Benjamin Otte [Tue, 1 Nov 2011 22:37:35 +0000 (23:37 +0100)]
localfileinfo: Split out attributes that don't need a stat
This will be used in the next commit.
Benjamin Otte [Tue, 1 Nov 2011 22:21:42 +0000 (23:21 +0100)]
localfileenumerator: Take the type from the readdir() call
That way, we can avoid stat() calls for all enumerated files in various
cases. In particular in the autocompletion code in the GTK filechooser.
Benjamin Otte [Tue, 1 Nov 2011 19:12:27 +0000 (20:12 +0100)]
gio: Add tests for g_file_attribute_matcher_subtract()
Benjamin Otte [Tue, 1 Nov 2011 19:11:47 +0000 (20:11 +0100)]
fileinfo: Add g_file_attribute_matcher_subtract()
Added as public API so I can write tests, the use case is local.
Benjamin Otte [Tue, 1 Nov 2011 18:12:17 +0000 (19:12 +0100)]
tests: Add tests for fileattributematchers
Benjamin Otte [Tue, 1 Nov 2011 18:06:39 +0000 (19:06 +0100)]
fileinfo: Add g_file_attribute_matcher_to_string()
This is to be mainly used for debugging and tests.
Benjamin Otte [Tue, 1 Nov 2011 16:50:22 +0000 (17:50 +0100)]
fileinfo: Store namespace::* as 0th attribute
This way, we can get_attribute() namespaces. This will be important in
the next commit.
Benjamin Otte [Tue, 1 Nov 2011 12:58:09 +0000 (13:58 +0100)]
fileinfo: Change the way attribute matchers are created
We now sort the matchers and remove unnecessary duplicates (like
removing standard:type when we already match standard:*), so that we can
do more complex operations on them easily in later commits.
Benjamin Otte [Tue, 1 Nov 2011 11:47:33 +0000 (12:47 +0100)]
fileinfo: Remove optimization for GFileAttributeMatcher
It makes code more complicated, in particular the code I'm about to add.
Lucas De Marchi [Thu, 6 Oct 2011 14:18:03 +0000 (11:18 -0300)]
Fix G*_TO_POINTER casts on 32 bits
If we don't do the cast to the proper size in 32 bits, things like below
doesn't work:
uint8_t u = 20;
void *p;
p = GUINT_TO_POINTER(u);
Signed-off-by: Colin Walters <walters@verbum.org>
https://bugzilla.gnome.org/show_bug.cgi?id=661546
Alexander Larsson [Fri, 4 Nov 2011 15:45:41 +0000 (16:45 +0100)]
win32: Make g_get_monotonic_clock lockless
Alexander Larsson [Fri, 4 Nov 2011 14:42:38 +0000 (15:42 +0100)]
win32: Use timeGetTime as monotonic base
This allows apps that need it to increase timer accuracy
using timeBeginPeriod
Alexander Larsson [Fri, 4 Nov 2011 10:27:00 +0000 (11:27 +0100)]
win32: Add a monotonic timer
Piotr Drąg [Tue, 15 Nov 2011 19:46:17 +0000 (20:46 +0100)]
Updated POTFILES.in
Daniel Mustieles [Tue, 15 Nov 2011 16:37:39 +0000 (17:37 +0100)]
Updated Spanish translation
Javier Jardón [Tue, 15 Nov 2011 12:11:55 +0000 (12:11 +0000)]
docs: Remove tmpl directory
We use inline comments now
Sjoerd Simons [Sun, 13 Nov 2011 16:49:57 +0000 (17:49 +0100)]
Fix 1bit mutex test on platforms with 32 bit aligned pointers
The 1 bit mutex tests asserts: ((gsize) ptrs) % 8, ==, 0), which fails
when the platform only aligns porters to 32 bits (e.g. S390 and
powerpc).
I'm not sure why this assertion was placed here, but given
that internally g_pointer_bit_trylock uses g_atomic_int_or internally
change the assertion so it only requires the alignment to be a multiple
of sizeof(int)
Matthias Clasen [Tue, 15 Nov 2011 04:30:18 +0000 (23:30 -0500)]
Avoid unused variable warnings in g_ascii_strtod
Matthias Clasen [Tue, 15 Nov 2011 03:18:13 +0000 (22:18 -0500)]
GThread: handle thread names safely
Avoid a race condition where the string may be freed before
the new thread got around to using it. Also add a test for
thread names.
https://bugzilla.gnome.org/show_bug.cgi?id=663381
Matthias Clasen [Tue, 15 Nov 2011 02:18:12 +0000 (21:18 -0500)]
Switch to using template-free docs build
All docs are finally inline, so we can do this switch.
