Christian Dywan [Tue, 20 Jul 2010 13:01:03 +0000 (15:01 +0200)]
Don't pretend to free lock if it was never allocated
Christian Dywan [Tue, 21 Sep 2010 12:43:39 +0000 (14:43 +0200)]
Always get current time in _dbus_connection_block_pending_call
Christian Dywan [Tue, 20 Jul 2010 09:53:15 +0000 (11:53 +0200)]
Take care to free windows_sid in cache_peer_loginfo_string
Christian Dywan [Tue, 20 Jul 2010 10:50:23 +0000 (12:50 +0200)]
Free envvar and args in pass_info in the oom case
Ralf Habacker [Tue, 28 Sep 2010 06:17:38 +0000 (08:17 +0200)]
When launching dbus-daemon be more verbose in error case.
Ralf Habacker [Mon, 27 Sep 2010 20:21:29 +0000 (22:21 +0200)]
Keep unix autolaunch functions in sync.
Ralf Habacker [Sun, 26 Sep 2010 18:05:11 +0000 (20:05 +0200)]
Added documentation for autolaunch transport.
Ralf Habacker [Fri, 27 Aug 2010 15:03:16 +0000 (17:03 +0200)]
Fixed case when no scope attribute is used.
Ralf Habacker [Fri, 27 Aug 2010 14:51:44 +0000 (16:51 +0200)]
Extended autolaunch protocol with scope attribute.
Christian Dywan [Thu, 23 Sep 2010 12:38:15 +0000 (14:38 +0200)]
Add return_if_fail watch != NULL in public DBusWatch API
Reviewed-by: Havoc Pennington <hp@pobox.com>
Signed-off-by: Will Thompson <will.thompson@collabora.co.uk>
Ralf Habacker [Fri, 10 Sep 2010 07:14:06 +0000 (09:14 +0200)]
Include cmake subdirectory in source distribution.
Marcus Brinkmann [Thu, 9 Sep 2010 12:15:45 +0000 (14:15 +0200)]
Fix typo in creating temp file.
Marcus Brinkmann [Mon, 6 Sep 2010 13:50:16 +0000 (15:50 +0200)]
Fix use of _dbus_make_file_world_readable.
Will Thompson [Wed, 8 Sep 2010 11:11:50 +0000 (12:11 +0100)]
Explicitly require the daemon to respect destination=''
Ralf Habacker [Wed, 8 Sep 2010 07:13:42 +0000 (09:13 +0200)]
Unix compile fix.
Lennart Poettering [Tue, 7 Sep 2010 00:57:54 +0000 (02:57 +0200)]
post release version bump
Havoc Pennington [Tue, 31 Aug 2010 16:00:26 +0000 (12:00 -0400)]
fix some "unused result" warnings in dbus-launch and dbus-spawn
Needed to avoid -Werror
Havoc Pennington [Tue, 31 Aug 2010 15:59:09 +0000 (11:59 -0400)]
Change failure of setsockopt SO_REUSEADDR to be a warning, not set an error
It was previously setting an error but then continuing, which is
not allowed since errors can't be "piled up"
Lennart Poettering [Mon, 6 Sep 2010 12:52:05 +0000 (14:52 +0200)]
build-sys: bump version number for the 1.4.0 release
Kay Sievers [Fri, 3 Sep 2010 13:12:19 +0000 (15:12 +0200)]
systemd: enable the dbus service unconditionally
Since D-Bus is a core part of the OS don't make this service optional.
