platform/upstream/glib.git
14 years agoG{Byte,Ptr,}Array: move docs from tmpl to .c
Ryan Lortie [Sun, 31 Jan 2010 19:30:08 +0000 (14:30 -0500)]
G{Byte,Ptr,}Array: move docs from tmpl to .c

14 years agoGList, GSList: move docs from tmpl to .c
Ryan Lortie [Sun, 31 Jan 2010 18:07:16 +0000 (13:07 -0500)]
GList, GSList: move docs from tmpl to .c

14 years agoremove trailing whitespace breaking .gitignore
Ryan Lortie [Sun, 31 Jan 2010 06:26:49 +0000 (01:26 -0500)]
remove trailing whitespace breaking .gitignore

14 years agoGIOChannel: move docs from tmpl to .c
Ryan Lortie [Sun, 31 Jan 2010 05:39:40 +0000 (00:39 -0500)]
GIOChannel: move docs from tmpl to .c

14 years agoGTree, GNode: move docs from tmpl to .c
Ryan Lortie [Sun, 31 Jan 2010 05:27:28 +0000 (00:27 -0500)]
GTree, GNode: move docs from tmpl to .c

14 years agoGThreadPool: move docs from tmpl to .c
Ryan Lortie [Sun, 31 Jan 2010 05:18:09 +0000 (00:18 -0500)]
GThreadPool: move docs from tmpl to .c

14 years agoGRelation: move docs from tmpl to .c
Ryan Lortie [Sun, 31 Jan 2010 05:16:14 +0000 (00:16 -0500)]
GRelation: move docs from tmpl to .c

14 years agoGCache: move docs from tmpl to .c
Ryan Lortie [Sun, 31 Jan 2010 05:10:58 +0000 (00:10 -0500)]
GCache: move docs from tmpl to .c

14 years agoGChecksum: move docs from tmpl to .c
Ryan Lortie [Sun, 31 Jan 2010 05:05:48 +0000 (00:05 -0500)]
GChecksum: move docs from tmpl to .c

14 years agoGCompletion: move docs from tmpl to .c
Ryan Lortie [Sun, 31 Jan 2010 05:03:33 +0000 (00:03 -0500)]
GCompletion: move docs from tmpl to .c

14 years agoFix typo in introduction to threads in Glib
Christian Dywan [Mon, 1 Feb 2010 11:37:23 +0000 (12:37 +0100)]
Fix typo in introduction to threads in Glib

14 years agoUpdated Norwegian bokmål translation
Kjartan Maraas [Sun, 31 Jan 2010 18:57:59 +0000 (19:57 +0100)]
Updated Norwegian bokmål translation

14 years agoFix shadowing of G_VALUE_COLLECT variables in G_VALUE_COLLECT_INIT
Philip Withnall [Sun, 31 Jan 2010 16:15:32 +0000 (16:15 +0000)]
Fix shadowing of G_VALUE_COLLECT variables in G_VALUE_COLLECT_INIT

Closes: bgo#608602

14 years agodata{list,sets}, quarks: move docs from tmpl to .c
Ryan Lortie [Sun, 31 Jan 2010 04:34:19 +0000 (23:34 -0500)]
data{list,sets}, quarks: move docs from tmpl to .c

14 years agoGPattern: move docs from tmpl to .c
Ryan Lortie [Sun, 31 Jan 2010 04:10:34 +0000 (23:10 -0500)]
GPattern: move docs from tmpl to .c

14 years agoString chunks: move docs from tmpl to .c
Ryan Lortie [Sun, 31 Jan 2010 04:04:52 +0000 (23:04 -0500)]
String chunks: move docs from tmpl to .c

14 years agoMove allocator and memchunk docs from tmpl to .c
Ryan Lortie [Sun, 31 Jan 2010 03:58:43 +0000 (22:58 -0500)]
Move allocator and memchunk docs from tmpl to .c

14 years agoGSequence: move docs from tmpl to .c
Ryan Lortie [Sun, 31 Jan 2010 03:19:29 +0000 (22:19 -0500)]
GSequence: move docs from tmpl to .c

14 years agoGShell: move docs from tmpl to .c
Ryan Lortie [Sun, 31 Jan 2010 03:15:35 +0000 (22:15 -0500)]
GShell: move docs from tmpl to .c

14 years agoGTimer: move docs from tmpl to .c
Ryan Lortie [Sun, 31 Jan 2010 03:12:22 +0000 (22:12 -0500)]
GTimer: move docs from tmpl to .c

14 years agoFix up gthread.c documentation
Ryan Lortie [Sun, 31 Jan 2010 02:55:03 +0000 (21:55 -0500)]
Fix up gthread.c documentation

The last incomplete commit got pushed accidentally.  This finishes it up.