Matthias Clasen [Mon, 14 Nov 2011 12:44:52 +0000 (07:44 -0500)]
Move remaining docs inline
This introduces a fake source file just for holding
docs that have no good place elsewhere. Not great, but
better than templates.
Matthias Clasen [Tue, 15 Nov 2011 02:06:30 +0000 (21:06 -0500)]
Move GModule docs inline
Matthias Clasen [Sun, 13 Nov 2011 20:37:56 +0000 (15:37 -0500)]
Move GDate docs inline
Matthias Clasen [Sun, 13 Nov 2011 05:43:50 +0000 (00:43 -0500)]
Move i18n docs inline
Matthias Clasen [Sun, 13 Nov 2011 05:26:57 +0000 (00:26 -0500)]
Move slice and hook docs inline
Dan Winship [Tue, 15 Nov 2011 00:46:02 +0000 (19:46 -0500)]
GSocketClient: improve error messages on connection failure
Include the hostname (or proxy hostname if it was the connection to
the proxy server that failed) in the GError message when
g_socket_client_connect* fail.
https://bugzilla.gnome.org/show_bug.cgi?id=661266
Dan Winship [Mon, 14 Nov 2011 23:27:34 +0000 (18:27 -0500)]
GUnixInput/OutputStream: fix blocking methods to always block
Previously, if you created a GUnixInputStream or GUnixOutputStream
from a non-blocking file descriptor, it might sometimes return
G_IO_ERROR_WOULD_BLOCK from g_input_stream_read/g_output_stream_write,
which is wrong. Fix that. (Use the GPollableInput/OutputStream methods
if you want non-blocking I/O.)
Also, add a test for this to gio/tests/unix-streams.
Also, fix the GError messages to say "Error reading from file
descriptor", etc instead of "Error reading from unix" (which was
presumably from a bad search and replace job).
https://bugzilla.gnome.org/show_bug.cgi?id=626866
Dan Winship [Mon, 14 Nov 2011 18:55:14 +0000 (13:55 -0500)]
Fix glib/tests/utils for g_parse_debug_string() changes
Dan Winship [Sun, 12 Jun 2011 19:59:36 +0000 (15:59 -0400)]
add GNetworkMonitor, for... monitoring the network
Add GNetworkMonitor and its associated extension point, provide a base
implementation that always claims the network is available, and a
netlink-based implementation built on top of that that actually tracks
the network state.
https://bugzilla.gnome.org/show_bug.cgi?id=620932
Dan Winship [Sat, 1 Oct 2011 12:31:54 +0000 (08:31 -0400)]
GInetAddressMask: new type for internet address range matching
Eg, for matching a GInetAddress to a range like "10.0.0.0/8" or
"fe80::/10"
https://bugzilla.gnome.org/show_bug.cgi?id=620932
Dan Winship [Sun, 13 Nov 2011 21:39:45 +0000 (16:39 -0500)]
gunixinput/outputstream: fix docs
Don't try to link to the docs of undocumented private classes
Dan Winship [Tue, 27 Apr 2010 20:54:18 +0000 (16:54 -0400)]
GUnixInputStream, GUnixOutputStream: support ordinary files better
If the fd is not a pipe or socket, fall back to using threads to do
async I/O rather than poll, since poll doesn't work the way you want
for ordinary files.
https://bugzilla.gnome.org/show_bug.cgi?id=606913
Matthias Clasen [Sun, 13 Nov 2011 03:52:24 +0000 (22:52 -0500)]
Some more docs reshuffling
Matthias Clasen [Sun, 13 Nov 2011 02:54:42 +0000 (21:54 -0500)]
More consistent doc formatting
Move some things around, make capitalization of short descriptions
more consistent.
Matthias Clasen [Sun, 13 Nov 2011 00:13:44 +0000 (19:13 -0500)]
Cleanups to debug env var handling
Matthias Clasen [Sat, 12 Nov 2011 23:36:52 +0000 (18:36 -0500)]
Parse G_DEBUG only once
Matthias Clasen [Sat, 12 Nov 2011 15:26:31 +0000 (10:26 -0500)]
glib-unix.h: Add G_BEGIN/END_DECLS
The omission was pointed out in bug 663880.
