsson [Fri, 9 Mar 2012 10:32:45 +0000 (11:32 +0100)]
Fix crash in signal emission optimization

The trace of the signal emission of the was calculating the instance
type after the instance was unrefed. Fix this by keeping the instance type around.

12 years agosignal emission: Allocate instance_and_param with alloca
Alexander Larsson [Fri, 9 Mar 2012 08:32:06 +0000 (09:32 +0100)]
signal emission: Allocate instance_and_param with alloca

This Gives a 5% performance increase in e.g. the emit-handled test.

12 years agoUse native glib-genmarshal when cross-compiling
Erik van Pienbroek [Thu, 8 Mar 2012 19:29:18 +0000 (20:29 +0100)]
Use native glib-genmarshal when cross-compiling

The ability to cross-compile glib got broken after the
merge of the 'signal-performance' branch as the assumption
was made that the generated glib-genmarshal can be executed
on the host (which isn't valid when cross-compiling).

Fixed this by using the just-built glib-genmarshal for normal
compilations and the native (host) glib-genmarshal when doing a
cross-compilation as was also done in several other areas of GLib

Tested for host=x86_64-unknown-linux-gnu, target=x86_64-unknown-linux-gnu
and host=x86_64-unknown-linux-gnu, target=i686-w64-mingw32

https://bugzilla.gnome.org/show_bug.cgi?id=671676

12 years agogio-querymodules: unlink instead of writing empty cache
Ryan Lortie [Thu, 8 Mar 2012 16:55:09 +0000 (11:55 -0500)]
gio-querymodules: unlink instead of writing empty cache

If there are no modules installed then the most appropriate thing is to
have no cachefile instead of an empty one.  This unbreaks the "clean
directory after 'make uninstall'" check that automake does.

https://bugzilla.gnome.org/show_bug.cgi?id=671664

12 years agomsvc_recommended_pragmas.h: Silence C4819 warnings
Chun-wei Fan [Thu, 8 Mar 2012 09:18:43 +0000 (17:18 +0800)]
msvc_recommended_pragmas.h: Silence C4819 warnings

This warning appears when unicode chars that cannot be displayed in the
current Windows code page is used anywhere in the file, including comment
blocks.  We probably don't need to see these, especially as problems
caused by such characters are manifested as other warnings or errors,
for example, the need to add BOM to a file when compiling code with
complex script on Windows.

12 years agoAssamese translation completed
Nilamdyuti Goswami [Thu, 8 Mar 2012 08:38:11 +0000 (14:08 +0530)]
Assamese translation completed

12 years agoUpdated Traditional Chinese translation(Hong Kong and Taiwan)
Chao-Hsiung Liao [Thu, 8 Mar 2012 07:56:14 +0000 (15:56 +0800)]
Updated Traditional Chinese translation(Hong Kong and Taiwan)

12 years agoUpdate config.h.win32(.in)
Chun-wei Fan [Thu, 8 Mar 2012 07:34:39 +0000 (15:34 +0800)]
Update config.h.win32(.in)

Remove the config for ENABLE_REGEX, as GRegex is now included in all builds.

12 years agoglib-genmarshal.c: Use G_VA_COPY instead of va_copy()
Chun-wei Fan [Thu, 8 Mar 2012 07:04:17 +0000 (15:04 +0800)]
glib-genmarshal.c: Use G_VA_COPY instead of va_copy()

In the code generation portion, va_copy() is not universally available,
so use the existing G_VA_COPY macro that in turn calls va_copy() if it
is available or call an appropriate emulation otherwise.

12 years agoUse G_VA_COPY instead of va_copy()
Chun-wei Fan [Thu, 8 Mar 2012 06:51:22 +0000 (14:51 +0800)]
Use G_VA_COPY instead of va_copy()

va_vopy() is not universally available in all compilers, so make use of
the existing G_VA_COPY macro which either calls va_copy() if it is
available, or emulates it if otherwise.

12 years agogobject: Fix a compiler warning
Emmanuele Bassi [Wed, 7 Mar 2012 12:53:14 +0000 (12:53 +0000)]
gobject: Fix a compiler warning

Shuffle the parenthesis around.