14 years agomerge GVariantTypeInfo
Ryan Lortie [Sun, 31 Jan 2010 01:15:25 +0000 (20:15 -0500)]
merge GVariantTypeInfo

14 years agoRevert accidental commit "dirt"
Ryan Lortie [Sat, 30 Jan 2010 21:00:51 +0000 (16:00 -0500)]
Revert accidental commit "dirt"

This reverts commit c46bc6ba1fc8c5045c9a93b95e78e81936e2d954.

14 years agoGIO: Remove trailing "." from Since: tags in docs
Ryan Lortie [Sat, 30 Jan 2010 19:06:12 +0000 (14:06 -0500)]
GIO: Remove trailing "." from Since: tags in docs

Typing "Since: 2.22." results in the creation of an api-index-2.22..xml
file (ie: the last '.' is taken to be part of the version number).

14 years agodirt
Ryan Lortie [Sat, 30 Jan 2010 18:00:01 +0000 (13:00 -0500)]
dirt

14 years agomove threads docs to .c
Ryan Lortie [Sat, 30 Jan 2010 17:32:09 +0000 (12:32 -0500)]
move threads docs to .c

14 years agoGIOChannel: fill in missing docs
Ryan Lortie [Sat, 30 Jan 2010 18:29:23 +0000 (13:29 -0500)]
GIOChannel: fill in missing docs

14 years agog_strlcpy doc: small fixup
Ryan Lortie [Sat, 30 Jan 2010 17:30:53 +0000 (12:30 -0500)]
g_strlcpy doc: small fixup

prevent the <note> from getting sucked in as part of the Returns:

14 years agogbitlock: fix gtkdoc brokenness
Ryan Lortie [Sat, 30 Jan 2010 17:13:50 +0000 (12:13 -0500)]
gbitlock: fix gtkdoc brokenness

14 years agoGRand: move docs from tmpl to inline comments
Ryan Lortie [Sat, 30 Jan 2010 06:00:50 +0000 (01:00 -0500)]
GRand: move docs from tmpl to inline comments

14 years agoUpdated Spanish translation
Jorge González [Sat, 30 Jan 2010 11:37:00 +0000 (12:37 +0100)]
Updated Spanish translation

14 years agogvarianttype: fix a trivial missing const
Ryan Lortie [Sat, 30 Jan 2010 05:00:48 +0000 (00:00 -0500)]
gvarianttype: fix a trivial missing const

14 years agodon't #include <glib.h> from other public headers
Ryan Lortie [Sat, 30 Jan 2010 04:56:04 +0000 (23:56 -0500)]
don't #include <glib.h> from other public headers

fix up some problems that were hidden by that

14 years agoUpdated French translation
Claude Paroz [Fri, 29 Jan 2010 21:52:58 +0000 (22:52 +0100)]
Updated French translation

14 years agoUpdate POTFILES.in to make l10n.gnome.org happy
Andre Klapper [Fri, 29 Jan 2010 21:33:06 +0000 (22:33 +0100)]
Update POTFILES.in to make l10n.gnome.org happy

14 years agoimprove gitignore
Ryan Lortie [Fri, 29 Jan 2010 03:43:51 +0000 (22:43 -0500)]
improve gitignore

14 years agoGHash: move docs from tmpl to inline comments
Ryan Lortie [Fri, 29 Jan 2010 03:36:48 +0000 (22:36 -0500)]
GHash: move docs from tmpl to inline comments

14 years agoSince: markers for g_bit_*lock()
Ryan Lortie [Fri, 29 Jan 2010 00:12:55 +0000 (19:12 -0500)]
Since: markers for g_bit_*lock()

14 years agoBug 548967 - 1 bit mutex lock: add tests
Ryan Lortie [Thu, 28 Jan 2010 23:41:19 +0000 (18:41 -0500)]
Bug 548967 - 1 bit mutex lock: add tests

Add a test case for the new API.

Always check the emulated futex(2) implementation, even on systems with
futex support.

14 years agoBug 548967 - 1 bit mutex lock
Ryan Lortie [Thu, 28 Jan 2010 16:32:07 +0000 (11:32 -0500)]
Bug 548967 - 1 bit mutex lock

Add support for a mutex lock that consumes only one bit of storage
inside of an integer on systems that support futexes.  Futex is emulated
(at a higher cost) on systems that don't have it -- but only in the
contended case.