Lennart Poettering [Sun, 5 Sep 2010 21:43:24 +0000 (23:43 +0200)]
autoconf: update lt version info in preparation for 0.4.2
Lennart Poettering [Sun, 5 Sep 2010 21:37:25 +0000 (23:37 +0200)]
NEWS: update for 1.4.0
Lennart Poettering [Sun, 5 Sep 2010 21:30:47 +0000 (23:30 +0200)]
systemd: update sd-daemon.[ch] from upstream systemd
Colin Walters [Fri, 3 Sep 2010 18:46:53 +0000 (14:46 -0400)]
Merge branch 'dbus-1.2'
Conflicts:
dbus/dbus-sysdeps-unix.c
Colin Walters [Fri, 3 Sep 2010 14:18:25 +0000 (10:18 -0400)]
Make dbus-uuidgen atomic
A Red Hat QA engineer hit in practice a race condition in dbus-uuidgen
where it could leave an empty file.
dbus-uuidgen (_dbus_create_uuid_file_exclusively) formerly created an
empty file in the path to the uuid, then filled it in. At some point,
the internal libdbus _dbus_string_save_to_file became atomic on Unix
at least (doing the save to temp file, fsync(), rename() dance).
So _dbus_create_uuid_file_exclusively doesn't need to create the file
beforehand anymore. However, it *does* need the file to be
world-readable, unlike all other consumers of
_dbus_string_save_to_file. So add a "world_readable" argument.
Lennart Poettering [Sat, 7 Aug 2010 19:04:59 +0000 (21:04 +0200)]
sd-daemon: update from upstream
This updates the sd-daemon.[ch] copy from upstream, which makes sure
thet the code compiles even if SOCK_CLOEXEC is not supported.
Romain Pokrzywka [Wed, 11 Aug 2010 12:05:31 +0000 (14:05 +0200)]
Cleanup: removed unused code; fixed Winsock error handling; fixed code indentation.
Ralf Habacker [Tue, 10 Aug 2010 09:26:42 +0000 (11:26 +0200)]
Fixed wince build problem reported by aheinecke at intevation dot de.
Ralf Habacker [Tue, 10 Aug 2010 07:17:33 +0000 (09:17 +0200)]
Fixed bug not creating bus/messagebus-config.
Yaakov Selkowitz [Mon, 9 Aug 2010 07:01:52 +0000 (09:01 +0200)]
Delete stale pid file on cygwin.
Yaakov Selkowitz [Mon, 9 Aug 2010 06:50:33 +0000 (08:50 +0200)]
Avoid -Werror problems on cygwin.
Yaakov Selkowitz [Mon, 9 Aug 2010 06:32:04 +0000 (08:32 +0200)]
Applied patches from cygwin port.
Romain Pokrzywka [Sun, 8 Aug 2010 21:54:35 +0000 (23:54 +0200)]
fix socket descriptor leak in _dbus_connect_tcp_socket_with_nonce
Marcus Brinkmann [Fri, 11 Jun 2010 00:04:36 +0000 (02:04 +0200)]
Wrap OutputDebugString for Windows CE builds, and add cast for MSVC.
Christian Dywan [Thu, 15 Jul 2010 11:20:04 +0000 (13:20 +0200)]
Add unit tests for dbus_message_get_path_decomposed
Christian Dywan [Thu, 15 Jul 2010 11:18:33 +0000 (13:18 +0200)]
Assert in _dbus_decompose_path if path is NULL or len is wrong
Lennart Poettering [Tue, 13 Jul 2010 00:50:02 +0000 (02:50 +0200)]
systemd: update systemd glue code
Lennart Poettering [Mon, 12 Jul 2010 23:35:03 +0000 (01:35 +0200)]
systemd: remove dbus pid file before starting dbus
Lennart Poettering [Sat, 10 Jul 2010 00:31:04 +0000 (02:31 +0200)]
man: add documentation for four new switches to dbus-daemon
This adds explanations for --address and --systemd-activation which have
recently been added to the daemon. Also adds coumentation for --nofork
and --introspect which has been previously forgotten.
