Ralf Habacker [Mon, 15 Aug 2011 15:33:42 +0000 (17:33 +0200)]
Merge branch 'dbus-1.4'
Simon McVittie [Mon, 15 Aug 2011 08:52:04 +0000 (09:52 +0100)]
Fix compilation of (no-op) alignment assertions on non-gcc
do {} while (0) isn't valid at file scope, so the non-gcc code path
was broken; in particular, this affected MSVC.
Reviewed-by:Ralf Habacker <ralf.habacker@freenet.de>
Simon McVittie [Mon, 15 Aug 2011 09:26:48 +0000 (10:26 +0100)]
Merge branch 'dbus-1.4'
Simon McVittie [Mon, 15 Aug 2011 09:26:42 +0000 (10:26 +0100)]
NEWS
Chris Mayo [Mon, 15 Aug 2011 09:23:04 +0000 (10:23 +0100)]
Install diagrams in htmldir
diagram.png and diagram.svg are linked from dbus-tutorial.html but recent
changes to doc/Makefile.am mean that if htmldir is set with configure they are
installed such that the relative link from dbus-tutorial.html does not work.
Reviewed-by: Simon McVittie <simon.mcvittie@collabora.co.uk>
Simon McVittie [Mon, 15 Aug 2011 08:56:03 +0000 (09:56 +0100)]
Merge branch 'dbus-1.4'
Ralf Habacker [Sat, 13 Aug 2011 20:08:26 +0000 (22:08 +0200)]
Fix compilation on MSVC, which doesn't understand "inline" with its C99 meaning.
This fix is limited to msvc compilers only.
Reviewed-by:Simon McVittie <simon.mcvittie@collabora.co.uk>
Simon McVittie [Fri, 12 Aug 2011 09:45:24 +0000 (10:45 +0100)]
NEWS
Simon McVittie [Fri, 12 Aug 2011 09:45:15 +0000 (10:45 +0100)]
dbus-daemon-eavesdrop test: apply the same Windows fixes as dbus-daemon test
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=40003
Simon McVittie [Fri, 5 Aug 2011 12:14:54 +0000 (13:14 +0100)]
Fix linking of libdbus-testutils.la with Automake targeting Windows
This regressed in commit
9875480ed - libdbus-testutils.la needs static
linking too.
Simon McVittie [Fri, 12 Aug 2011 09:26:13 +0000 (10:26 +0100)]
Merge branch 'dbus-1.4'
Simon McVittie [Fri, 12 Aug 2011 09:26:04 +0000 (10:26 +0100)]
NEWS
Simon McVittie [Fri, 5 Aug 2011 12:55:09 +0000 (13:55 +0100)]
Fix compiler warnings when cross-compiling dbus-daemon test for mingw32
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=40003
Reviewed-by: Colin Walters <walters@verbum.org>
Simon McVittie [Thu, 11 Aug 2011 10:54:52 +0000 (11:54 +0100)]
Merge branch 'dbus-1.4'
Conflicts:
NEWS
Simon McVittie [Thu, 11 Aug 2011 10:53:45 +0000 (11:53 +0100)]
NEWS for 1.4.x
Simon McVittie [Tue, 21 Jun 2011 16:02:10 +0000 (17:02 +0100)]
dbus-marshal-basic: assert that the size and (on gcc) alignment of all types is as expected
The __alignof__() checks here will fail if compiling for an architecture
where more-than-natural alignment is needed.
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=39636
Reviewed-by: Will Thompson <will.thompson@collabora.co.uk>
Simon McVittie [Tue, 21 Jun 2011 15:19:21 +0000 (16:19 +0100)]
Add _DBUS_STATIC_ASSERT (based on GLib's G_STATIC_ASSERT) and use it
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=39636
Reviewed-by: Will Thompson <will.thompson@collabora.co.uk>
Simon McVittie [Thu, 11 Aug 2011 10:26:58 +0000 (11:26 +0100)]
Merge branch 'dbus-1.4', omitting the DBUS_INLINE stuff
fd.o #40000 had effectively already been fixed (slightly differently)
on master.
Conflicts:
cmake/config.h.cmake
Steve Grubb [Mon, 1 Aug 2011 18:28:27 +0000 (14:28 -0400)]
capng: Drop supplemental groups when switching to daemon user
This patch ensures the daemon process doesn't inherit any supplemental
groups for the root user from an administrator login via an init
script.