Chun-wei Fan [Sat, 12 Nov 2011 02:38:31 +0000 (10:38 +0800)]
gstrfuncs.c: Fix variable declaration
Declare variables at top of block to stop C89 compilers from complaining.
Simon McVittie [Fri, 11 Nov 2011 14:41:50 +0000 (14:41 +0000)]
GDBusWorker: tolerate read errors while closing
My previous fix for GNOME#662100 was incomplete: it seems that with some
timings, the stream can be closed with an async read in-flight. This
can make the read fail immediately with G_IO_ERROR_CLOSED instead of
becoming cancelled.
This happens reliably on an embedded device, and rarely on my laptop;
repeating the test 100 times in quick succession reliably reproduces
the bug on my laptop.
It seems as though what we really want is to ignore read errors, once
we've established that we want to close the connection anyway - this
means that after asking to close, you're immune to exit-on-close,
which seems like a good rule.
An additional subtlety is that continuing to read after we know we
want to close is still required, otherwise we'll never emit ::closed.
Bug: https://bugzilla.gnome.org/show_bug.cgi?id=662100
Bug-NB: NB#287088
Signed-off-by: Simon McVittie <simon.mcvittie@collabora.co.uk>
Reviewed-by: Colin Walters <walters@verbum.org>
Simon McVittie [Thu, 10 Nov 2011 19:06:43 +0000 (19:06 +0000)]
GDBusWorker: debug on read errors if transport debugging is enabled
Bug: https://bugzilla.gnome.org/show_bug.cgi?id=662100
Bug-NB: NB#287088
Signed-off-by: Simon McVittie <simon.mcvittie@collabora.co.uk>
Reviewed-by: Colin Walters <walters@verbum.org>
Simon McVittie [Thu, 3 Nov 2011 18:33:08 +0000 (18:33 +0000)]
gdbus-exit-on-close test: don't leak a variant
Bug: https://bugzilla.gnome.org/show_bug.cgi?id=662100
Bug-NB: NB#287088
Signed-off-by: Simon McVittie <simon.mcvittie@collabora.co.uk>
Reviewed-by: Colin Walters <walters@verbum.org>
Simon McVittie [Thu, 3 Nov 2011 18:32:53 +0000 (18:32 +0000)]
gdbus-exit-on-close test: optionally be more verbose for debugging
Bug: https://bugzilla.gnome.org/show_bug.cgi?id=662100
Bug-NB: NB#287088
Signed-off-by: Simon McVittie <simon.mcvittie@collabora.co.uk>
Reviewed-by: Colin Walters <walters@verbum.org>
Dan Winship [Sat, 3 Sep 2011 13:03:17 +0000 (09:03 -0400)]
gio: more implementations of GFileDescriptorBased
Implement GFileDescriptorBased in GSocket{Input,Output}Stream and
GUnix{Input,Output}Stream.
https://bugzilla.gnome.org/show_bug.cgi?id=616852
Kristian Rietveld [Wed, 9 Nov 2011 07:32:33 +0000 (08:32 +0100)]
Need to include xlocale.h on Darwin for strtod_l
This fixes the build of gstrfuncs.c on Darwin.
Marc-André Lureau [Mon, 7 Nov 2011 19:04:15 +0000 (20:04 +0100)]
Add G_SIGNAL_DEPRECATED
Similar to G_PARAM_DEPRECATED. It will warn only for users of the
signals, so a signal can still be emited without warning, for
compatibility reasons.
Apparently, there is no way user flags could have been used before,
so that shouldn't break anyone.
https://bugzilla.gnome.org/show_bug.cgi?id=663581
Dan Winship [Mon, 7 Nov 2011 18:50:00 +0000 (13:50 -0500)]
gio/tests/proxy: don't use deprecated functions
Dan Winship [Mon, 7 Nov 2011 18:48:30 +0000 (13:48 -0500)]
Fix g_variant_get_gtype() deprecation message
Instead of:
warning: ‘g_variant_get_gtype’ is deprecated (declared at ../../gobject/glib-types.h:242): Use '((GType) ((21) << (2)))' instead [-Wdeprecated-declarations]
show:
warning: ‘g_variant_get_gtype’ is deprecated (declared at ../../gobject/glib-types.h:242): Use ''G_VARIANT_GET_TYPE'' instead [-Wdeprecated-declarations]
Also, document the macro-expansion problem in the
G_GNUC_DEPRECATED_FOR docs
Benjamin Otte [Sun, 16 Oct 2011 19:58:29 +0000 (12:58 -0700)]
docs: Improve g_variant_loop() docs
Typo fix and changing "is not recommended" to the nice version of "don't
do this", aka "causes undefined behavior".