Lennart Poettering [Sat, 10 Jul 2010 00:22:59 +0000 (02:22 +0200)]
systemd: add a nice description string to the .socket file
Lennart Poettering [Fri, 9 Jul 2010 18:37:24 +0000 (20:37 +0200)]
systemd: install systemd socket and service file by default
Lennart Poettering [Fri, 9 Jul 2010 19:47:13 +0000 (21:47 +0200)]
build-sys: update pkg-config.m4
Lennart Poettering [Wed, 2 Jun 2010 02:08:39 +0000 (04:08 +0200)]
activation: optionally, use systemd for system bus activation
Colin Walters [Fri, 9 Jul 2010 15:48:53 +0000 (11:48 -0400)]
Fix use of $servicename in status
See https://bugzilla.redhat.com/show_bug.cgi?id=521797
We have two init scripts, and commit
05283235632bc8891aaa3047563cb4
patched the "wrong" (i.e. slackware) one. Patch one that ships
with real operating systems too.
Lennart Poettering [Tue, 25 May 2010 00:27:44 +0000 (02:27 +0200)]
bus: add --address switch
This is allows overriding of the listening address on the command line,
which is particularly useful for systemd socket-based activation.
Lennart Poettering [Tue, 25 May 2010 00:25:04 +0000 (02:25 +0200)]
systemd: add systemd-style socket-activation
This is a pretty straightforward implementation of systemd-style socket
activation using the reference implementation of the
$LISTEN_FDS/$LISTEN_PID env var parsing as supplied by systemd upstream.
Sledz [Tue, 29 Jun 2010 21:04:15 +0000 (23:04 +0200)]
sysdeps-unix: enable address reuse on TCP sockets
This helps to avoid 'Address already in use' problems if server has been shut
down, and then restarted while sockets are still active on its port.
Signed-off-by: Steffen Sledz <sledz@dresearch.de>
Thiago Macieira [Tue, 22 Jun 2010 13:13:23 +0000 (15:13 +0200)]
Fix the reentrancy issue reported on bug 17754.
Patch based on patch by Havoc Pennington, with the references that
this is temporary removed.
Patch based on one from Olivier Hochreutiner <olivier.hochreutiner
gmail.com>
* dbus/dbus-connection.c (protected_change_timeout): remove the
elaborate nonworking hack to try to drop locks and just keep the
locks; this isn't right either, but at least is correct, though
it puts restrictions on apps.
* dbus/dbus-connection.c (protected_change_watch): make the same
change as for timeouts
* dbus/dbus-connection.c (dbus_connection_set_timeout_functions):
don't drop the lock here; add documentation of the problem to API
docs
(dbus_connection_set_watch_functions): same
* dbus/dbus-connection.c (dbus_connection_get_data)
(dbus_connection_set_data): introduce a separate slot_mutex
protecting connection->slot_list so these two functions can be
called inside watch and timeout functions. Not sure this
is going to be a good idea.
* dbus/dbus-connection.c (dbus_connection_unref)
(dbus_connection_ref): avoid using connection lock in ref/unref
so these can also be used in watch and timeout functions
Lennart Poettering [Wed, 23 Jun 2010 14:37:58 +0000 (16:37 +0200)]
build-sys: open 1.3.2 development cycle
Lennart Poettering [Wed, 23 Jun 2010 14:13:43 +0000 (16:13 +0200)]
build-sys: add make target for updating AUTHORS file
Lennart Poettering [Wed, 23 Jun 2010 02:00:28 +0000 (04:00 +0200)]
build-sys: prepare release 1.3.1
Lennart Poettering [Wed, 23 Jun 2010 01:14:27 +0000 (03:14 +0200)]
pkg-config: add a couple of additional variables to .pc
We already export the session services dir in the .pc file, add the
system services dir too, as well as the interfaces dir, so that packages
can drop their files into the right place.
Eventually we need to move the system-services dir to a place that is
relocatable to /lib/dbus, and this change makes the location queriable
for other projects, as a first step towards that goal.
Will Thompson [Tue, 22 Jun 2010 16:25:20 +0000 (17:25 +0100)]
Merge branch 'dbus-1.2'
Christian Dywan [Mon, 21 Jun 2010 05:13:47 +0000 (07:13 +0200)]
Implement doctype check in the trivial parser
This is needed for the libxml parser backend.