This is only an issue for pre-systemd systems.
https://bugzilla.redhat.com/show_bug.cgi?id=726953
Colin Walters [Mon, 1 Aug 2011 16:59:32 +0000 (12:59 -0400)]
autogen.sh: Don't run configure if NOCONFIGURE is set
See http://people.gnome.org/~walters/docs/build-api.txt
Ralf Habacker [Fri, 5 Aug 2011 22:36:42 +0000 (00:36 +0200)]
Clean up cygwin readme
Ralf Habacker [Fri, 5 Aug 2011 22:36:42 +0000 (00:36 +0200)]
Clean up cygwin readme
Ralf Habacker [Fri, 5 Aug 2011 21:26:41 +0000 (23:26 +0200)]
Clean author list
Ralf Habacker [Fri, 5 Aug 2011 21:52:49 +0000 (23:52 +0200)]
Reduce duplicates from authors list
Some duplicates are there by accident mostly because of the git
learning curve (as in my case). Other duplicates could be outdated
and may be candidates for additional mappings.
Ralf Habacker [Fri, 5 Aug 2011 21:52:49 +0000 (23:52 +0200)]
Reduce duplicates from authors list
Some duplicates are there by accident mostly because of the git
learning curve (as in my case). Other duplicates could be outdated
and may be candidates for additional mappings.
Ralf Habacker [Fri, 5 Aug 2011 21:26:41 +0000 (23:26 +0200)]
Clean author list
Ralf Habacker [Fri, 5 Aug 2011 21:01:10 +0000 (23:01 +0200)]
Win32 compile fix.
msvc compilers define 'inline' only for c++ code, so wrap it
with a platform independent DBUS_INLINE define in cmake
generated config.h.
Ralf Habacker [Fri, 5 Aug 2011 20:37:06 +0000 (22:37 +0200)]
Win32 compile fix.
Do not limit ACTION_RELOAD to unix only because it is used in the
crossplattform function handle_reload_watch().
Simon McVittie [Fri, 5 Aug 2011 14:07:24 +0000 (15:07 +0100)]
Merge branch 'dbus-1.4' and update NEWS further
Simon McVittie [Fri, 5 Aug 2011 14:03:39 +0000 (15:03 +0100)]
NEWS
Simon McVittie [Fri, 5 Aug 2011 13:52:37 +0000 (14:52 +0100)]
Merge branch 'dbus-1.4'
Conflicts:
bus/main.c
Simon McVittie [Thu, 4 Aug 2011 16:16:59 +0000 (17:16 +0100)]
Add a regression test for fd.o #38005
(The dbus-1.4 version.)
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=39836
Reviewed-by: Will Thompson <will.thompson@collabora.co.uk>
Simon McVittie [Thu, 4 Aug 2011 12:13:24 +0000 (13:13 +0100)]
relay test: don't leak the server's address
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=39836
Reviewed-by: Will Thompson <will.thompson@collabora.co.uk>
Simon McVittie [Wed, 29 Jun 2011 16:29:16 +0000 (17:29 +0100)]
_dbus_server_new_for_domain_socket: don't try to unlink abstract sockets
Our abstract socket names look like filenames (/tmp/dbus-MwozdykBNK or
whatever), so if we incorrectly unlink the abstract socket name, in
highly unlikely circumstances we could accidentally unlink a non-abstract
socket belonging to another process!
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=38656
Reviewed-by: Will Thompson <will.thompson@collabora.co.uk>
Simon McVittie [Wed, 29 Jun 2011 15:44:33 +0000 (16:44 +0100)]
bus signal_handler: call _exit in the unlikely event that the pipe is full or invalid
On OSs with abstract sockets, this is close enough. On OSs without
abstract sockets, this results in failing to clean up Unix sockets
in /tmp if someone has sent us thousands of SIGHUP signals since we
last entered the main loop - I think that's acceptable.
The reload pipe should never get closed, but if it is for some reason,
we want a SIGTERM after that to cause an exit too.
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=38656
Reviewed-by: Will Thompson <will.thompson@collabora.co.uk>
Simon McVittie [Wed, 29 Jun 2011 15:25:08 +0000 (16:25 +0100)]
bus signal_handler: comment why it's OK if the reload pipe gets full
Backported to dbus-1.4, originally part of commit
c7ef3ead558e147.
