platform/upstream/glib.git
13 years agoAdd glib credentials support to OpenBSD.
Antoine Jacoutot [Fri, 27 May 2011 13:51:08 +0000 (15:51 +0200)]
Add glib credentials support to OpenBSD.

https://bugzilla.gnome.org/show_bug.cgi?id=650885

13 years agoDon't hardcode path to true(1).
Antoine Jacoutot [Fri, 27 May 2011 15:39:38 +0000 (17:39 +0200)]
Don't hardcode path to true(1).

true(1) is not always installed under /bin so leave system() find
it in the PATH.

https://bugzilla.gnome.org/show_bug.cgi?id=651219

13 years agoPreserve consistency when removing all nodes from a hash table
Matthias Clasen [Fri, 27 May 2011 01:52:50 +0000 (21:52 -0400)]
Preserve consistency when removing all nodes from a hash table

During the recent refactorings of GHashTable a bug was introduced
where removing all nodes from a hash table would leave tombstones
behind, but make the counts appear like there are none.

Reported and tracked down by Carlos Garnacho,
https://bugzilla.gnome.org/show_bug.cgi?id=651141

This commit also adds a test that checks the internal consistency
of GHashTable over several insert/remove/remove-all operations.

13 years agomain: Use public function in documentation
David Schleef [Tue, 24 May 2011 21:35:20 +0000 (14:35 -0700)]
main: Use public function in documentation

https://bugzilla.gnome.org/show_bug.cgi?id=651009

13 years agoGDBusProxy: Mention gdbus-codegen in docs
David Zeuthen [Thu, 26 May 2011 14:18:44 +0000 (10:18 -0400)]
GDBusProxy: Mention gdbus-codegen in docs

Signed-off-by: David Zeuthen <davidz@redhat.com>
13 years agoGDBusProxy: Fix incorrect locking
David Zeuthen [Thu, 26 May 2011 14:07:38 +0000 (10:07 -0400)]
GDBusProxy: Fix incorrect locking

Pointed out by mclasen.

Signed-off-by: David Zeuthen <davidz@redhat.com>
13 years agoGDBusProxy: Add locking and notes/guarantees about MT safety
David Zeuthen [Thu, 26 May 2011 13:54:47 +0000 (09:54 -0400)]
GDBusProxy: Add locking and notes/guarantees about MT safety

This was discussed in

 https://bugzilla.gnome.org/show_bug.cgi?id=651133

Signed-off-by: David Zeuthen <davidz@redhat.com>
13 years agoGDBusProxy: Fix race condition when unsubscribing from signals
David Zeuthen [Thu, 26 May 2011 13:26:29 +0000 (09:26 -0400)]
GDBusProxy: Fix race condition when unsubscribing from signals

This was reported in bug 651133.

https://bugzilla.gnome.org/show_bug.cgi?id=651133

Signed-off-by: David Zeuthen <davidz@redhat.com>
13 years agoCorrect the annotation for g_file_enumerator_get_container()
Richard Hughes [Thu, 26 May 2011 11:37:24 +0000 (12:37 +0100)]
Correct the annotation for g_file_enumerator_get_container()

We don't ref the returned object, and alex has verified the gvfs implementation.

13 years agogthread: build unix tests only on unix
Marc-André Lureau [Wed, 11 May 2011 14:42:14 +0000 (16:42 +0200)]
gthread: build unix tests only on unix

The build fails when compiling with mingw32. Correct the build of a
unix tests on unix only.

https://bugzilla.gnome.org/show_bug.cgi?id=649973

13 years agogdbus-codegen: Fix up example to not use non-existant method
David Zeuthen [Tue, 24 May 2011 11:34:30 +0000 (07:34 -0400)]
gdbus-codegen: Fix up example to not use non-existant method

Signed-off-by: David Zeuthen <davidz@redhat.com>
13 years agogdbus-codegen: Properly render @param and %TRUE in generated Docbook
David Zeuthen [Tue, 24 May 2011 04:27:43 +0000 (00:27 -0400)]
gdbus-codegen: Properly render @param and %TRUE in generated Docbook