Thiago Macieira [Tue, 22 Jun 2010 12:54:38 +0000 (14:54 +0200)]
Revert "Bug 17754 - Fixed race condition in protected_change_timeout."
This reverts commit
b82ff17d7dd9689ab664e4a6c2a86c6f3dbfba69 and
9280395330c833dd98d0811d31349c26a84b2f6a.
Conflicts:
dbus/dbus-connection.c
Will Thompson [Thu, 17 Jun 2010 14:05:46 +0000 (15:05 +0100)]
Clarify GetConnectionUnixUser documentation.
(This makes it match the freshly-documented GetConnectionUnixProcessID.)
Will Thompson [Sun, 10 Jan 2010 11:27:33 +0000 (11:27 +0000)]
Document ListQueuedOwners in the spec
Will Thompson [Tue, 4 Aug 2009 10:22:33 +0000 (11:22 +0100)]
Document GetConnectionUnixProcessID in the spec
Will Thompson [Sun, 10 Jan 2010 11:11:40 +0000 (11:11 +0000)]
Fix a broken <xref/>
Will Thompson [Thu, 17 Jun 2010 14:14:22 +0000 (15:14 +0100)]
Revert "Bug 28460 - Refactored dbus configuration access."
This reverts commit
6f9077ee870ad02119facf83d1293301b4535c3b, which
broke the build on Unix (see previous patch) and introduced a bunch of
warnings.
Will Thompson [Thu, 17 Jun 2010 13:45:16 +0000 (14:45 +0100)]
Fix dbus-sysdeps.h on amd64, again.
HAVE_STDINT_H is defined in config.h, but that's not much use if
config.h isn't included. The new file dbus/dbus-config.c includes this
header without including config.h first, so fails to build.
Since dbus-sysdeps.h is internal, we can include config.h with impunity,
rather than relying on our callers to do so. Also, there's no need to
include stdint.h twice.
Ralf Habacker [Mon, 14 Jun 2010 16:27:04 +0000 (18:27 +0200)]
Bug 28460 - Refactored dbus configuration access.
Libdbus uses several config variables. On unix these settings are read from
environment variables by using _dbus_getenv.
On other platforms like wince there are no environment variables available and
_dbus_getenv needs an emulation for those plattforms (see
dbus/dbus-sysdeps-wince-glue.c)
To cleanup this emulation the appended patch adds a config api by adding
_dbus_config_... functions.
Also having all client config related functions listed in one header file
provides a good overview about which config attributes are available.
The default implementation retrieves the config values from environment
variables. For other os this could be easily extended or replaced by.
Ralf Habacker [Fri, 11 Jun 2010 22:02:57 +0000 (00:02 +0200)]
Use dbus provided wait function.
Plácido Revilla [Fri, 11 Jun 2010 16:31:19 +0000 (18:31 +0200)]
Bug 17754 - Fixed race condition in protected_change_timeout.
Johannes Carlsson [Wed, 9 Jun 2010 09:58:06 +0000 (11:58 +0200)]
Corrected thread problem causing some calls to hang for 25s
Since the connection lock is released for a short while in
_dbus_connection_acquire_io_path there can already be a method return
received by another thread. The fix is to do an extra check after the
I/O path has been aquired both.
Approved-by: Thiago Macieira
Ralf Habacker [Fri, 11 Jun 2010 08:52:43 +0000 (10:52 +0200)]
Do not set DBUS_SESSION_BUS_DEFAULT_ADDRESS unconditional in code, it is already defined in build system.
Ralf Habacker [Fri, 11 Jun 2010 08:30:35 +0000 (10:30 +0200)]
Keep cmake default session and system bus address setup in sync with autotools.
Christian Dywan [Wed, 5 May 2010 11:58:24 +0000 (13:58 +0200)]
Fix x64 build after WinCE intptr_t usage changes.