Benjamin Otte [Mon, 25 Jul 2011 11:08:59 +0000 (13:08 +0200)]
docs: Tell which function to use to avoid errors.
Chun-wei Fan [Mon, 7 Nov 2011 06:14:45 +0000 (14:14 +0800)]
Update VS property sheets
gmain.h was added into the list of deprecated headers, so "install"
that as weel.
Matthias Clasen [Mon, 7 Nov 2011 05:14:28 +0000 (00:14 -0500)]
Add a test for CDATA in gschema.xml files
Matthias Clasen [Mon, 7 Nov 2011 05:13:56 +0000 (00:13 -0500)]
Accept CDATA in gschema.xml files
Alexander Larsson [Thu, 3 Nov 2011 16:27:08 +0000 (17:27 +0100)]
win32: Remove some debug spew on startup
This is not needed anymore and was causing problems for pkg-config.
Matthias Clasen [Thu, 3 Nov 2011 05:48:54 +0000 (01:48 -0400)]
Make the default log handler more useful
We make the default log handler only print default and informational
messages if the log domain is explicitly requested.
https://bugzilla.gnome.org/show_bug.cgi?id=661926
Matthias Clasen [Thu, 3 Nov 2011 04:39:31 +0000 (00:39 -0400)]
Move old deprecated gmain api into deprecated/
Matthias Clasen [Thu, 3 Nov 2011 04:30:56 +0000 (00:30 -0400)]
Remove single-include guards in deprecated/
Not allowed to include these headers single anymore, either.
Matthias Clasen [Thu, 3 Nov 2011 04:16:41 +0000 (00:16 -0400)]
GIO: Don't use G_DISABLE_DEPRECATED for functions
Matthias Clasen [Thu, 3 Nov 2011 04:09:32 +0000 (00:09 -0400)]
GObject: Don't use G_DISABLE_DEPRECATED for functions
Matthias Clasen [Thu, 3 Nov 2011 04:05:29 +0000 (00:05 -0400)]
Update deprecation docs
Matthias Clasen [Thu, 3 Nov 2011 04:04:15 +0000 (00:04 -0400)]
Don't use G_DISABLE_DEPRECATED guards around deprecated functions
I'm leaving the old-style deprecation guards in place around
deprecated macros, enumeration values, etc, for now.
Dan Winship [Fri, 28 Oct 2011 19:38:52 +0000 (15:38 -0400)]
glib-mkenums: add --identifier-prefix and --symbol-prefix args
Allow passing --identifier-prefix and --symbol-prefix to glib-mkenums,
with the same meanings as in g-ir-scanner, to allow fixing up the enum
name parsing globally rather than needing to add a /<* *>/ override to
each enum.
https://bugzilla.gnome.org/show_bug.cgi?id=661797
Dan Winship [Fri, 28 Oct 2011 19:38:27 +0000 (15:38 -0400)]
docs: don't refer to a deprecated glib-mkenums flag
Refer to the new flag instead
https://bugzilla.gnome.org/show_bug.cgi?id=661797
Dan Winship [Fri, 14 Oct 2011 18:32:11 +0000 (14:32 -0400)]
glib-mkenums.in: fix @ENUMPREFIX@ with /*< underscore_name=... >*/
For enums defined with an underscore_name override, @ENUMPREFIX@
would include a trailing "_". Fix.
https://bugzilla.gnome.org/show_bug.cgi?id=661797
Aleksander Morgado [Thu, 27 Oct 2011 14:54:02 +0000 (16:54 +0200)]
gdbus: avoid warning when finalizing a GDBusObjectManagerClient
If the GDBusObjectManagerClient doesn't get a name owner during its lifetime,
`on_control_proxy_g_signal' will never be connected to any signal, so we
shouldn't dump any warning in that case.
Fixes https://bugzilla.gnome.org/show_bug.cgi?id=662858
David Zeuthen [Thu, 27 Oct 2011 14:30:58 +0000 (10:30 -0400)]
g_bus_own_name: fix race when unowning a name immediately after owning it
... and also add a test to verify that the fix works.
https://bugzilla.gnome.org/show_bug.cgi?id=662808
Signed-off-by: David Zeuthen <davidz@redhat.com>
Ryan Lortie [Wed, 26 Oct 2011 19:10:33 +0000 (15:10 -0400)]
Revert "gthread: G_STATIC_MUTEX_INIT: Fix this for the non-win32 case."