Signed-off-by: David Zeuthen <davidz@redhat.com>
13 years agogdbus-codegen: Handle unexpected XML tags
David Zeuthen [Tue, 24 May 2011 03:22:04 +0000 (23:22 -0400)]
gdbus-codegen: Handle unexpected XML tags

This was reported in bug 650874. Add tests.

https://bugzilla.gnome.org/show_bug.cgi?id=650874

Signed-off-by: David Zeuthen <davidz@redhat.com>
13 years agogdbusintrospection: fix introspection of DBus Introspection structures
Giovanni Campagna [Thu, 21 Apr 2011 21:45:59 +0000 (23:45 +0200)]
gdbusintrospection: fix introspection of DBus Introspection structures

Correctly mark fields as arrays (requires changing gobject-introspection
to pick those).

https://bugzilla.gnome.org/show_bug.cgi?id=646635

13 years agoRevert "gsettings-tool: warn if setting a value fails"
Ray Strode [Mon, 23 May 2011 17:39:09 +0000 (13:39 -0400)]
Revert "gsettings-tool: warn if setting a value fails"

This reverts commit ea57feff96f13bbd4d03a76040a4ddfad2677310.

It makes the gsettings tool fail any time it tries to set a
key to a value, that the key already has.

https://bugzilla.gnome.org/show_bug.cgi?id=641768

13 years agoFix build with gcc-2.*.
Antoine Jacoutot [Mon, 23 May 2011 17:11:10 +0000 (19:11 +0200)]
Fix build with gcc-2.*.

https://bugzilla.gnome.org/show_bug.cgi?id=650884

13 years agoUpdated Irish translation.
Seán de Búrca [Mon, 23 May 2011 06:35:46 +0000 (00:35 -0600)]
Updated Irish translation.

13 years agoCosmetic changes
Matthias Clasen [Mon, 23 May 2011 04:48:10 +0000 (00:48 -0400)]
Cosmetic changes

Use inc/dec for refcounting.

13 years agoCosmetics
Matthias Clasen [Mon, 23 May 2011 04:40:33 +0000 (00:40 -0400)]
Cosmetics

Use g_atomic_int_inc/dec instead of add(...,1/-1), since
this is the way refcounting is done elsewhere. Some other
cosmetic changes.

13 years agoSimply symbols files
Matthias Clasen [Mon, 23 May 2011 04:21:06 +0000 (00:21 -0400)]
Simply symbols files

The grouping in files/headers is not used anymore, and
the function attributes neither. Adapt abicheck scripts
and .def file generation rules accordingly.

13 years agoImprove atomic ops implementation
Matthias Clasen [Sun, 22 May 2011 20:31:43 +0000 (16:31 -0400)]
Improve atomic ops implementation

When using gcc builtins for atomic operations, provide them
as macros, so gcc can see the builtins and do optimizations.
This change gives considerable speedups in bitlocks, which
use atomic operations heavily, see bug 650458.

Also, don't define G_ATOMIC_OP_MEMORY_BARRIER_NEEDED unconditionally
when using gcc builtins.

https://bugzilla.gnome.org/show_bug.cgi?id=617491

13 years agoAdd some atomic ops tests
Matthias Clasen [Sun, 22 May 2011 05:29:22 +0000 (01:29 -0400)]
Add some atomic ops tests

13 years agoFix HAVE_FUTEX definition
Matthias Clasen [Sun, 22 May 2011 04:33:05 +0000 (00:33 -0400)]
Fix HAVE_FUTEX definition

13 years agoTrivial: fix a typo
Matthias Clasen [Sun, 22 May 2011 00:56:04 +0000 (20:56 -0400)]
Trivial: fix a typo

13 years agoRevert "g_key_file_has_key_full: Clarify the docs"
Matthias Clasen [Sat, 21 May 2011 15:14:05 +0000 (11:14 -0400)]
Revert "g_key_file_has_key_full: Clarify the docs"

This reverts commit 4ddb84999d47816624f9ae7dcba181cbccd5c4d9.