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=33336
Reviewed-by: Thiago Macieira <thiago@kde.org>
Simon McVittie [Wed, 29 Jun 2011 15:20:09 +0000 (16:20 +0100)]
bus signal_handler: don't pretend to be portable away from Unix
Signals are POSIX but not ISO C, so guard it with DBUS_UNIX.
dbus-sysdeps-util-win doesn't actually implement _dbus_set_signal_handler
anyway, so not compiling this code on non-Unix seems more honest.
Backported to dbus-1.4, originally part of commit
c7ef3ead558e147.
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=33336
Reviewed-by: Thiago Macieira <thiago@kde.org>
Simon McVittie [Wed, 29 Jun 2011 15:17:20 +0000 (16:17 +0100)]
Conditionalize inclusion of unistd.h
Windows doesn't have it, for instance.
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=38656
Reviewed-by: Will Thompson <will.thompson@collabora.co.uk>
Brian Cameron [Wed, 29 Jun 2011 15:15:51 +0000 (16:15 +0100)]
Handle SIGTERM again, to avoid leaving /tmp/dbus-XXX sockets on Solaris
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=38656
Reviewed-by: Simon McVittie <simon.mcvittie@collabora.co.uk>
Simon McVittie [Fri, 5 Aug 2011 13:18:27 +0000 (14:18 +0100)]
Merge branch 'dbus-1.4'
Simon McVittie [Wed, 27 Jul 2011 14:59:53 +0000 (15:59 +0100)]
cmake: always enable GNU and recent-POSIX extensions, like autoconf does
If the C library is (e)glibc, this allows use of POSIX, BSD, SVID, GNU,
etc., extensions to ISO C, regardless of using -ansi or not.
Not doing this broke the cmake build on Linux since commit
18b08180,
which added AC_USE_SYSTEM_EXTENSIONS to configure.ac (and removed
_GNU_SOURCE from files that use it) without also updating the cmake
build system. SO_PEERCRED is defined unconditionally, but struct ucred
is considered to be a GNU extension, so can't be used under _GNU_SOURCE.
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=29228
Reviewed-by: Will Thompson <will.thompson@collabora.co.uk>
Simon McVittie [Thu, 14 Jul 2011 17:48:26 +0000 (18:48 +0100)]
_dbus_lookup_session_address_launchd: don't define if not used
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=39231
Reviewed-by: Will Thompson <will.thompson@collabora.co.uk>
Simon McVittie [Thu, 14 Jul 2011 17:48:03 +0000 (18:48 +0100)]
message-factory: move static functions only called by #if 0 code into another #if 0
The comment "This is really expensive and doesn't add too much coverage"
is copied from the commented-out use of this code.
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=39231
Reviewed-by: Will Thompson <will.thompson@collabora.co.uk>
Simon McVittie [Thu, 14 Jul 2011 17:47:18 +0000 (18:47 +0100)]
dbus-memory: move variable only used on Windows into the conditional
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=39231
Reviewed-by: Will Thompson <will.thompson@collabora.co.uk>
Simon McVittie [Thu, 14 Jul 2011 17:46:57 +0000 (18:46 +0100)]
dbus-marshal-recursive-util: remove unused function call
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=39231
Reviewed-by: Will Thompson <will.thompson@collabora.co.uk>
Simon McVittie [Thu, 14 Jul 2011 17:46:38 +0000 (18:46 +0100)]
config-parser: don't try to get CommonProgramFiles from the environment on Unix
It's unused on Unix, and gcc warns.
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=39231
Reviewed-by: Will Thompson <will.thompson@collabora.co.uk>
Simon McVittie [Thu, 14 Jul 2011 17:46:06 +0000 (18:46 +0100)]
activation-helper: remove unused variable from the real build
It's only used in the test build.
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=39231
Reviewed-by: Will Thompson <will.thompson@collabora.co.uk>
Simon McVittie [Thu, 14 Jul 2011 17:45:45 +0000 (18:45 +0100)]
activation-helper: check_permissions: in the test build, don't have unused variables
In the real build they're used to check permissions.
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=39231
Reviewed-by: Will Thompson <will.thompson@collabora.co.uk>
Simon McVittie [Thu, 14 Jul 2011 17:45:13 +0000 (18:45 +0100)]
_dbus_get_current_time: move struct timeval into the non-monotonic branch
It's only used there.