12 years agotests: fix GAction test for headless run
Cosimo Cecchi [Wed, 7 Mar 2012 03:10:53 +0000 (22:10 -0500)]
tests: fix GAction test for headless run

12 years agoAssamese translation completed
Nilamdyuti Goswami [Tue, 6 Mar 2012 16:13:24 +0000 (21:43 +0530)]
Assamese translation completed

12 years agoUpdated Korean translation
Seong-ho Cho [Tue, 6 Mar 2012 14:50:59 +0000 (23:50 +0900)]
Updated Korean translation

12 years agoFix unused variable warning
Alexander Larsson [Tue, 6 Mar 2012 10:35:17 +0000 (11:35 +0100)]
Fix unused variable warning

12 years agoGUnix{Input,Output}Stream: Only release cancelable poll-fd if we allocated it
David Zeuthen [Mon, 5 Mar 2012 17:57:32 +0000 (12:57 -0500)]
GUnix{Input,Output}Stream: Only release cancelable poll-fd if we allocated it

Otherwise we get criticals a'la

 GLib-GIO-CRITICAL **: g_cancellable_release_fd: assertion `cancellable->priv->fd_refcount > 0' failed

when reading/writing to certain kinds of file descriptors.

Patch reviewed by Dan Winship on IRC.

Signed-off-by: David Zeuthen <davidz@redhat.com>
12 years agobump version
Matthias Clasen [Mon, 5 Mar 2012 16:43:20 +0000 (11:43 -0500)]
bump version

12 years agoUse G_DEFINE_CONSTRUCTOR for the glib_init constructor
Alexander Larsson [Mon, 5 Mar 2012 14:05:07 +0000 (15:05 +0100)]
Use G_DEFINE_CONSTRUCTOR for the glib_init constructor

No need to hand-roll this when we have a macro version.

12 years ago2.31.20 2.31.20
Matthias Clasen [Mon, 5 Mar 2012 14:01:20 +0000 (09:01 -0500)]
2.31.20

12 years agoUpdate glib-genmarshal docs
Matthias Clasen [Mon, 5 Mar 2012 13:32:07 +0000 (08:32 -0500)]
Update glib-genmarshal docs

A bunch of new commandline options sneaked in without docs...

12 years agoFix a compiler warning
Matthias Clasen [Mon, 5 Mar 2012 12:54:47 +0000 (07:54 -0500)]
Fix a compiler warning

12 years agogobject: Fix symbol lists
Matthias Clasen [Mon, 5 Mar 2012 12:34:58 +0000 (07:34 -0500)]
gobject: Fix symbol lists

12 years agoMore updates
Matthias Clasen [Mon, 5 Mar 2012 12:02:53 +0000 (07:02 -0500)]
More updates

12 years agoMerge branch 'signal-performance'
Alexander Larsson [Mon, 5 Mar 2012 11:29:45 +0000 (12:29 +0100)]
Merge branch 'signal-performance'

12 years agoFix distcheck
Matthias Clasen [Mon, 5 Mar 2012 04:42:45 +0000 (23:42 -0500)]
Fix distcheck

12 years agoupdate Simplified Chinese (zh_CN) translation
YunQiang Su [Mon, 5 Mar 2012 09:44:58 +0000 (17:44 +0800)]
update Simplified Chinese (zh_CN) translation

12 years agoFix distcheck
Matthias Clasen [Mon, 5 Mar 2012 03:25:50 +0000 (22:25 -0500)]
Fix distcheck

12 years agoUpdates for 2.31.20
Matthias Clasen [Mon, 5 Mar 2012 02:48:26 +0000 (21:48 -0500)]
Updates for 2.31.20

12 years agoFix binding properties of the same object
Matthias Clasen [Mon, 5 Mar 2012 01:49:04 +0000 (20:49 -0500)]
Fix binding properties of the same object

This problem was pointed out in bug 639873.
The patch here is based on Matt Barnes patch,
I've added a testcase as well.

12 years agoRun gdb under libtool so that local libraries are used
Nirbheek Chauhan [Tue, 21 Feb 2012 14:12:22 +0000 (19:42 +0530)]
Run gdb under libtool so that local libraries are used

If this isn't done, the system-wide glib is used for this test,
and that can cause test failures.

https://bugzilla.gnome.org/621368

12 years agoUpdated Slovenian translation
Matej Urbančič [Sun, 4 Mar 2012 20:23:29 +0000 (21:23 +0100)]
Updated Slovenian translation

12 years agoUpdated Slovenian translation
Martin Srebotnjak [Sun, 4 Mar 2012 20:20:59 +0000 (21:20 +0100)]
Updated Slovenian translation