13 years agog_key_file_has_key_full: Clarify the docs
Matthias Clasen [Sat, 21 May 2011 15:12:08 +0000 (11:12 -0400)]
g_key_file_has_key_full: Clarify the docs

13 years agoEnforce rules about modifying hash tables in callbacks
Matthias Clasen [Fri, 20 May 2011 19:07:08 +0000 (15:07 -0400)]
Enforce rules about modifying hash tables in callbacks

We have the infrastructure to do this, so lets do it.
Also add tests for find and foreach to the testsuite.
Bug 650688

13 years agoBe more careful when calling destroy notifies
Matthias Clasen [Fri, 20 May 2011 03:50:03 +0000 (23:50 -0400)]
Be more careful when calling destroy notifies

If we are, we can allow modification of the hash table
from destroy notifies.

https://bugzilla.gnome.org/show_bug.cgi?id=650459

13 years agoGObject: move GLib type definitions to a separate header
Giovanni Campagna [Sun, 3 Apr 2011 13:51:04 +0000 (15:51 +0200)]
GObject: move GLib type definitions to a separate header

This way it is possible to pull them into the GLib GIR file.

Some cleanups by Colin Walters <walters@verbum.org>

https://bugzilla.gnome.org/show_bug.cgi?id=646635

13 years agoWhitespace fixes
Ryan Lortie [Wed, 18 May 2011 21:56:50 +0000 (17:56 -0400)]
Whitespace fixes

13 years agoGApplication: fix remote action states
Ryan Lortie [Wed, 18 May 2011 21:53:07 +0000 (17:53 -0400)]
GApplication: fix remote action states

Fix up remote_action_info_new_from_iter() to do its job better and use
it from all places that it's appropriate.

Closes #650236.

13 years agoAdd debug output to GApplication actions example
Ryan Lortie [Wed, 18 May 2011 20:12:04 +0000 (16:12 -0400)]
Add debug output to GApplication actions example

Enough to clearly demonstrate bug #650236.

13 years agogatomic: #include gthread.h to fix compilation on nonoptimized architectures
Jasper Lievisse Adriaanse [Wed, 18 May 2011 16:01:44 +0000 (12:01 -0400)]
gatomic: #include gthread.h to fix compilation on nonoptimized architectures

The fallback case of implementing atomic integers with mutexes needs
gthread.h.

Commit message written by Colin Walters <walters@verbum.org>

https://bugzilla.gnome.org/show_bug.cgi?id=631153

13 years agogtype: Remove unused variable
Colin Walters [Wed, 18 May 2011 15:57:21 +0000 (11:57 -0400)]
gtype: Remove unused variable

13 years agog_key_file_has_key_full: New function to fix g_key_file_has_key()'s GError semantics
Colin Walters [Mon, 16 May 2011 19:30:31 +0000 (15:30 -0400)]
g_key_file_has_key_full: New function to fix g_key_file_has_key()'s GError semantics

See https://bugzilla.gnome.org/show_bug.cgi?id=649657 for discussion
of why it's bad for bindings for gerror return values to both signal
errors and carry meaning.

https://bugzilla.gnome.org/show_bug.cgi?id=650345

13 years agoFix short day names in Hungarian translation
Gabor Kelemen [Wed, 18 May 2011 10:39:31 +0000 (12:39 +0200)]
Fix short day names in Hungarian translation

13 years agoUpdate Visual C++ README.txt
Chun-wei Fan [Wed, 18 May 2011 03:14:00 +0000 (11:14 +0800)]
Update Visual C++ README.txt

Correct the unpacking path of GLib...

13 years agoFix gsettings tool string handling
Ryan Lortie [Tue, 17 May 2011 15:58:46 +0000 (11:58 -0400)]
Fix gsettings tool string handling

There are some bugs caused by the way that gsettings-tool currently
attempts to help the user when they leave the quotes off of a string
value that they are setting.

