Simon McVittie [Thu, 28 Jul 2011 15:24:11 +0000 (16:24 +0100)]
Merge branch 'dbus-1.4'
Simon McVittie [Thu, 28 Jul 2011 15:23:19 +0000 (16:23 +0100)]
Add a note about bootstrapping new platforms to README
Simon McVittie [Tue, 26 Apr 2011 16:37:21 +0000 (17:37 +0100)]
Upgrade the type system into its own top-level section of the spec
The type system can be used independently, for instance in GVariant
(although GVariant's binary encoding is in fact not the same).
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=38252
Reviewed-by: Lennart Poettering <lennart@poettering.net>
Simon McVittie [Thu, 28 Jul 2011 10:32:46 +0000 (11:32 +0100)]
Merge branch 'dbus-1.4'
Simon McVittie [Tue, 14 Jun 2011 09:10:24 +0000 (10:10 +0100)]
man pages: replace all unescaped hyphen/minus characters with \-
In a man page, "-" officially means a typographical (Unicode) hyphen,
which frequently breaks the ability to copy and paste code examples from
a man page. "\-" means the ASCII hyphen/minus character. See
<http://lintian.debian.org/tags/hyphen-used-as-minus-sign.html> for
more details.
Rather than trying to distinguish between hyphens, em-dashes and
hyphen/minus, I just replaced all ambiguous hyphens with \- by applying
this vim command repeatedly until it didn't find anything:
%s/\(^\|[^\\]\)-/\1\\-/g
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=38284
Reviewed-by: Lennart Poettering <lennart@poettering.net>
Simon McVittie [Fri, 10 Jun 2011 15:45:34 +0000 (16:45 +0100)]
Fix spelling/grammatical mistakes detected by Debian's lintian(1)
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=38284
Reviewed-by: Lennart Poettering <lennart@poettering.net>
Simon McVittie [Wed, 27 Jul 2011 17:54:35 +0000 (18:54 +0100)]
Merge branch 'dbus-1.4'
Sascha Silbe [Sun, 17 Oct 2010 12:38:20 +0000 (14:38 +0200)]
list_allows_user: fix debug output for groups
Signed-off-by: Sascha Silbe <sascha-pgp@silbe.org>
Reviewed-by: Simon McVittie <simon.mcvittie@collabora.co.uk>
Simon McVittie [Tue, 26 Jul 2011 17:00:57 +0000 (18:00 +0100)]
Merge branch 'dbus-1.4'
Conflicts:
dbus/dbus-message.c
Simon McVittie [Tue, 26 Jul 2011 16:59:23 +0000 (17:59 +0100)]
NEWS
Simon McVittie [Thu, 14 Jul 2011 16:20:21 +0000 (17:20 +0100)]
DBusMessage: always access refcount atomically, even for assertions/initial ref
Reviewed-by: Cosimo Alfarano <cosimo.alfarano@collabora.co.uk>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=38005
Simon McVittie [Thu, 14 Jul 2011 16:26:52 +0000 (17:26 +0100)]
dbus-memory: use atomic accesses to block count, even for assertions
Reviewed-by: Cosimo Alfarano <cosimo.alfarano@collabora.co.uk>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=38005
Simon McVittie [Tue, 26 Jul 2011 10:51:13 +0000 (11:51 +0100)]
Merge branch 'dbus-1.4'
Simon McVittie [Tue, 26 Jul 2011 10:51:04 +0000 (11:51 +0100)]
NEWS
Simon McVittie [Thu, 14 Jul 2011 16:23:21 +0000 (17:23 +0100)]
DBusConnection: use atomic accesses to refcount in assertions/initial ref
Reviewed-by: Cosimo Alfarano <cosimo.alfarano@collabora.co.uk>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=38005
Simon McVittie [Thu, 14 Jul 2011 16:14:29 +0000 (17:14 +0100)]
dbus_message_ref: avoid unused variable if not asserting
Reviewed-by: Cosimo Alfarano <cosimo.alfarano@collabora.co.uk>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=38005
Simon McVittie [Thu, 14 Jul 2011 16:14:06 +0000 (17:14 +0100)]
_dbus_connection_close_if_only_one_ref: use _dbus_atomic_get
Reviewed-by: Cosimo Alfarano <cosimo.alfarano@collabora.co.uk>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=38005
Simon McVittie [Thu, 14 Jul 2011 16:07:08 +0000 (17:07 +0100)]
Add _dbus_atomic_get implemented in terms of inc, dec
Reviewed-by: Cosimo Alfarano <cosimo.alfarano@collabora.co.uk>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=38005
Simon McVittie [Thu, 23 Jun 2011 12:39:00 +0000 (13:39 +0100)]
DBusObjectTree: always access refcount atomically
Reviewed-by: Cosimo Alfarano <cosimo.alfarano@collabora.co.uk>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=38005
Simon McVittie [Thu, 23 Jun 2011 12:38:31 +0000 (13:38 +0100)]
dbus_message_unref: make an assertion more strict
We've just decremented the refcount, so it should have been at least 1
before we did that.