14 years agoglib docs: Delete and ignore empty .tmpl files
Ryan Lortie [Thu, 28 Jan 2010 17:39:46 +0000 (12:39 -0500)]
glib docs: Delete and ignore empty .tmpl files

14 years agoReturn NULL from g_mapped_file_get_contents()
Ryan Lortie [Thu, 28 Jan 2010 02:30:26 +0000 (21:30 -0500)]
Return NULL from g_mapped_file_get_contents()

In the case of an empty file, return NULL instead of "".  This means
that the return result of g_mapped_file_get_contents() will always be
page-aligned.

14 years agoUpdated Slovenian translation
Matej Urbančič [Wed, 27 Jan 2010 12:16:13 +0000 (13:16 +0100)]
Updated Slovenian translation

14 years agoImprove type documentation
Behdad Esfahbod [Tue, 26 Jan 2010 21:12:20 +0000 (16:12 -0500)]
Improve type documentation

14 years agoAdd gtk-doc comments with Since tags to GMemoryOutputStream properties
Christian Dywan [Tue, 26 Jan 2010 19:45:40 +0000 (20:45 +0100)]
Add gtk-doc comments with Since tags to GMemoryOutputStream properties

14 years ago[glib/gvarianttype] Move some documentation to header file
Javier Jardón [Tue, 26 Jan 2010 18:15:48 +0000 (19:15 +0100)]
[glib/gvarianttype] Move some documentation to header file

Also, change G_VARIANT_TYPE(str) for G_VARIANT_TYPE(type_string)
so It's correctly documented now.

14 years ago[glib/gvariantype] Added missing "Since: 2.24" tag
Javier Jardón [Tue, 26 Jan 2010 18:00:41 +0000 (19:00 +0100)]
[glib/gvariantype] Added missing "Since: 2.24" tag

14 years agoalways use our own internal assertion message symbol
Martin Pitt [Tue, 26 Jan 2010 10:22:31 +0000 (11:22 +0100)]
always use our own internal assertion message symbol

Re-using glibc's __abort_msg symbol causes linking problems, since the symbol
is declared private. Always use our own__glib_abort_msg symbol to store
assertion messages, to avoid compatibility and linking problems.

Also fix the test case to work with out of tree builds (such as "make
distcheck"), and re-enable it.

https://bugzilla.gnome.org/show_bug.cgi?id=594872

14 years agoBump version
Matthias Clasen [Mon, 25 Jan 2010 19:16:09 +0000 (14:16 -0500)]
Bump version

14 years ago2.23.2 2.23.2
Matthias Clasen [Mon, 25 Jan 2010 19:14:14 +0000 (14:14 -0500)]
2.23.2

14 years agoDisable yet another failing test that was recently added...
Matthias Clasen [Mon, 25 Jan 2010 19:13:48 +0000 (14:13 -0500)]
Disable yet another failing test that was recently added...

14 years agoMinor docs update
Matthias Clasen [Mon, 25 Jan 2010 17:46:35 +0000 (12:46 -0500)]
Minor docs update

14 years agoAnother few failing tests disabled
Matthias Clasen [Mon, 25 Jan 2010 17:43:10 +0000 (12:43 -0500)]
Another few failing tests disabled

14 years agoDisable a failing test, tsk, tsk.
Matthias Clasen [Mon, 25 Jan 2010 17:29:09 +0000 (12:29 -0500)]
Disable a failing test, tsk, tsk.

14 years agoMention GVariant
Matthias Clasen [Mon, 25 Jan 2010 17:00:06 +0000 (12:00 -0500)]
Mention GVariant

14 years agoUpdates
Matthias Clasen [Mon, 25 Jan 2010 16:57:47 +0000 (11:57 -0500)]
Updates

14 years agoadd testcase for GVariantType
Ryan Lortie [Mon, 25 Jan 2010 16:31:56 +0000 (11:31 -0500)]
add testcase for GVariantType

14 years agomerge GVariantType
Ryan Lortie [Mon, 25 Jan 2010 16:26:27 +0000 (11:26 -0500)]
merge GVariantType

GVariantType is the type system for GVariant

14 years agoMove the boxed private type data to TypeNode
Benjamin Otte [Wed, 20 Jan 2010 20:06:30 +0000 (21:06 +0100)]
Move the boxed private type data to TypeNode

This way we don't need to keep a custom array that we bsearch on (and
that isn't threadsafe) but can use the gtype.c machinery that is
threadsafe. And fast, too!

https://bugzilla.gnome.org/show_bug.cgi?id=554887

14 years agoPut calls to registered copy/free functions into separate functions
Benjamin Otte [Wed, 20 Jan 2010 19:02:24 +0000 (20:02 +0100)]
Put calls to registered copy/free functions into separate functions

This eases cleaning up these functions.