Simplify the code to make it more robust and add some comments about why
it should be done this way.

https://bugzilla.gnome.org/show_bug.cgi?id=649915

13 years agogobject: Speed up g_type_from_name()
Benjamin Otte [Tue, 17 May 2011 12:58:39 +0000 (14:58 +0200)]
gobject: Speed up g_type_from_name()

The hash table used exclusively for looking up types by name used to map
quarks => types. But we can easily make it map strings => types, which
avoids the quark lookup. And that in trun avoids taking a lock and
consulting another hash table. So this change should make
g_type_from_name() roughly twice as fast.

13 years agokeyfile: fill parse buffer in line sized chunks
John Lindgren [Tue, 17 May 2011 03:03:30 +0000 (23:03 -0400)]
keyfile: fill parse buffer in line sized chunks

When loading a keyfile the incoming bytes are fed
to a line buffer to get parsed each time a new line
is encountered.

The code that fills the line buffer does it inefficiently,
one byte at a time.

This commit changes that code to look ahead at the incoming
bytes for the next '\n' character and then fill the line buffer
all at once.

https://bugzilla.gnome.org/show_bug.cgi?id=650211

13 years agokeyfile: avoid needless allocations on file load
John Lindgren [Tue, 17 May 2011 02:27:11 +0000 (22:27 -0400)]
keyfile: avoid needless allocations on file load

When loading a key file, the keys and values of individual lines
are allocated once when copied and trimmed from the parse buffer
and allocated/copied again when added to the lookup map.

This commit avoids the second pair of allocations by introducing
a new function g_key_file_add_key_value_pair that gives the
lookup map direct ownership of the key and value copied from the
parse buffer.

https://bugzilla.gnome.org/show_bug.cgi?id=650211

13 years agoRename a problematic test binary
Matthias Clasen [Mon, 16 May 2011 18:38:45 +0000 (14:38 -0400)]
Rename a problematic test binary

Having a binary called printf is apparently asking for trouble,
so lets not do that.

http://bugzilla.gnome.org/show_bug.cgi?id=650078

13 years agoAdd gdbus autogenerated xml files to the ignore file
Emmanuele Bassi [Mon, 16 May 2011 15:10:06 +0000 (16:10 +0100)]
Add gdbus autogenerated xml files to the ignore file

13 years agom4macros/gsettings.m4: Use AS_HELP_STRING instead the deprecated AC_HELP_STRING
Javier Jardón [Mon, 16 May 2011 11:51:59 +0000 (12:51 +0100)]
m4macros/gsettings.m4: Use AS_HELP_STRING instead the deprecated AC_HELP_STRING

13 years agoautogen.sh: Use autoreconf instead custom script
Javier Jardón [Mon, 16 May 2011 10:51:29 +0000 (11:51 +0100)]
autogen.sh: Use autoreconf instead custom script

13 years agogdbus-codegen: Ensure that generated skeletons are MT-safe
David Zeuthen [Sun, 15 May 2011 15:45:37 +0000 (11:45 -0400)]
gdbus-codegen: Ensure that generated skeletons are MT-safe

For example, if setting a property on a skeleton from another thread
than where it was constructed, the idle handler responsible for
emitting the PropertiesChanged() signal could run immediately and
clear skeleton->priv->changed_properties_idle_source causing
g_source_unref() to be called with a NULL pointer. This race was
easily be fixed by adding a lock to the skeleton object.

In addition to fixing this race, also move the code for setting up the
idle handler to a class handler for the GObject::notify signal. This
change allows use of g_object_freeze_notify() and g_object_thaw_notify()
to perform atomic property changes from another thread than the one
that the skeleton was created in.