Johannes Carlsson [Wed, 9 Jun 2010 09:58:06 +0000 (11:58 +0200)]
Corrected thread problem causing some calls to hang for 25s
Since the connection lock is released for a short while in
_dbus_connection_acquire_io_path there can already be a method return
received by another thread. The fix is to do an extra check after the
I/O path has been aquired both.
Approved-by: Thiago Macieira
Ralf Habacker [Mon, 7 Jun 2010 13:46:33 +0000 (15:46 +0200)]
Fixed compiler warnings.
Ralf Habacker [Mon, 7 Jun 2010 12:17:13 +0000 (14:17 +0200)]
Minor include file fix.
When including dbus/dbus-errors.h to get error related functions,
the related error constants should be available too.
Ralf Habacker [Mon, 7 Jun 2010 12:23:44 +0000 (14:23 +0200)]
Keep in sync with autotools.
Ralf Habacker [Mon, 7 Jun 2010 11:56:06 +0000 (13:56 +0200)]
Keep in sync with autotools.
Will Thompson [Fri, 21 May 2010 10:48:28 +0000 (11:48 +0100)]
Merge branch 'dbus-1.2'
Conflicts:
bus/main.c
configure.in
Will Thompson [Thu, 20 May 2010 15:08:33 +0000 (16:08 +0100)]
kqueue set_watched_dirs: fix termination condition
num_fds is the number of elements of dirs currently in use. This bug
meant that encountering a previously un-watched directory would cause j
to increment forever, and so dirs[j] would eventually segfault.
(I've checked the corresponding code for inotify, and it's correct. I
wonder if some of the duplication could be eliminated.)
Thanks to Pablo Martí Gamboa <pmarti@warp.es> for reporting this issue!
Thiago Macieira [Tue, 18 May 2010 12:59:06 +0000 (14:59 +0200)]
Fix building on Unix
Luiz Augusto Von Dentz [Wed, 12 May 2010 18:01:00 +0000 (21:01 +0300)]
Fix memory allocation when reusing message from cache
Signed-off-by: Thiago Macieira <thiago@kde.org>
David Zeuthen [Thu, 13 May 2010 14:25:22 +0000 (10:25 -0400)]
Add an extra parameter to the PropertiesChanged() signal
It was pointed out on the mailing list that it would be useful to know
that a given property has changed without conveying its value. Because
without this parameter a true_no_value property could change, however
there is no way for a client-side proxy to know _what_ property it was
(only that some property changed).
With the parameter, however, a client-side proxy can reliably discard
a cached property value.
Also rename the "true_no_value" to "invalidates" as the spec is now
using this language.
Also allow using the annotation in the enclosed interface name.
Also rename the annotation name so it uses Property in its name
instead of Properties. This is to be more consistent with the existing
org.freedesktop.DBus.Method.NoReply annotation which uses Method, not
Methods.
Signed-off-by: David Zeuthen <davidz@redhat.com>
David Zeuthen [Fri, 7 May 2010 16:50:47 +0000 (12:50 -0400)]
Add an optional PropertiesChanged() signal
Some notes about this new signal
- The PropertiesChanged() signal is optional. An application can
convey support for this signal by either including or excluding it
from the returned introspection data much like apps not supporting
(or predating) the GetAll() method does not include GetAll() in the
introspection data.
- An object can use PropertiesChanged() but opt out of using it for
one or more properties by using the
org.freedesktop.DBus.Properties.EmitsChangedSignal
annotation on the properties in question
- Applications can start using this new signal without breaking
compatibility with clients relying on existing D-Bus API.
The intent of the patch is simply to standardize existing behavior
- EggDBus has a very similar signal called EggDBusPropertiesChanged()
(also on the org.freedesktop.DBus.Properties interface)
- NetworkManager has a PropertiesChanged() signal on each different
interface (e.g. not org.fd.D.P) that it implements
- GDBus, an implementation of the D-Bus protocol in GLib, already
implements this signal
Signed-off-by: David Zeuthen <davidz@redhat.com>
Marcus Brinkmann [Wed, 12 May 2010 12:47:18 +0000 (14:47 +0200)]
Fix _IOLBF for Windows CE and also define _IONBF if needed.