This reverts commit
52fd106724aa79ad57ecaa7ad9d340f8c89da06d.
Murray Cumming [Wed, 26 Oct 2011 09:38:13 +0000 (11:38 +0200)]
gthread: G_STATIC_MUTEX_INIT: Fix this for the non-win32 case.
Also initialize the unused member. This was correct before but
was broken when the ifndef was moved inside the GStaticMutex
struct:
http://git.gnome.org/browse/glib/commit/glib/gthread.h?id=
24652730a9faaedb19b9e90024077eb7f75a6907
This avoids a compiler warning.
Damien Lespinau [Tue, 25 Oct 2011 02:25:59 +0000 (22:25 -0400)]
Add TAGS files to .gitignore
"make tags" is the best way to produce tags in an automake project and
yet very few people are using it so the tags files it generates are never
added to the .gitignore files.
Add TAGS files to .gitignore then.
https://bugzilla.gnome.org/show_bug.cgi?id=615424
Robert Nagy [Tue, 25 Oct 2011 01:18:48 +0000 (21:18 -0400)]
Fix some build issues on OpenBSD
Adapt to some OpenBSD header quirks.
Patch by Robert Nagy.
https://bugzilla.gnome.org/show_bug.cgi?id=661386
Sjoerd Simons [Fri, 21 Oct 2011 13:40:56 +0000 (15:40 +0200)]
Fix document generation in out of tree builds
https://bugzilla.gnome.org/show_bug.cgi?id=662390
Will Thompson [Thu, 20 Oct 2011 15:07:03 +0000 (16:07 +0100)]
g_parse_debug_string: invert flags given besides "all"
Any flags specified as well as "all" are subtracted from the result,
allowing the user to specify FOO_DEBUG="all,bar,baz" to mean "give me
debugging information for everything except bar and baz".
https://bugzilla.gnome.org/show_bug.cgi?id=642452
Will Thompson [Wed, 16 Feb 2011 10:48:22 +0000 (10:48 +0000)]
Test g_parse_debug_string ("all")
https://bugzilla.gnome.org/show_bug.cgi?id=642452
Matthias Clasen [Mon, 24 Oct 2011 17:43:11 +0000 (13:43 -0400)]
Fix 'make check' with clang
Patch by Elias Pipping
https://bugzilla.gnome.org/show_bug.cgi?id=662491
Matthias Clasen [Mon, 24 Oct 2011 17:25:15 +0000 (13:25 -0400)]
Fix g_hash_table_iter_replace
When reusing an existing key 'internally', we must avoid calling
the key_destroy function on the old key.
https://bugzilla.gnome.org/show_bug.cgi?id=662544
Simon McVittie [Fri, 21 Oct 2011 10:42:10 +0000 (11:42 +0100)]
GDBusConnection: document that this object is (meant to be) thread-safe
Bug: https://bugzilla.gnome.org/show_bug.cgi?id=661992
Signed-off-by: Simon McVittie <simon.mcvittie@collabora.co.uk>
Reviewed-by: David Zeuthen <davidz@redhat.com>
Simon McVittie [Fri, 21 Oct 2011 14:46:00 +0000 (15:46 +0100)]
GDBusConnection: make the closed flag atomic (but still lock to write)
Strictly speaking, neither of the two uses that aren't under the lock
*needs* to be atomic, but it seems better to be obviously correct (and
we save another 4 bytes of struct).
One of these uses is in g_dbus_connection_is_closed(), any use of which
is inherently a race condition anyway.
The other is g_dbus_connection_flush_sync, which as far as I can tell
just needs a best-effort check, to not waste effort on a connection that
has been closed for a while (but I could be wrong).
I removed the check for the closed flag altogether in
g_dbus_connection_send_message_with_reply_unlocked, because it turns out
to be redundant with one in g_dbus_connection_send_message_unlocked,
which is called immediately after.
g_dbus_connection_close_sync held the lock to check the closed flag,
which is no longer needed.
As far as I can tell, the only reason why the lock is still desirable
when setting the closed flag is so that remove_match_rule can't fail
by racing with close notification from the worker thread - but
on_worker_closed needs to hold the lock anyway, to deal with other
data structures, so there's no point in trying to eliminate the
requirement to hold the lock.