Signed-off-by: David Zeuthen <davidz@redhat.com>
13 years agodocs: Improve punctuation in some of the GArray method documentation
Philip Withnall [Sun, 15 May 2011 14:31:16 +0000 (15:31 +0100)]
docs: Improve punctuation in some of the GArray method documentation

13 years agodocs: Add the cute trick comment to all sort functions
Benjamin Otte [Sun, 15 May 2011 14:16:30 +0000 (16:16 +0200)]
docs: Add the cute trick comment to all sort functions

Thanks to Sebastian Dröge for pointing out there's more than
g_array_sort() and I hadn't in fact added the comment to g_array_sort()
in 80928ea403cfff5ffaa4bc73e84e3d4ebcb09123

13 years agodocs: Add a cute trick for achieving a stable sort
Benjamin Otte [Sun, 15 May 2011 14:11:51 +0000 (16:11 +0200)]
docs: Add a cute trick for achieving a stable sort

This trick is inspired by
http://www.gnu.org/s/hello/manual/libc/Array-Sort-Function.html
http://jeffreystedfast.blogspot.com/2011/04/optimizing-merge-sort.html

13 years agoUpdated Esperanto translation
Kristjan SCHMIDT [Sun, 15 May 2011 08:16:59 +0000 (10:16 +0200)]
Updated Esperanto translation

13 years agogdbus-codegen: Use (skip) annotation on return values
David Zeuthen [Fri, 13 May 2011 18:23:14 +0000 (14:23 -0400)]
gdbus-codegen: Use (skip) annotation on return values

This was discussed in bug 649657.

https://bugzilla.gnome.org/show_bug.cgi?id=649657

Signed-off-by: David Zeuthen <davidz@redhat.com>
13 years agoUpdated Hebrew translation.
Yaron Shahrabani [Fri, 13 May 2011 05:38:35 +0000 (08:38 +0300)]
Updated Hebrew translation.

13 years agoGDBusConnection: Document that sockets are frobbed into non-blocking mode
David Zeuthen [Thu, 12 May 2011 19:36:34 +0000 (15:36 -0400)]
GDBusConnection: Document that sockets are frobbed into non-blocking mode

Signed-off-by: David Zeuthen <davidz@redhat.com>
13 years agogdbus-codegen: Mark members in generated class instance structs as private
David Zeuthen [Thu, 12 May 2011 01:16:41 +0000 (21:16 -0400)]
gdbus-codegen: Mark members in generated class instance structs as private

Signed-off-by: David Zeuthen <davidz@redhat.com>
13 years agoAdd g_dbus_method_invocation_take_error() convenience method
David Zeuthen [Thu, 12 May 2011 00:57:54 +0000 (20:57 -0400)]
Add g_dbus_method_invocation_take_error() convenience method

Similar in spirit to g_simple_async_result_take_error().

Signed-off-by: David Zeuthen <davidz@redhat.com>
13 years agogdbus-codegen: Fix annotation parsing
Colin Walters [Wed, 11 May 2011 21:13:57 +0000 (17:13 -0400)]
gdbus-codegen: Fix annotation parsing

It takes 3 options.

13 years agogdbus-codegen: Quote substitution variables
Maarten Bosmans [Wed, 11 May 2011 19:47:00 +0000 (15:47 -0400)]
gdbus-codegen: Quote substitution variables

Avoids failure if they contain spaces.

Commit message written by Colin Walters <walters@verbum.org>

https://bugzilla.gnome.org/show_bug.cgi?id=649775

13 years agogdbus-codegen: Drop dependency on argparse
Colin Walters [Wed, 11 May 2011 19:02:23 +0000 (15:02 -0400)]
gdbus-codegen: Drop dependency on argparse

This is only in Python 2.7 or as a separate module.  Use optparse
which exists since 2.3.

https://bugzilla.gnome.org/show_bug.cgi?id=649988

13 years agogbase64: fix introspection annotations
Giovanni Campagna [Sun, 3 Apr 2011 15:19:48 +0000 (17:19 +0200)]
gbase64: fix introspection annotations

Makes the g_base64_* functions usable from introspection GLib bindings
(gjs, currently, as both vala and pygobject use manual bindings for
GLib)

https://bugzilla.gnome.org/show_bug.cgi?id=646635

13 years agogdatetime: Quiet GCC warning about possibly-uninitialized
Colin Walters [Wed, 11 May 2011 17:19:05 +0000 (13:19 -0400)]
gdatetime: Quiet GCC warning about possibly-uninitialized