12 years agobuild: include glib-compile-resources.xml in the dist tarball
Michael Biebl [Sat, 3 Mar 2012 23:26:10 +0000 (00:26 +0100)]
build: include glib-compile-resources.xml in the dist tarball

https://bugzilla.gnome.org/show_bug.cgi?id=671281

12 years agobuild: Fix failure on "make distclean"
Michael Biebl [Sat, 3 Mar 2012 20:39:35 +0000 (21:39 +0100)]
build: Fix failure on "make distclean"

https://bugzilla.gnome.org/show_bug.cgi?id=671270

12 years agofastpath: Do allow NO_RECURSE fastpath for NOP emissions
Alexander Larsson [Sat, 3 Mar 2012 20:36:10 +0000 (21:36 +0100)]
fastpath: Do allow NO_RECURSE fastpath for NOP emissions

This fixes a performance regression wrt the old NOP emission
handler.

12 years agoglib/gmacros.h: Fix up attribute to avoid namespace collisions
Shota Akiyama [Thu, 23 Feb 2012 10:52:12 +0000 (11:52 +0100)]
glib/gmacros.h: Fix up attribute to avoid namespace collisions

https://bugzilla.gnome.org/show_bug.cgi?id=670751

12 years agogspawn: rename G_SPAWN_ERROR_2BIG to be more bindings-friendly
Dan Winship [Fri, 2 Mar 2012 15:22:11 +0000 (10:22 -0500)]
gspawn: rename G_SPAWN_ERROR_2BIG to be more bindings-friendly

Rename G_SPAWN_ERROR_2BIG to G_SPAWN_ERROR_TOO_BIG (while keeping the
old name for compatibility), to fix problems with language bindings
where the old name translates into something that would be
syntactically invalid due to starting with a digit.

https://bugzilla.gnome.org/show_bug.cgi?id=671025

12 years agoGApplication: fix name unregister logic
Ryan Lortie [Sat, 3 Mar 2012 16:40:51 +0000 (11:40 -0500)]
GApplication: fix name unregister logic

We were attempting to unregister our ownership of our D-Bus name even in
the case that we were non-unique (ie: we didn't actually own the name).
Rework the logic a bit to prevent that: for non-unique, we leave
impl->bus_name as NULL and we only register/unregister if it is
non-NULL.

12 years agoUpdated Vietnamese translation
Nguyễn Thái Ngọc Duy [Sat, 3 Mar 2012 15:49:18 +0000 (22:49 +0700)]
Updated Vietnamese translation

12 years agopo/vi: import from Damned Lies
Nguyễn Thái Ngọc Duy [Sat, 3 Mar 2012 15:35:16 +0000 (22:35 +0700)]
po/vi: import from Damned Lies

12 years agoUpdated Basque language
Inaki Larranaga Murgoitio [Sat, 3 Mar 2012 13:43:05 +0000 (14:43 +0100)]
Updated Basque language

12 years agoBug 640202 Impossible to build multi-architecture libraries
Peter O'Gorman [Fri, 2 Mar 2012 21:36:27 +0000 (13:36 -0800)]
Bug 640202 Impossible to build multi-architecture libraries

Hides ELF-only linker flag -export-dynamic from non-ELF linkers.

12 years agoMask out G_SIGNAL_TYPE_STATIC_SCOPE when comparing types
Alexander Larsson [Fri, 2 Mar 2012 19:22:29 +0000 (20:22 +0100)]
Mask out G_SIGNAL_TYPE_STATIC_SCOPE when comparing types

12 years agobuild: Fix rules for marshalers.[ch]
Emmanuele Bassi [Fri, 2 Mar 2012 17:06:36 +0000 (17:06 +0000)]
build: Fix rules for marshalers.[ch]

12 years agoAdd more tests for signal emissions
Alexander Larsson [Fri, 2 Mar 2012 15:16:17 +0000 (16:16 +0100)]
Add more tests for signal emissions

12 years agoOptimize single-handler va_marshaller case
Alexander Larsson [Wed, 22 Feb 2012 18:36:05 +0000 (19:36 +0100)]
Optimize single-handler va_marshaller case

When there is only one closure handling a signal emission and
it doesn't have a bunch of complicated features enabled we
can short circuit the va_args collection into GValues and call the
callback via the va_marshaller directly.

https://bugzilla.gnome.org/show_bug.cgi?id=661140

12 years agoUse builtin marshallers and va_marshallers if possible
Alexander Larsson [Fri, 2 Mar 2012 10:20:45 +0000 (11:20 +0100)]
Use builtin marshallers and va_marshallers if possible

If the signal argumment types matches a built in standard
marshaller we use the va_marshaller for that, and also the
normal marshaller if NULL was specified (as its faster than
the generic one).