Bug: https://bugzilla.gnome.org/show_bug.cgi?id=661992
Signed-off-by: Simon McVittie <simon.mcvittie@collabora.co.uk>
Reviewed-by: David Zeuthen <davidz@redhat.com>
Simon McVittie [Fri, 21 Oct 2011 15:24:29 +0000 (16:24 +0100)]
GDBusConnection: document which properties are protected by @lock
Also, a few that don't need to be.
Bug: https://bugzilla.gnome.org/show_bug.cgi?id=661992
Signed-off-by: Simon McVittie <simon.mcvittie@collabora.co.uk>
Reviewed-by: David Zeuthen <davidz@redhat.com>
Simon McVittie [Mon, 17 Oct 2011 18:27:23 +0000 (19:27 +0100)]
GDBusConnection: access the exit-on-close flag atomically
This isn't strictly necessary, because in every location where it's
checked, if the reading thread misses an update from another thread,
it's indistinguishable from the reading thread having been scheduled
before the writing thread, which is an unavoidable race condition that
callers need to cope with anyway. On the other hand, merging exit_on_close
into atomic_flags gives the least astonishing semantics to library users
and saves 4 bytes of struct, and if you're accessing exit-on-close often
enough for it to be a performance concern, you're probably doing it wrong.
Bug: https://bugzilla.gnome.org/show_bug.cgi?id=661992
Signed-off-by: Simon McVittie <simon.mcvittie@collabora.co.uk>
Reviewed-by: David Zeuthen <davidz@redhat.com>
Simon McVittie [Mon, 17 Oct 2011 12:27:03 +0000 (13:27 +0100)]
Annotate GDBusConnection private functions with thread/lock status
The thread shared between all GDBusWorker instances was variously called
the "worker thread" or "message handler thread", which I mostly changed to
"the GDBusWorker thread" to avoid ambiguity.
Bug: https://bugzilla.gnome.org/show_bug.cgi?id=661992
Signed-off-by: Simon McVittie <simon.mcvittie@collabora.co.uk>
Reviewed-by: David Zeuthen <davidz@redhat.com>
Yaron Shahrabani [Sat, 22 Oct 2011 10:39:15 +0000 (12:39 +0200)]
Updated Hebrew translation.
Matej Urbančič [Fri, 21 Oct 2011 19:24:06 +0000 (21:24 +0200)]
Updated Slovenian translation
Ryan Lortie [Fri, 21 Oct 2011 18:05:10 +0000 (14:05 -0400)]
GDBusMessage: fix leak of GError
As part of the deserialisation process of a zero-length array in the
DBus wire format, parse_value_from_blob() recursively calls itself with
the expectation of failing (as can be seen by the assert immediately
following).
It passes &local_error to this always-failing call and then fails to
free it (indeed, to use it at all). The result is that the GError is
leaked.
Fix it by passing in NULL instead, so that the GError is never created
in the first place.
https://bugzilla.gnome.org/show_bug.cgi?id=662411
Ryan Lortie [Fri, 21 Oct 2011 19:00:24 +0000 (15:00 -0400)]
GVariant format string docs: fix maybe types
The documentation for maybe types failed to mention 'a' as one of the
types that was handled with a single pointer for which NULL means
"nothing". Correct that omission.
Problem caught by Shaun McCance.
Matthias Clasen [Fri, 21 Oct 2011 18:57:47 +0000 (14:57 -0400)]
Make g_ascii_strtod behave as documented
The docs explicitly state that we reset errno, so lets do that,
even if we just wrap strtod_l. Also move the argument check
out of the ifdef.
https://bugzilla.gnome.org/show_bug.cgi?id=662398
Simon McVittie [Wed, 19 Oct 2011 16:23:30 +0000 (17:23 +0100)]
GDBusConnection: check for initializedness in most public API
The only exceptions are those of the trivial getters/setters that don't
already need the initialization check for its secondary role as a memory
barrier (this is consistent with GSocket, where trivial getters/setters
don't check):
* g_dbus_connection_set_exit_on_close
* g_dbus_connection_get_exit_on_close
* g_dbus_connection_is_closed
g_dbus_connection_set_exit_on_close needs to be safe for
use before initialization anyway, so it can be set at construct-time.