Ralf Habacker [Mon, 10 May 2010 20:56:11 +0000 (22:56 +0200)]
Fixed assert raised with msvc 2008 on windows.
Marcus Brinkmann [Wed, 5 May 2010 15:05:24 +0000 (17:05 +0200)]
Fix infinite recursion on Windows CE.
Marcus Brinkmann [Mon, 3 May 2010 17:16:35 +0000 (19:16 +0200)]
Fix GetModuleFileNameA and RegQueryValueExA for Windows CE.
Ray Strode [Mon, 10 May 2010 14:12:45 +0000 (10:12 -0400)]
Add clarification to spec about UpdateActivationEnvironment
UpdateActivationEnvironment takes a a{ss}. This means only
valid UTF-8 can be used. Environment variables are normally
ascii, but in theory have no specific encoding to them. This
means that certain valid environment variables can't be sent
to the bus for updating its activation environment.
This commit just adds a note to the spec explaining this
restriction.
Marcus Brinkmann [Wed, 21 Apr 2010 12:25:29 +0000 (14:25 +0200)]
Move definition of _WIN32_WCE to configure.in (cmake defines it already).
Romain Pokrzywka [Tue, 4 May 2010 13:41:35 +0000 (15:41 +0200)]
fix {u}intptr_t usage on wince with msvc
Ralf Habacker [Wed, 5 May 2010 05:56:50 +0000 (07:56 +0200)]
Send system log messages to debug port for now.
Ralf Habacker [Mon, 3 May 2010 10:06:22 +0000 (12:06 +0200)]
Added missing implementation for _dbus_system_logv().
Ralf Habacker [Mon, 3 May 2010 09:56:40 +0000 (11:56 +0200)]
Cleanup debug messages.
Marcus Brinkmann [Thu, 29 Apr 2010 23:44:28 +0000 (01:44 +0200)]
Defensively protect against bogus argv argument on W32CE.
Ralf Habacker [Thu, 29 Apr 2010 14:21:05 +0000 (16:21 +0200)]
Disabled launch helper activation tests on windows.
Launch helper application isn't supported on windows yet.
Ralf Habacker [Thu, 29 Apr 2010 14:10:31 +0000 (16:10 +0200)]
Be more silent with porting problems.
Fridrich Štrba [Mon, 3 May 2010 08:12:05 +0000 (10:12 +0200)]
Don't truncate pointers on Windows x64 platform
Fridrich Štrba [Wed, 28 Apr 2010 13:09:14 +0000 (15:09 +0200)]
Change to assume executable extension when none is given (except on wince)
Fridrich Štrba [Wed, 28 Apr 2010 12:19:00 +0000 (14:19 +0200)]
Pass the environment to CreateProcessA correctly and be more defensive
Ralf Habacker [Tue, 27 Apr 2010 09:22:32 +0000 (11:22 +0200)]
Merge branch 'master' of ssh://git.freedesktop.org/git/dbus/dbus
Ralf Habacker [Tue, 27 Apr 2010 09:19:56 +0000 (11:19 +0200)]
Minor name spelling fix.
David Zeuthen [Mon, 26 Apr 2010 22:00:07 +0000 (18:00 -0400)]
Fix spelling error
This broke 'make install' since dbus-daemon-launch-helper didn't get
installed.
Signed-off-by: David Zeuthen <davidz@redhat.com>
Tomas Pelka [Fri, 23 Apr 2010 16:00:47 +0000 (12:00 -0400)]
[legacy init script] Fix the use of $servicename
Ralf Habacker [Thu, 22 Apr 2010 09:51:23 +0000 (11:51 +0200)]
Added cmake cross compile wrapper.