12 years agoAdd g_signal_set_va_marshaller
Alexander Larsson [Wed, 22 Feb 2012 18:34:05 +0000 (19:34 +0100)]
Add g_signal_set_va_marshaller

This lets you set a va_marshaller on your signal which will be
propagated to all closures for the signal. Also, automatically
uses the generica va_marshaller if you specify a NULL c_marshaller.

https://bugzilla.gnome.org/show_bug.cgi?id=661140

12 years agoAdd _g_closure_is_void to check for NULL vfuncs
Alexander Larsson [Tue, 28 Feb 2012 14:48:20 +0000 (15:48 +0100)]
Add _g_closure_is_void to check for NULL vfuncs

https://bugzilla.gnome.org/show_bug.cgi?id=661140

12 years agoSupport generating va marshallers in glib-genmarshal
Alexander Larsson [Wed, 22 Feb 2012 18:32:36 +0000 (19:32 +0100)]
Support generating va marshallers in glib-genmarshal

https://bugzilla.gnome.org/show_bug.cgi?id=661140

12 years agoAdd optional support for varargs marshallers to GClosure
Alexander Larsson [Wed, 22 Feb 2012 18:29:14 +0000 (19:29 +0100)]
Add optional support for varargs marshallers to GClosure

These closures support being invoked on a va_args which can
be useful as you can then avoid boxing the va_args into
GValues in certain cases.

https://bugzilla.gnome.org/show_bug.cgi?id=661140

12 years agoAdd GRealClosure and move meta_marshallers there
Alexander Larsson [Wed, 22 Feb 2012 16:46:21 +0000 (17:46 +0100)]
Add GRealClosure and move meta_marshallers there

This means we're not abusing the notifiers for meta_marshallres,
and we're able to later cleanly add other fields to GClosure.

We still have to leave the ABI intact for the GClosure->meta_marshal
bit, as old G_CLOSURE_N_NOTIFIERS macro instances still accesses it.
However, we always set it to zero to keep those macros working.

https://bugzilla.gnome.org/show_bug.cgi?id=661140

12 years agotests: Add generic and empty signal emission performace tests
Alexander Larsson [Wed, 22 Feb 2012 18:44:24 +0000 (19:44 +0100)]
tests: Add generic and empty signal emission performace tests

generic means it uses the generic marshaller
empty means the vfunc pointer is NULL

https://bugzilla.gnome.org/show_bug.cgi?id=661140

12 years agotests: performance: add emit-handled/emit-unhandled tests
Lionel Landwerlin [Tue, 27 Sep 2011 08:15:17 +0000 (09:15 +0100)]
tests: performance: add emit-handled/emit-unhandled tests

Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@linux.intel.com>
12 years agoGSequence: add note about lookup requiring sorting
Jörn Magens [Wed, 29 Feb 2012 17:29:50 +0000 (12:29 -0500)]
GSequence: add note about lookup requiring sorting

g_sequence_lookup() only works on sorted sequences, but it's quite easy
to create unsorted sequences.  Add a note to the documentation that the
sequence must be sorted in order for g_sequence_lookup() to work.

https://bugzilla.gnome.org/show_bug.cgi?id=670969

12 years agoUpdated Hebrew translation.
Yaron Shahrabani [Fri, 2 Mar 2012 12:15:58 +0000 (14:15 +0200)]
Updated Hebrew translation.

12 years agogobject/gtype.c: Emphasize use of priv in member and use g_return_val_if_fail
Christian Dywan [Sat, 26 Mar 2011 14:48:54 +0000 (14:48 +0000)]
gobject/gtype.c: Emphasize use of priv in member and use g_return_val_if_fail

Fixes: https://bugzilla.gnome.org/show_bug.cgi?id=597785

Signed-off-by: Javier Jardón <jjardon@gnome.org>
12 years agoAdd G_VOLUME_IDENTIFIER_KIND_CLASS identifier
David Zeuthen [Thu, 1 Mar 2012 19:06:43 +0000 (14:06 -0500)]
Add G_VOLUME_IDENTIFIER_KIND_CLASS identifier

If an application (such as Nautilus) wants to show a sidebar with
devices group into different groups such as "Devices" and "Network",
it's currently up to the application itself to do the classification
(for example by looking at the URI scheme for the activation root,
e.g. smb://).