One optimization in value_set_internal() was lost in the process. It
shouldn't cause too many issues when all is said and done.

https://bugzilla.gnome.org/show_bug.cgi?id=554887

14 years agoUpdated Bulgarian translation
Alexander Shopov [Mon, 25 Jan 2010 09:58:18 +0000 (11:58 +0200)]
Updated Bulgarian translation

14 years agoTypo fixes
Matthias Clasen [Mon, 25 Jan 2010 01:50:51 +0000 (20:50 -0500)]
Typo fixes

14 years agoUpdated Bengali translation
Jamil Ahmed [Sun, 24 Jan 2010 16:54:12 +0000 (22:54 +0600)]
Updated Bengali translation

14 years agogio: .gitignore gio-querymodules
Ryan Lortie [Sat, 23 Jan 2010 06:07:46 +0000 (01:07 -0500)]
gio: .gitignore gio-querymodules

14 years agoBug 448888 - don't init g_slice for always-malloc
Ryan Lortie [Sat, 23 Jan 2010 01:18:58 +0000 (20:18 -0500)]
Bug 448888 - don't init g_slice for always-malloc

14 years ago[tests] Remove C++ style comments
Alberto Garcia [Wed, 20 Jan 2010 17:53:08 +0000 (18:53 +0100)]
[tests] Remove C++ style comments

It makes the IBM XL C Compiler (the 'native' non-free compiler
on the AIX 5.3 and 6.1 platform) stop compiling with syntax error.

Fixes https://bugzilla.gnome.org/show_bug.cgi?id=581300

Signed-off-by: Javier Jardón <jjardon@gnome.org>
14 years ago[configure] More -lresolv-checking-fixing
Dan Winship [Mon, 18 Jan 2010 14:46:19 +0000 (09:46 -0500)]
[configure] More -lresolv-checking-fixing

https://bugzilla.gnome.org/show_bug.cgi?id=586150

14 years agoEnsure values are memset to 0 when calling G_VALUE_COLLECT_INIT()
Benjamin Otte [Mon, 18 Jan 2010 13:03:16 +0000 (14:03 +0100)]
Ensure values are memset to 0 when calling G_VALUE_COLLECT_INIT()

The reason we need to enforce this is that the GTypeValueTable
documentation explicitly states that memory is memset to 0 when the
value_init function is called.

https://bugzilla.gnome.org/show_bug.cgi?id=603590

14 years ago[configure] Fix res_query check for OS X
Dan Winship [Wed, 13 Jan 2010 16:24:28 +0000 (11:24 -0500)]
[configure] Fix res_query check for OS X

https://bugzilla.gnome.org/show_bug.cgi?id=586150

14 years agoUpgrade to version 0.8 of libasyncns
Javier Jardón [Tue, 17 Nov 2009 20:15:15 +0000 (21:15 +0100)]
Upgrade to version 0.8 of libasyncns

Fixes https://bugzilla.gnome.org/show_bug.cgi?id=602240

14 years agoUpdated Basque language
Inaki Larranaga Murgoitio [Wed, 13 Jan 2010 16:29:39 +0000 (17:29 +0100)]
Updated Basque language

14 years agoUpdated Norwegian bokmål translation
Kjartan Maraas [Wed, 13 Jan 2010 15:20:56 +0000 (16:20 +0100)]
Updated Norwegian bokmål translation

14 years agoproperly abort instead of looping infinitely
Sven Herzberg [Wed, 13 Jan 2010 13:47:58 +0000 (14:47 +0100)]
properly abort instead of looping infinitely

Fixes: Bug 568760 - nautilus freezes due to a bug in garray.c:322

* glib/garray.c: increase the size of potential return values by
  using an unsigned result; properly check if we still handle valid size
  proposals, return the original request if there's no usable size left
* tests/array-test.c: reproduce the error condition of the bug report

14 years agoFix docs in previous commit
Alexander Larsson [Wed, 13 Jan 2010 09:25:52 +0000 (10:25 +0100)]
Fix docs in previous commit

14 years agoRemove additional thread support in performance test
Alexander Larsson [Wed, 13 Jan 2010 09:24:09 +0000 (10:24 +0100)]
Remove additional thread support in performance test

We're always enabling threads now so this is not needed.