Bug: https://bugzilla.gnome.org/show_bug.cgi?id=661689
Signed-off-by: Simon McVittie <simon.mcvittie@collabora.co.uk>
Reviewed-by: David Zeuthen <davidz@redhat.com>
Simon McVittie [Thu, 20 Oct 2011 19:02:02 +0000 (20:02 +0100)]
GDBusConnection: document use while uninitialized as undefined behaviour
Bug: https://bugzilla.gnome.org/show_bug.cgi?id=661689
Bug: https://bugzilla.gnome.org/show_bug.cgi?id=662208
Signed-off-by: Simon McVittie <simon.mcvittie@collabora.co.uk>
Reviewed-by: David Zeuthen <davidz@redhat.com>
Simon McVittie [Thu, 20 Oct 2011 12:12:26 +0000 (13:12 +0100)]
GDBusConnection: check for initialization where needed for thread-safety
Also document which fields require such a check in order to have correct
threading semantics.
This usage doesn't matches the GInitable documentation, which suggests
use of a GError - but using an uninitialized GDBusConnection is
programming error, and not usefully recoverable. (The GInitable
documentation may have been a mistake - GNOME#662208.) Also, not all of
the places where we need it can raise a GError.
The check serves a dual purpose: it turns a non-deterministic crash into
a deterministic critical warning, and is also a memory barrier for
thread-safety. All of these functions dereference or return fields that
are meant to be protected by FLAG_INITIALIZED, so they could crash or
return an undefined value to their caller without this, if called from a
thread that isn't the one that called initable_init() (although I can't
think of any way to do that without encountering a memory barrier,
undefined behaviour, or a race condition that leads to undefined
behaviour if the non-initializing thread wins the race).
One exception is that initable_init() itself makes a synchronous call.
We deal with that by passing new internal flags up the call stack, to
reassure g_dbus_connection_send_message_unlocked() that it can go ahead.
Bug: https://bugzilla.gnome.org/show_bug.cgi?id=661689
Bug: https://bugzilla.gnome.org/show_bug.cgi?id=661992
Signed-off-by: Simon McVittie <simon.mcvittie@collabora.co.uk>
Reviewed-by: David Zeuthen <davidz@redhat.com>
Simon McVittie [Fri, 21 Oct 2011 15:02:22 +0000 (16:02 +0100)]
GDBusConnection: replace is_initialized with an atomic flag
The comment implied that even failed initialization would set
is_initialized = TRUE, but this wasn't the case - failed initialization
would only set initialization_error, and it was necessary to check both.
It turns out the documented semantics are nicer than the implemented
semantics, since this lets us use atomic operations, which are also
memory barriers, to avoid needing separate memory barriers or locks
for initialization_error (and other members that are read-only after
construction).
I expect to need more than one atomically-accessed flag to fix thread
safety, so instead of a minimal implementation I've turned is_initialized
into a flags word.
Signed-off-by: Simon McVittie <simon.mcvittie@collabora.co.uk>
Bug: https://bugzilla.gnome.org/show_bug.cgi?id=661689
Bug: https://bugzilla.gnome.org/show_bug.cgi?id=661992
Reviewed-by: David Zeuthen <davidz@redhat.com>
Chun-wei Fan [Thu, 20 Oct 2011 06:42:51 +0000 (14:42 +0800)]
Fix VS 2008 property sheet
Missed a required ';'-sorry about this. DOH! :|
Matthias Clasen [Thu, 20 Oct 2011 01:03:43 +0000 (21:03 -0400)]
Deprecate g_atexit
This function was just not a good idea to begin with.
Its documentation gives plenty of reason not to use it.
Cosimo Cecchi [Tue, 18 Oct 2011 14:48:31 +0000 (10:48 -0400)]
desktop-app-info: annotate g_desktop_app_info_get_keywords()
https://bugzilla.gnome.org/show_bug.cgi?id=662096
Ryan Lortie [Wed, 19 Oct 2011 19:26:25 +0000 (15:26 -0400)]
more NEWS
Ryan Lortie [Wed, 19 Oct 2011 19:07:22 +0000 (15:07 -0400)]
The usual docs unbreaking...
Ryan Lortie [Wed, 19 Oct 2011 14:11:54 +0000 (10:11 -0400)]
configure.ac: join the cargo-cult of _cv_
For some reason, the setting of g_atomic_lock_free wasn't making it down
to the lower part of the configure script where glibconfig.h was being
generated when building using mingw32-configure.