This patch adds a new identifier G_VOLUME_IDENTIFIER_KIND_CLASS that
can be set by volume monitors and used by applications.

See https://bugzilla.gnome.org/show_bug.cgi?id=668295

Signed-off-by: David Zeuthen <davidz@redhat.com>
12 years agoUpdated Telugu Translations
Krishnababu Krothapalli [Thu, 1 Mar 2012 08:29:41 +0000 (13:59 +0530)]
Updated Telugu Translations

12 years agog_uri_unescape_segment: Mention (allow-none)/%NULL in documentation
Colin Walters [Wed, 29 Feb 2012 17:37:24 +0000 (12:37 -0500)]
g_uri_unescape_segment: Mention (allow-none)/%NULL in documentation

12 years agoVersion the deprecation of g_value_array_get_type()
Emmanuele Bassi [Wed, 29 Feb 2012 17:52:39 +0000 (17:52 +0000)]
Version the deprecation of g_value_array_get_type()

Just like the rest of the GValueArray API.

12 years agogsocket: fix compilation when ENOTSOCK and WSANOTSOCK are both defined
Andoni Morales Alastruey [Wed, 29 Feb 2012 10:19:21 +0000 (11:19 +0100)]
gsocket: fix compilation when ENOTSOCK and WSANOTSOCK are both defined

12 years agoUpdate Visual C++ projects
Chun-wei Fan [Wed, 29 Feb 2012 03:58:16 +0000 (11:58 +0800)]
Update Visual C++ projects

Update the build support of the included PCRE as we are now including
PCRE 8.30 with the GLib distribution.

Also "install" the new gversionmacros.h header file.

12 years agogio-2.0.pc: add gdbus_codegen variable
Aleksander Morgado [Tue, 28 Feb 2012 00:14:29 +0000 (01:14 +0100)]
gio-2.0.pc: add gdbus_codegen variable

https://bugzilla.gnome.org/show_bug.cgi?id=670922

12 years agoUpdated Galician translations
Fran Diéguez [Mon, 27 Feb 2012 23:32:18 +0000 (00:32 +0100)]
Updated Galician translations

12 years agoGDBusConnection: Don't leak message reply in error path
David Zeuthen [Mon, 27 Feb 2012 19:53:13 +0000 (14:53 -0500)]
GDBusConnection: Don't leak message reply in error path

This was reported in bug 670909.

https://bugzilla.gnome.org/show_bug.cgi?id=670909

Signed-off-by: David Zeuthen <davidz@redhat.com>
12 years agoGDBusMenuModel: plug a leak of a GHashTable
Ryan Lortie [Mon, 27 Feb 2012 19:00:55 +0000 (14:00 -0500)]
GDBusMenuModel: plug a leak of a GHashTable

12 years agoFixup docs comment
Christian Persch [Mon, 27 Feb 2012 15:25:50 +0000 (16:25 +0100)]
Fixup docs comment

12 years agoregex: unicode: Update to Unicode 6.1.0
Christian Persch [Mon, 13 Feb 2012 13:20:55 +0000 (14:20 +0100)]
regex: unicode: Update to Unicode 6.1.0

Update some more PCRE tables to Unicode 6.1.0. Should be completely
updated now.

12 years agoFix doc build again
Matthias Clasen [Mon, 27 Feb 2012 12:16:11 +0000 (07:16 -0500)]
Fix doc build again

The way gtk-doc works, we need compiling.sgml in both
content_files (to make sure it ends up in the disted tarball)
and in expand_content_files (to have references expanded).

12 years agoImprove the wording of the message for G_UNAVAILABLE
Matthias Clasen [Mon, 27 Feb 2012 11:48:01 +0000 (06:48 -0500)]
Improve the wording of the message for G_UNAVAILABLE

The version number we have here is the one the function appeared
in, not the max-allowed version. Therefore 'unavailable before
VERSION' makes more sense than 'unavailable for VERSION'.