14 years agogobject: Use new G_VALUE_COLLECT_INIT variant
Edward Hervey [Wed, 2 Dec 2009 10:50:02 +0000 (11:50 +0100)]
gobject: Use new G_VALUE_COLLECT_INIT variant

Makes g_object_new_valist 20% to 30% faster (against 2321e5a).

Profiled against the pan newsreader which uses a variant of simple
and complex object creation.

https://bugzilla.gnome.org/show_bug.cgi?id=603590

14 years agogsignal: Use new G_VALUE_COLLECT_INIT variant
Edward Hervey [Wed, 2 Dec 2009 10:49:49 +0000 (11:49 +0100)]
gsignal: Use new G_VALUE_COLLECT_INIT variant

Makes g_signal_emit_valist from 15% to 20% faster.

Results reported from profiling the pan newsreader which uses a variant
of simple and complex signal emissions (i.e no args or various args)

https://bugzilla.gnome.org/show_bug.cgi?id=603590

14 years agogvaluecollector: Add variant of G_VALUE_COLLECT for most used cases.
Edward Hervey [Wed, 2 Dec 2009 10:48:18 +0000 (11:48 +0100)]
gvaluecollector: Add variant of G_VALUE_COLLECT for most used cases.

Most callers of G_VALUE_COLLECT previously had to initialize the GValue
and then G_VALUE_COLLECT would still go through a cleanup phase.

The new variant allows passing a unitialized GValue along with a GType
and speedup the initialization/collection process.

https://bugzilla.gnome.org/show_bug.cgi?id=603590

14 years agoMention thread changes in NEWS
Alexander Larsson [Tue, 12 Jan 2010 20:55:15 +0000 (21:55 +0100)]
Mention thread changes in NEWS

14 years agoUse unconditional thread calls in gio and gobject
Alexander Larsson [Tue, 12 Jan 2010 20:28:23 +0000 (21:28 +0100)]
Use unconditional thread calls in gio and gobject

If threads are available we always enable threads in gobject, which
means all gio/gobject code can enable the unconditional thread calls.

This is a minor optimization since we avoid a bunch of unnecessary
is-threads-enabled checks.

https://bugzilla.gnome.org/show_bug.cgi?id=606775

14 years agoMake thread calls unconditional if G_THREADS_MANDATORY is set
Alexander Larsson [Tue, 12 Jan 2010 20:26:52 +0000 (21:26 +0100)]
Make thread calls unconditional if G_THREADS_MANDATORY is set

Since gobject now initialized threads unconditionally if threads are
available it makes no sense to have each thread operation
conditionally check if threads are enables, so allow this to be avoided.

https://bugzilla.gnome.org/show_bug.cgi?id=606775

14 years agoEnable threads in g_type_init()
Alexander Larsson [Tue, 12 Jan 2010 20:22:45 +0000 (21:22 +0100)]
Enable threads in g_type_init()

This means threads will be supported for all gobject libraries/applications
and initialized early enough to not cause any problems.

This solves the problem of libraries needing threadsafety. Previosly
they just called g_threads_init() anyway, which often works but sometimes
breaks in unexpected ways.

See this thread for more details:
http://mail.gnome.org/archives/gtk-devel-list/2009-November/msg00208.html

https://bugzilla.gnome.org/show_bug.cgi?id=606775

14 years agoRelax g_thread_init() requirements
Alexander Larsson [Tue, 12 Jan 2010 20:16:59 +0000 (21:16 +0100)]
Relax g_thread_init() requirements

We now allow g_thread_init(NULL) to be called after other glib calls (with
some minor limitations). This is mainly a documentation change as this
really was already possible.

We also allow g_thread_init() to be called multiple times. Only the
first call actually initializes the threading system, further calls
are ignored (but print a warning if the argument is not NULL).

https://bugzilla.gnome.org/show_bug.cgi?id=606775

14 years agoImplement lazy loading of fam plugin
Alexander Larsson [Tue, 12 Jan 2010 12:09:10 +0000 (13:09 +0100)]
Implement lazy loading of fam plugin

14 years agoAdd gio-querymodule program
Alexander Larsson [Tue, 12 Jan 2010 12:11:22 +0000 (13:11 +0100)]
Add gio-querymodule program

This can be used to update the giomodule.cache file in directories with
giomodules in order to support lazy module loading.