13 years agoGSettings: fix example in the docs
Ryan Lortie [Tue, 10 May 2011 13:25:54 +0000 (15:25 +0200)]
GSettings: fix example in the docs

13 years agoGDBusMethodInvocation: fix introspection annotations
Giovanni Campagna [Tue, 3 May 2011 16:54:45 +0000 (18:54 +0200)]
GDBusMethodInvocation: fix introspection annotations

In particular (transfer full), to avoid a refcounting bug and a
segmentation fault.

https://bugzilla.gnome.org/show_bug.cgi?id=646635

13 years ago[gi] Add missing allow-none annotations to GDBusConnection
Martin Pitt [Sun, 8 May 2011 12:12:23 +0000 (14:12 +0200)]
[gi] Add missing allow-none annotations to GDBusConnection

13 years agoUpdated Russian translation
Yuri Kozlov [Sun, 20 Mar 2011 07:50:35 +0000 (10:50 +0300)]
Updated Russian translation

13 years agoUpdated Spanish translation
Daniel Mustieles [Sat, 7 May 2011 10:38:56 +0000 (12:38 +0200)]
Updated Spanish translation

13 years agoGSettings: keep resets on loss of writability
Ryan Lortie [Sat, 7 May 2011 09:42:46 +0000 (11:42 +0200)]
GSettings: keep resets on loss of writability

Don't drop resets from a GDelayedSettingsBackend when the writability
changes.  Resets will always succeed, even against non-writable keys and
some people (gnome-screensaver) are using them in a way that they would
want them not to be forgotten.

13 years agoAllow NULL to be stored into GSettings trees
Ryan Lortie [Sat, 7 May 2011 09:41:19 +0000 (11:41 +0200)]
Allow NULL to be stored into GSettings trees

This is used for resets (in theory) and currently it crashes when
freeing the tree.

13 years agogdbus: Add a --timeout option
David Zeuthen [Fri, 6 May 2011 19:25:49 +0000 (15:25 -0400)]
gdbus: Add a --timeout option

Signed-off-by: David Zeuthen <davidz@redhat.com>
13 years agogdbus-codegen: fix typo in man page
Ray Strode [Fri, 6 May 2011 17:00:38 +0000 (13:00 -0400)]
gdbus-codegen: fix typo in man page

s/on_handle_hello-world/on_handle_hello_world/

13 years agogdbus-codegen: gracefully handle property get on proxy with no cached property
David Zeuthen [Fri, 6 May 2011 14:32:42 +0000 (10:32 -0400)]
gdbus-codegen: gracefully handle property get on proxy with no cached property

This can easily happen if the owner of the remote object vanishes. Of
course, when that happens, user code is already notified (by e.g. the
notify::g-name-owner signal) so it can avoid using the proxy but
requiring that is a bit harsh. IOW, before this patch this critical
error was printed

 GLib-GIO-CRITICAL **: g_dbus_gvariant_to_gvalue: assertion `value != NULL' failed

when that happened. With this patch, we just avoid setting the GValue
so the user will get the default value for its type instead. So, for
example, if the user code is getting a GVariant property on such a
defunct proxy, then he gets a NULL back. So unless said user code
checks the return value, criticals will still be printed if the NULL
GVariant is used for anything interesting.

Signed-off-by: David Zeuthen <davidz@redhat.com>
13 years agoDocumentation additions
Matthias Clasen [Fri, 6 May 2011 11:14:48 +0000 (07:14 -0400)]
Documentation additions

Add Since tags to recently introduced test function typedefs.
https://bugzilla.gnome.org/show_bug.cgi?id=649506

13 years agoAdd two annotations for g-i
Dan Winship [Thu, 5 May 2011 19:18:22 +0000 (15:18 -0400)]
Add two annotations for g-i

13 years agoAdd tests for new GDateTime functionality
Matthias Clasen [Thu, 5 May 2011 18:52:19 +0000 (14:52 -0400)]
Add tests for new GDateTime functionality

13 years agoExpand the g_date_time_format documentation
Matthias Clasen [Thu, 5 May 2011 17:40:51 +0000 (13:40 -0400)]
Expand the g_date_time_format documentation

Mention strftime, and explain supported modifiers.