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=39231
Reviewed-by: Will Thompson <will.thompson@collabora.co.uk>
Simon McVittie [Thu, 14 Jul 2011 17:44:46 +0000 (18:44 +0100)]
Remove various never-used variables
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=39231
Reviewed-by: Will Thompson <will.thompson@collabora.co.uk>
Simon McVittie [Thu, 14 Jul 2011 17:43:04 +0000 (18:43 +0100)]
test-service, test-shell-service: check what RequestName returned
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=39231
Reviewed-by: Will Thompson <will.thompson@collabora.co.uk>
Simon McVittie [Thu, 14 Jul 2011 17:42:42 +0000 (18:42 +0100)]
dbus-monitor: remove unused SIGINT handler
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=39231
Reviewed-by: Will Thompson <will.thompson@collabora.co.uk>
Steve Grubb [Mon, 1 Aug 2011 18:28:27 +0000 (14:28 -0400)]
capng: Drop supplemental groups when switching to daemon user
This patch ensures the daemon process doesn't inherit any supplemental
groups for the root user from an administrator login via an init
script.
This is only an issue for pre-systemd systems.
https://bugzilla.redhat.com/show_bug.cgi?id=726953
Colin Walters [Mon, 1 Aug 2011 16:59:32 +0000 (12:59 -0400)]
autogen.sh: Don't run configure if NOCONFIGURE is set
See http://people.gnome.org/~walters/docs/build-api.txt
Simon McVittie [Fri, 29 Jul 2011 16:20:55 +0000 (17:20 +0100)]
micro version
Simon McVittie [Fri, 29 Jul 2011 15:46:38 +0000 (16:46 +0100)]
spec v0.18 and D-Bus 1.5.6
Simon McVittie [Thu, 21 Jul 2011 16:02:47 +0000 (17:02 +0100)]
spec: mention that buses may have a security policy, but leave it implementation-specific
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=39450
Reviewed-by: Thiago Macieira <thiago@kde.org>
Simon McVittie [Thu, 21 Jul 2011 15:24:01 +0000 (16:24 +0100)]
Move the explanation of message routing to the Message Routing section, leaving behind a summary
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=39450
Reviewed-by: Thiago Macieira <thiago@kde.org>
Simon McVittie [Fri, 29 Jul 2011 15:31:48 +0000 (16:31 +0100)]
Define eavesdropping, and document the eavesdrop match
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=39450
Reviewed-by: Thiago Macieira <thiago@kde.org>
Simon McVittie [Thu, 21 Jul 2011 15:17:35 +0000 (16:17 +0100)]
spec: define unicast messages and broadcast signals, and explicitly allow unicast signals
I believe that the wording of the spec has always allowed unicast signals,
but most bindings assume that signals are broadcasts, so it seems worth
saying specifically that this feature exists and can be useful.
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=39450
Reviewed-by: Thiago Macieira <thiago@kde.org>
Simon McVittie [Thu, 21 Jul 2011 15:14:02 +0000 (16:14 +0100)]
spec: make the Match Rules section true again
The spec previously claimed that only messages matching the client's
match rules would be received. This is not actually true: messages
listing a client as their DESTINATION are always delivered (security
policy permitting).
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=39450
Reviewed-by: Thiago Macieira <thiago@kde.org>
Simon McVittie [Fri, 29 Jul 2011 15:28:32 +0000 (16:28 +0100)]
Merge branch 'dbus-1.4'
Conflicts:
NEWS
configure.ac
Simon McVittie [Fri, 29 Jul 2011 15:17:32 +0000 (16:17 +0100)]
micro version
Simon McVittie [Fri, 29 Jul 2011 13:28:54 +0000 (14:28 +0100)]
1.4.14
Simon McVittie [Fri, 29 Jul 2011 12:43:09 +0000 (13:43 +0100)]
NEWS for 1.5
Simon McVittie [Fri, 29 Jul 2011 11:52:32 +0000 (12:52 +0100)]
Merge branch 'dbus-1.4'
Conflicts:
NEWS
Simon McVittie [Fri, 29 Jul 2011 11:51:42 +0000 (12:51 +0100)]
NEWS
Simon McVittie [Thu, 28 Jul 2011 15:58:05 +0000 (16:58 +0100)]
Implement _dbus_atomic_get directly, rather than via inc + dec
The Windows implementation is untested, but does at least (cross-)compile,
and matches what GLib does.