14 years agoUse the new lazy loading of modules for gio modules
Alexander Larsson [Tue, 12 Jan 2010 10:57:17 +0000 (11:57 +0100)]
Use the new lazy loading of modules for gio modules

14 years agoAdd support for lazy loading of giomodules
Alexander Larsson [Tue, 12 Jan 2010 10:36:12 +0000 (11:36 +0100)]
Add support for lazy loading of giomodules

Adds an optional query method to giomodules which should return all
possible extension points the module may implement.

Then we add a new call g_io_modules_scan_all_in_directory() similar to
g_io_modules_load_all_in_directory() that doesn't return all loaded
modules, thus allowing lazy loading.

In g_io_modules_scan_all_in_directory we look for an optional
giomodule.cache file and use the information in that to avoid
loading modules until they are needed for an extension point.

14 years agoFlush gcharsetconverter when needed.
Ignacio Casal Quinteiro [Tue, 12 Jan 2010 11:54:49 +0000 (12:54 +0100)]
Flush gcharsetconverter when needed.

This patch fixes bug #605686.

14 years agog_object_new(): skip varargs for simple calls
Kamal Mostafa [Sat, 2 Jan 2010 04:36:28 +0000 (20:36 -0800)]
g_object_new(): skip varargs for simple calls

g_object_new() avoids useless varargs processing for simple calls with no
properties.  Fixes https://bugzilla.gnome.org/show_bug.cgi?id=605883

14 years agoUpdated Swedish translation
Daniel Nylander [Sat, 9 Jan 2010 15:43:35 +0000 (16:43 +0100)]
Updated Swedish translation

14 years agoUpdated Spanish translation
Jorge González [Thu, 7 Jan 2010 19:26:48 +0000 (20:26 +0100)]
Updated Spanish translation

14 years agoUse GCC atomic builtin operations.
Hiroyuki Ikezoe [Thu, 7 Jan 2010 10:33:39 +0000 (19:33 +0900)]
Use GCC atomic builtin operations.

Fix for bug #531902.

14 years agodocs: switch to xi:inbclude for the content to save some more seconds
Stefan Kost [Thu, 7 Jan 2010 08:47:20 +0000 (10:47 +0200)]
docs: switch to xi:inbclude for the content to save some more seconds

14 years agodocs: switch to xi:inbclude for the content to save some more seconds
Stefan Kost [Thu, 7 Jan 2010 09:31:24 +0000 (11:31 +0200)]
docs: switch to xi:inbclude for the content to save some more seconds

14 years agodocs: use xi:include to use the generated indexes if available.
Stefan Kost [Thu, 7 Jan 2010 08:22:14 +0000 (10:22 +0200)]
docs: use xi:include to use the generated indexes if available.

This works transparantly for people with new enough gtk-doc (>=1.11) and
significantly improves the build time.

14 years agoDon't do pointer arithmetics on void*
Haakon Sporsheim [Thu, 7 Jan 2010 08:12:13 +0000 (10:12 +0200)]
Don't do pointer arithmetics on void*

For instance MSVC doesn't like that.

Signed-off-by: Tor Lillqvist <tml@iki.fi>
14 years agoAdd properties to GMemoryOutputStream
Matthias Clasen [Wed, 6 Jan 2010 22:37:11 +0000 (17:37 -0500)]
Add properties to GMemoryOutputStream

This helps bindings. Patch by Krzysztof Kosiński. See bug 605733.

14 years agoFix the filename roundtrip check
Matthias Clasen [Wed, 6 Jan 2010 18:23:48 +0000 (13:23 -0500)]
Fix the filename roundtrip check

The check was always failing because the code was comparing
the wrong strings. Reported in bug 605977.

14 years agoGUnixFDMessage: fix a small typo in an extra check
Ryan Lortie [Tue, 5 Jan 2010 23:49:23 +0000 (18:49 -0500)]
GUnixFDMessage: fix a small typo in an extra check

In the deserialise function, GUnixFDMessage was comparing 'level' to
both SOL_SOCKET and SCM_RIGHTS.  It is correct to compare 'type' to
SCM_RIGHTS.  The code passed tests only because:

1) it's a "should always be OK" double-check

2) SOL_SOCKET and SCM_RIGHTS, by chance, both have the value '1' on
   Linux systems.

14 years agoFix a memleak
Matthias Clasen [Tue, 5 Jan 2010 23:18:55 +0000 (18:18 -0500)]
Fix a memleak

An early exit in expand_application_parameters forgot to free
a GString. Reported by Steve Grubb.