13 years agoGDateTime: enhance g_date_time_format()
Matthias Clasen [Thu, 5 May 2011 17:16:30 +0000 (13:16 -0400)]
GDateTime: enhance g_date_time_format()

Make g_date_time_format() support some useful format modifiers from
strftime(). This commit adds support for the POSIX 'O' modifier
(alternative digits), as well as the -/_/0 padding modifiers, which
are a GNU strftime() extension.
https://bugzilla.gnome.org/show_bug.cgi?id=648678

13 years agoGSocket: Use MSG_CMSG_CLOEXEC
Colin Walters [Thu, 5 May 2011 17:16:54 +0000 (13:16 -0400)]
GSocket: Use MSG_CMSG_CLOEXEC

This ensures received file descriptors don't leak to child processes.

https://bugzilla.gnome.org/show_bug.cgi?id=649480

13 years agoUpdate VS2010 "install" project
Chun-wei Fan [Thu, 5 May 2011 07:09:51 +0000 (15:09 +0800)]
Update VS2010 "install" project

The dependencies on gsettings and glib-compile-schemas were missed-made
up for them here.

13 years agoUpdate gcancellable.c for Win32
Chun-wei Fan [Thu, 5 May 2011 05:40:38 +0000 (13:40 +0800)]
Update gcancellable.c for Win32

errno.h is still required for Windows in this file...

13 years agoUpdate VS property sheets
Chun-wei Fan [Thu, 5 May 2011 05:36:34 +0000 (13:36 +0800)]
Update VS property sheets

There are some new public headers added for GIO.  Make up for them here
in the install phase

13 years agoMSVC support update for GObject
Chun-wei Fan [Thu, 5 May 2011 04:42:44 +0000 (12:42 +0800)]
MSVC support update for GObject

Apparently GObject now depends on libffi.  Add it to the dependencies list.

For MSVC, use the libffi-3.0.10(rcx) versions.

13 years agoGLib MSVC Support updates (autotools)
Chun-wei Fan [Thu, 5 May 2011 04:35:45 +0000 (12:35 +0800)]
GLib MSVC Support updates (autotools)

Filter out *-unix.c from the source file list, as it is obviously not
usable (nor used) under VS 2008/2010

13 years agoPost-release bump
Matthias Clasen [Thu, 5 May 2011 01:26:11 +0000 (21:26 -0400)]
Post-release bump

13 years agoMore distcheck fixes 2.29.4
Matthias Clasen [Thu, 5 May 2011 00:45:46 +0000 (20:45 -0400)]
More distcheck fixes

13 years agoTry to fix distcheck
Matthias Clasen [Wed, 4 May 2011 23:44:05 +0000 (19:44 -0400)]
Try to fix distcheck

13 years agoPre-release bump
Matthias Clasen [Wed, 4 May 2011 18:53:08 +0000 (14:53 -0400)]
Pre-release bump

13 years agoUpdates
Matthias Clasen [Wed, 4 May 2011 18:33:03 +0000 (14:33 -0400)]
Updates

13 years agogdbus-codegen: don't assert non-negativity for unsigned values
Ray Strode [Wed, 4 May 2011 16:45:22 +0000 (12:45 -0400)]
gdbus-codegen: don't assert non-negativity for unsigned values

The code generator sprinkles a few asserts in its output of the form:

g_assert (prop_id - 1 >= 0 && prop_id - 1 < %d);\n

prop_id is unsigned, though, so this generates a compiler warning for
me.

This commit changes the code to merely check for prop_id != 0 instead of
prop_id - 1 >= 0

13 years agoGVariant: Better handling of invalid UTF-8
Ryan Lortie [Wed, 4 May 2011 16:12:31 +0000 (18:12 +0200)]
GVariant: Better handling of invalid UTF-8

g_variant_new_string() hits a g_return_if_fail() when given invalid
UTF-8.  That's certainly the right thing to do, but
g_variant_builder_add() uses this function internally and crashes when
it returns NULL due to the user passing a non-utf8 string.