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=38005
Reviewed-by: Lennart Poettering <lennart@poettering.net>
Simon McVittie [Wed, 15 Jun 2011 13:37:47 +0000 (14:37 +0100)]
Remove remnants of dbus-viewer
Parts of it were already missing, it wasn't compiled, and it depends on
dbus-glib and Gtk.
Simon McVittie [Wed, 15 Jun 2011 13:30:13 +0000 (14:30 +0100)]
tools: use AM_CPPFLAGS, and wrap its lines
Simon McVittie [Wed, 15 Jun 2011 13:27:34 +0000 (14:27 +0100)]
test, test/name-test: remove SOURCES where they can be inferred from the name of the executable
Simon McVittie [Wed, 15 Jun 2011 13:25:54 +0000 (14:25 +0100)]
test/name-test: remove useless CFLAGS
Simon McVittie [Wed, 15 Jun 2011 13:25:43 +0000 (14:25 +0100)]
test/name-test: use AM_CPPFLAGS instead of INCLUDES
Simon McVittie [Wed, 15 Jun 2011 13:23:58 +0000 (14:23 +0100)]
dbus: use AM_CPPFLAGS
Simon McVittie [Wed, 15 Jun 2011 13:22:42 +0000 (14:22 +0100)]
test: promote GLIB_CFLAGS, DBUS_GLIB_CFLAGS to top level and use AM_CPPFLAGS
INCLUDES is a deprecated way to get the same effect as AM_CPPFLAGS.
It's harmless to add extra -I directories to all the tests, even those
that use neither GLib nor dbus-glib, so we can simplify by setting these
AM_CPPFLAGS for the whole directory.
Simon McVittie [Wed, 15 Jun 2011 13:21:14 +0000 (14:21 +0100)]
test: reverse assumption regarding static/dynamic linking
If we change the default assumption to be that new tests will be
dynamically-linked to libdbus, those tests can be useful for
installcheck or even for installation. Accordingly, explicitly use
new variable $(static_cppflags) for all tests that need static linking.
Simon McVittie [Wed, 15 Jun 2011 13:14:34 +0000 (14:14 +0100)]
bus: use AM_CPPFLAGS to simplify preprocessor invocation
Everything in this directory is statically linked to libdbus-internal,
so we can make -DDBUS_STATIC_BUILD global. Also, merge INCLUDES into
AM_CPPFLAGS (it's an older name for the same functionality).
Simon McVittie [Wed, 15 Jun 2011 10:50:33 +0000 (11:50 +0100)]
Simplify linking for tests that use libdbus-testutils
This is the library used by tests that link libdbus-internal and DBusLoop.
By linking libdbus-internal into it, we can avoid having to repeat that
dependency all over the place - libtool and cmake both know how to follow
recursive dependencies.
In cmake, also use libdbus-testutils for more tests, in preference to
repeating its source files.
Simon McVittie [Wed, 15 Jun 2011 10:41:27 +0000 (11:41 +0100)]
Don't explicitly link thread and network libs into various tests
These tests get everything they need from the public or internal API of
libdbus-internal.la, and libtool knows how to pull in libraries'
dependencies, so we don't need explicit linking.
spawn-test and break-loader don't actually need test-utils.[ch]
either.
Simon McVittie [Fri, 29 Jul 2011 10:09:02 +0000 (11:09 +0100)]
Rename DBUS_CLIENT_LIBS to LIBDBUS_LIBS to more accurately reflect its meaning
This does still need to be in configure.ac, because it's common to
dbus/Makefile.am (linking the static/shared library) and dbus-1*.pc.in
(telling static library users which additional libraries they must link).
Simon McVittie [Wed, 15 Jun 2011 09:34:49 +0000 (10:34 +0100)]
tools: simplify linking
* dbus-send, dbus-uuidgen only need to link libdbus; libtool knows what
extra libraries libdbus depends on
* dbus-monitor uses a Winsock header (on Windows) so it needs
NETWORK_libs,but still doesn't need threads
* dbus-launch needs X (on Unix) but doesn't directly need threads or
networking
Simon McVittie [Wed, 15 Jun 2011 09:30:13 +0000 (10:30 +0100)]
tools: put bin_PROGRAMS one per line, eliminate unnecessary variable
Simon McVittie [Tue, 14 Jun 2011 16:44:35 +0000 (17:44 +0100)]
Move knowledge of what the bus daemon/launcher need to link into bus/
Simon McVittie [Tue, 14 Jun 2011 16:40:07 +0000 (17:40 +0100)]
Substitute cflags/libs per-library
This means each module can link whatever it needs to, localizing the
knowledge of which module needs which libraries into its own
Makefile.am.