If we prefix glib_cv_ to the start of the variable name (like everyone
else is doing) then it magically starts working.
I love you, automake.
Simon McVittie [Tue, 18 Oct 2011 15:19:43 +0000 (16:19 +0100)]
gdbus-non-socket test: avoid use of a GMainContext across a fork
See https://bugzilla.gnome.org/show_bug.cgi?id=658999 for why this would
be bad.
Bug: https://bugzilla.gnome.org/show_bug.cgi?id=662100
Signed-off-by: Simon McVittie <simon.mcvittie@collabora.co.uk>
Reviewed-by: David Zeuthen <davidz@redhat.com>
Simon McVittie [Tue, 18 Oct 2011 15:18:21 +0000 (16:18 +0100)]
gdbus-exit-on-close test: cover more possibilities
We didn't previously test anything except the implicit default of TRUE.
Now we test implicit TRUE, explicit TRUE, explicit FALSE, and
disconnecting at the local end (which regressed while fixing Bug #651268).
Also avoid some questionable use of a main context, which fell foul of
Bug #658999 and caused this test to be disabled in master.
Bug: https://bugzilla.gnome.org/show_bug.cgi?id=662100
Bug-NB: NB#287088
Signed-off-by: Simon McVittie <simon.mcvittie@collabora.co.uk>
Reviewed-by: David Zeuthen <davidz@redhat.com>
Simon McVittie [Wed, 19 Oct 2011 09:49:56 +0000 (10:49 +0100)]
GDBusWorker: if a read was cancelled it means we closed the connection
This was a regression caused by my previous work on GDBusWorker thread-safety
(Bug #651268). The symptom is that if you disconnect a GDBusConnection
locally, the default implementation of GDBusConnection::closed
terminates your process, even though it shouldn't do that for
locally-closed connections; this is because GDBusWorker didn't think a
cancelled read was a local close.
Bug: https://bugzilla.gnome.org/show_bug.cgi?id=662100
Bug-NB: NB#287088
Signed-off-by: Simon McVittie <simon.mcvittie@collabora.co.uk>
Reviewed-by: David Zeuthen <davidz@redhat.com>
Simon McVittie [Tue, 18 Oct 2011 15:27:21 +0000 (16:27 +0100)]
Revert "Disable two GDBus tests"
This reverts commit
05ef173466e32d8b3d212803e4a72239913a362d.
Bug: https://bugzilla.gnome.org/show_bug.cgi?id=662100
Signed-off-by: Simon McVittie <simon.mcvittie@collabora.co.uk>
Reviewed-by: David Zeuthen <davidz@redhat.com>
Ryan Lortie [Wed, 19 Oct 2011 03:26:00 +0000 (23:26 -0400)]
Fix bug in g_static_rec_mutex_unlock_full()
pthreads doesn't implement the _lock_full() and _unlock_full() calls on
recursive mutexes so we don't have it on GRecMutex either. Now that
we're using GRecMutex to implement GStaticRecMutex, we have to fake it
by keeping an internal counter of the number of locks and calling
g_rec_mutex_unlock() the appropriate number of times.
The code to do this looked like:
depth = mutex->depth;
while (mutex->depth--)
g_rec_mutex_unlock (rm);
return depth;
which unfortunately did one last decrement after mutex->depth was
already zero (leaving it equal to -1).
When locked the next time, the count would then increase from -1 to 0
and then the next _unlock_full() call would not do any calls to
g_rec_mutex_unlock(), leading to a deadlock.
https://bugzilla.gnome.org/show_bug.cgi?id=661914
Ryan Lortie [Tue, 18 Oct 2011 20:21:50 +0000 (16:21 -0400)]
gatomic: introduce G_ATOMIC_LOCK_FREE
We clean up the detection of if we should do 'real' atomic operations or
mutex-emulated ones with the introduction of a new (public) macro:
G_ATOMIC_LOCK_FREE. If defined, our atomic operations are guaranteed to
be done in hardware.
We need to use __GCC_HAVE_SYNC_COMPARE_AND_SWAP_4 to determine if our
compiler supports GCC-style atomic operations from the gatomic.h header
because we might be building a program against GLib using a different
set of compiler options (or a different compiler) than was used to build
GLib itself.
Unfortunately, this macro is not available on clang, so it has currently
regressed to using the mutex emulation. A bug about that has been
opened here:
http://llvm.org/bugs/show_bug.cgi?id=11174