We can protect the internal code by returning "[Invalid UTF-8]" in this
case while also making the problem easier to debug (and less fatal).

Closes #632631.

13 years agogdbus-codegen: Make sure we pass -Werror
Alexander Larsson [Wed, 4 May 2011 11:54:34 +0000 (13:54 +0200)]
gdbus-codegen: Make sure we pass -Werror

_g_value_equal might not always be used, so mark it G_GNUC_UNUSED
to make sure we're -Werror clean.

13 years agoUpdated Norwegian bokmål translation
Kjartan Maraas [Wed, 4 May 2011 09:06:44 +0000 (11:06 +0200)]
Updated Norwegian bokmål translation

13 years agoGDBusObjectManagerServer: Allow setting :connection property
David Zeuthen [Wed, 4 May 2011 07:37:46 +0000 (03:37 -0400)]
GDBusObjectManagerServer: Allow setting :connection property

.. and add a C setter to do this. Also make the C getter return a
reference since the property may be set from another thread. Also
change the constructor to _not_ take a GDBusConnection since this is
something you almost always want to do _after_ creating it. The
API/ABI break is fine as there has never been a GLib release with this
type.

https://bugzilla.gnome.org/show_bug.cgi?id=648959

Signed-off-by: David Zeuthen <davidz@redhat.com>
13 years agoRename g_unix_pipe_flags to g_unix_open_pipe
Colin Walters [Tue, 3 May 2011 20:31:34 +0000 (16:31 -0400)]
Rename g_unix_pipe_flags to g_unix_open_pipe

From IRC discussion, people liked this name more.

https://bugzilla.gnome.org/show_bug.cgi?id=649322

13 years agoRemove extraneous decorations
Matthias Clasen [Tue, 3 May 2011 23:15:46 +0000 (19:15 -0400)]
Remove extraneous decorations

Since in a doc comment takes no leading '@'.

13 years agoglib-unix.c: Don't leak FDs if pipe2() succeeds
Colin Walters [Tue, 3 May 2011 20:22:13 +0000 (16:22 -0400)]
glib-unix.c: Don't leak FDs if pipe2() succeeds

13 years agoOnly remove the lines that are no longer needed
Matthias Clasen [Tue, 3 May 2011 16:41:11 +0000 (12:41 -0400)]
Only remove the lines that are no longer needed

13 years agoRevert "[build] Remove stale LDADD lines from tests/Makefile.am to allow build"
Matthias Clasen [Tue, 3 May 2011 16:37:45 +0000 (12:37 -0400)]
Revert "[build] Remove stale LDADD lines from tests/Makefile.am to allow build"

This reverts commit b0874ce28e910121fb2d94f6f9183bfdc08e8eaf.

13 years agoUse g types in public API for consistency
Colin Walters [Tue, 3 May 2011 14:57:22 +0000 (10:57 -0400)]
Use g types in public API for consistency

13 years agoGCancellable: Use g_unix_pipe_flags with FD_CLOEXEC
Colin Walters [Mon, 2 May 2011 19:42:51 +0000 (15:42 -0400)]
GCancellable: Use g_unix_pipe_flags with FD_CLOEXEC

The old code was creating a pipe and setting FD_CLOEXEC
non-atomically.

https://bugzilla.gnome.org/show_bug.cgi?id=649225

13 years agog_unix_set_fd_nonblocking: New API to control file descriptor blocking state
Colin Walters [Tue, 3 May 2011 13:52:10 +0000 (09:52 -0400)]
g_unix_set_fd_nonblocking: New API to control file descriptor blocking state

And use it in relevant places in GLib.

https://bugzilla.gnome.org/show_bug.cgi?id=649225

13 years agoglib-unix.h: Unconditionally include unistd.h
Colin Walters [Mon, 2 May 2011 21:02:16 +0000 (17:02 -0400)]
glib-unix.h: Unconditionally include unistd.h

danw points out it's part of POSIX.

https://bugzilla.gnome.org/show_bug.cgi?id=649201