Simon McVittie [Tue, 14 Jun 2011 16:20:50 +0000 (17:20 +0100)]
Stop defining DBUS_API_SUBJECT_TO_CHANGE
Seriously.
Simon McVittie [Tue, 14 Jun 2011 15:56:13 +0000 (16:56 +0100)]
dbus: remove export_symbols_internal, it will always be empty
The whole point of libdbus-internal.la is that it exports all its symbols.
Simon McVittie [Wed, 15 Jun 2011 10:32:29 +0000 (11:32 +0100)]
Simplify linking by moving -export-dynamic to AM_LDFLAGS
Simon McVittie [Tue, 14 Jun 2011 15:37:24 +0000 (16:37 +0100)]
Remove DBUS_GLIB_CFLAGS from name-test subdirectory (not used/needed)
This appears to be left over from when dbus-glib was part of dbus.
Simon McVittie [Tue, 14 Jun 2011 15:36:14 +0000 (16:36 +0100)]
Remove DBUS_CLIENT_CFLAGS, which is always empty
Simon McVittie [Tue, 14 Jun 2011 15:35:08 +0000 (16:35 +0100)]
Remove DBUS_TEST_CFLAGS, which is empty
Simon McVittie [Wed, 15 Jun 2011 10:24:29 +0000 (11:24 +0100)]
Remove leftover reference to dbus-md5.[ch]
Simon McVittie [Wed, 15 Jun 2011 10:24:16 +0000 (11:24 +0100)]
Remove declarations for tests that no longer exist
Simon McVittie [Fri, 29 Jul 2011 09:45:43 +0000 (10:45 +0100)]
Fix compilation for removal of link cache
Lennart Poettering [Thu, 28 Jul 2011 19:45:35 +0000 (21:45 +0200)]
Merge branch 'dbus-1.4'
Lennart Poettering [Thu, 28 Jul 2011 03:36:33 +0000 (05:36 +0200)]
sysdeps-unix: remove some redundant assignments
Lennart Poettering [Thu, 28 Jul 2011 19:26:21 +0000 (21:26 +0200)]
Merge branch 'dbus-1.4'
Lennart Poettering [Fri, 11 Mar 2011 02:03:03 +0000 (03:03 +0100)]
activation: add /lib/dbus-1/system-services to the search path for services
In order to allow D-Bus usage during early boot (where /usr is not
accessible) also search for bus activation files in
/lib/dbus-1/system-services/. This is only a first step in the right
direction, before we really can boot without /usr we'd need to move all
current activation files (or possibly replace
/usr/dbus-1/system-services to a symlink to
/lib/dbus-1/system-services).
Simon McVittie [Fri, 29 Apr 2011 15:01:59 +0000 (16:01 +0100)]
_dbus_message_remove_counter: remove ability to return the link
Reviewed-by: Colin Walters <walters@verbum.org>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=34393
Simon McVittie [Fri, 11 Mar 2011 13:24:08 +0000 (13:24 +0000)]
Add a regression test that can reproduce fd.o #34393
The number of messages is arbitrary; the more messages, the more likely
the crash is. 2000 messages seem to cause it reliably on this laptop,
but I've set it to 10000 to be safe.
Reviewed-by: Colin Walters <walters@verbum.org>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=34393
Simon McVittie [Fri, 29 Apr 2011 14:08:53 +0000 (15:08 +0100)]
Remove the per-connection link cache
With fd.o#34393 fixed, retaking the lock to cache unused links
significantly adds to locking overhead (-18% throughput in a synthetic
benchmark on an ARM device). The cache is also unlimited in size, and
probably contributes to memory growth and fragmentation by not being
under the system malloc's control.
Fixing fd.o #34393, but also dropping this cache entirely, turns out to
lead to a 5% increase in throughput on the same synthetic benchmark.
Reviewed-by: Colin Walters <walters@verbum.org>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=34393