12 years agodocs/glib: Moving back compiling.sgml to content_files to fix make dist
Rico Tzschichholz [Mon, 27 Feb 2012 11:50:18 +0000 (12:50 +0100)]
docs/glib: Moving back compiling.sgml to content_files to fix make dist

12 years agoUpdated Bulgarian translation
Alexander Shopov [Mon, 27 Feb 2012 05:31:44 +0000 (07:31 +0200)]
Updated Bulgarian translation

12 years agoAdd a 'these are private' note for the version macros
Matthias Clasen [Mon, 27 Feb 2012 05:32:13 +0000 (00:32 -0500)]
Add a 'these are private' note for the version macros

12 years agoImprove documentation around deprecations
Matthias Clasen [Mon, 27 Feb 2012 05:18:47 +0000 (00:18 -0500)]
Improve documentation around deprecations

12 years agoAdd versioned deprecation annotation
Emmanuele Bassi [Tue, 21 Feb 2012 13:40:50 +0000 (13:40 +0000)]
Add versioned deprecation annotation

We start from GLib 2.26.

https://bugzilla.gnome.org/show_bug.cgi?id=670542

12 years agovalue array: Annotate with versioned deprecation
Emmanuele Bassi [Mon, 20 Feb 2012 16:41:30 +0000 (16:41 +0000)]
value array: Annotate with versioned deprecation

https://bugzilla.gnome.org/show_bug.cgi?id=670542

12 years agoAdd flexible API version boundaries
Emmanuele Bassi [Mon, 20 Feb 2012 16:20:15 +0000 (16:20 +0000)]
Add flexible API version boundaries

There are cases when it should be possible to define at compile time
what range of functions and types should be used, in order to get,
or restrict, the compiler warnings for deprecated or newly added
types or functions.

For instance, if GLib introduces a deprecation warning on a type in
version 2.32, application code can decide to specify the minimum and
maximum boundary of the used API to be 2.30; when compiling against
a new version of GLib, this would produce the following results:

  - all deprecations introduced prior to 2.32 would emit compiler
    warnings when used by the application code;
  - all deprecations introduced in 2.32 would not emit compiler
    warnings when used by the application code;
  - all new symbols introduced in 2.32 would emit a compiler warning.

Using this scheme it should be possible to have fairly complex
situations, like the following one:

  assuming that an application is compiled with:
    GLIB_VERSION_MIN_REQUIRED = GLIB_VERSION_2_30
    GLIB_VERSION_MAX_ALLOWED  = GLIB_VERSION_2_32

  and a GLib header containing:

    void function_A (void) GLIB_DEPRECATED_IN_2_26;
    void function_B (void) GLIB_DEPRECATED_IN_2_28;
    void function_C (void) GLIB_DEPRECATED_IN_2_30;
    void function_D (void) GLIB_AVAILABLE_IN_2_32;
    void function_E (void) GLIB_AVAILABLE_IN_2_34;

  any application code using the above functions will get the following
  compiler warnings:

    function_A: deprecated symbol warning
    function_B: deprecated symbol warning
    function_C: no warning
    function_D: no warning
    function_E: undefined symbol warning

This means that it should be possible to gradually port code towards
non-deprecated API gradually, on a per-release basis.

https://bugzilla.gnome.org/show_bug.cgi?id=670542

12 years agoregex: Remove --disable-regex option
Christian Persch [Sat, 11 Feb 2012 22:51:11 +0000 (23:51 +0100)]
regex: Remove --disable-regex option

https://bugzilla.gnome.org/show_bug.cgi?id=622149

12 years agodocs: Clarify how to unset a GFileInfo attribute
David King [Tue, 21 Feb 2012 11:19:20 +0000 (11:19 +0000)]
docs: Clarify how to unset a GFileInfo attribute

https://bugzilla.gnome.org/show_bug.cgi?id=592666

12 years agodocs: Clarify g_type_register_fundamental() behaviour
David King [Tue, 21 Feb 2012 10:05:08 +0000 (10:05 +0000)]
docs: Clarify g_type_register_fundamental() behaviour

https://bugzilla.gnome.org/show_bug.cgi?id=580873

12 years agoregex: Remove pcre_ucd.c
Christian Persch [Sun, 12 Feb 2012 20:22:13 +0000 (21:22 +0100)]
regex: Remove pcre_ucd.c

12 years agoregex: Use glib for unicode data
Christian Persch [Sun, 12 Feb 2012 20:20:33 +0000 (21:20 +0100)]
regex: Use glib for unicode data

Use g_unichar_type() and g_unichar_get_script() instead of pcre tables.