Reviewed-by: Cosimo Alfarano <cosimo.alfarano@collabora.co.uk>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=38005
Simon McVittie [Thu, 23 Jun 2011 12:35:55 +0000 (13:35 +0100)]
DBusMessageFilter: exclusively use atomic accesses to refcount
Reviewed-by: Cosimo Alfarano <cosimo.alfarano@collabora.co.uk>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=38005
Simon McVittie [Mon, 18 Jul 2011 18:42:35 +0000 (19:42 +0100)]
Merge branch 'dbus-1.4' and update NEWS for master
Conflicts:
NEWS
Simon McVittie [Mon, 18 Jul 2011 18:33:00 +0000 (19:33 +0100)]
NEWS
Simon McVittie [Tue, 12 Apr 2011 12:17:30 +0000 (13:17 +0100)]
Uninstall the installed docs
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=36156
Reviewed-by: Will Thompson <will.thompson@collabora.co.uk>
Simon McVittie [Tue, 12 Apr 2011 12:08:22 +0000 (13:08 +0100)]
Install man2html output to $(htmldir)
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=36156
Reviewed-by: Will Thompson <will.thompson@collabora.co.uk>
Simon McVittie [Tue, 12 Apr 2011 12:06:25 +0000 (13:06 +0100)]
Install diagram.*, system-activation.txt to the $(docdir)
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=36156
Reviewed-by: Will Thompson <will.thompson@collabora.co.uk>
Simon McVittie [Tue, 12 Apr 2011 12:01:44 +0000 (13:01 +0100)]
Install xmlto output to $(htmldir) automatically
Also rename HTML_FILES to XMLTO_OUTPUT, as a more self-describing name:
we have other HTML files which are not the output from xmlto.
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=36156
Reviewed-by: Will Thompson <will.thompson@collabora.co.uk>
Simon McVittie [Tue, 12 Apr 2011 11:54:22 +0000 (12:54 +0100)]
Install dbus.devhelp to $(htmldir), not $(apidir), and remove it in clean
It needs to be in $(htmldir) so that its relative paths will work.
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=36156
Reviewed-by: Will Thompson <will.thompson@collabora.co.uk>
Simon McVittie [Tue, 12 Apr 2011 11:52:30 +0000 (12:52 +0100)]
configure.ac: fix check for xsltproc
Pre-setting XSLTPROC interferes with AC_CHECK_PROGS letting the user
override choice of program via the environment.
Reviewed-by: Will Thompson <will.thompson@collabora.co.uk>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=36156
Simon McVittie [Wed, 15 Jun 2011 11:43:15 +0000 (12:43 +0100)]
Use EXEEXT when running tests from another directory, and skip bus-test-launch-helper on non-Unix
This is necessary when cross-compiling from Linux to mingw32 and running
the resulting tests under Wine. (This partially works! Some tests fail,
though.)
Reviewed-by: Will Thompson <will.thompson@collabora.co.uk>
Simon McVittie [Mon, 6 Jun 2011 15:12:05 +0000 (16:12 +0100)]
Build docs after running tests, and remove redundant DIST_SUBDIRS
If DIST_SUBDIRS isn't set, it defaults to SUBDIRS, so it's just noise.
Running tests before building documentation is an easy way to speed up the
hack/make check/fix cycle, by not wasting time rebuilding the
documentation (which is often slow) until all the tests compile and pass.
https://bugs.freedesktop.org/show_bug.cgi?id=34405
Reviewed-by: Will Thompson <will.thompson@collabora.co.uk>
Simon McVittie [Mon, 13 Jun 2011 10:11:31 +0000 (11:11 +0100)]
Run bus tests from test/
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=34405
Reviewed-by: Will Thompson <will.thompson@collabora.co.uk>
Simon McVittie [Mon, 13 Jun 2011 10:08:28 +0000 (11:08 +0100)]
Run dbus-test from test/, not dbus/
This avoids failures caused by test/data not having been set up yet.
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=34405
Reviewed-by: Will Thompson <will.thompson@collabora.co.uk>
Simon McVittie [Mon, 6 Jun 2011 14:57:45 +0000 (15:57 +0100)]
dbus tests: don't warn on skipped tests
We should treat _dbus_warn as fatal when we run tests, like bus/ does.
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=34405
Reviewed-by: Will Thompson <will.thompson@collabora.co.uk>
Simon McVittie [Mon, 6 Jun 2011 14:01:52 +0000 (15:01 +0100)]
test/Makefile.am: list tests one per line
This improves clarity, and makes conflicts less likely and merges more
obviously correct.
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=34405
Reviewed-by: Will Thompson <will.thompson@collabora.co.uk>
Simon McVittie [Wed, 26 Jan 2011 18:24:03 +0000 (18:24 +0000)]
Opt-in to modern OS features on Solaris, to get file descriptor passing
Surely it's broken to define SCM_RIGHTS but not the necessary structs to
do something useful with it, but whatever.
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=33465
Reviewed-by: Will Thompson <will.thompson@collabora.co.uk>
Simon McVittie [Wed, 22 Jun 2011 11:31:24 +0000 (12:31 +0100)]
_dbus_open_unix_socket: make static
It isn't called from outside its translation unit.
Simon McVittie [Wed, 22 Jun 2011 11:31:08 +0000 (12:31 +0100)]
_dbus_open_tcp_socket: remove
It's only implemented on Unix, internal to dbus, and never called.
Simon McVittie [Mon, 18 Jul 2011 14:15:45 +0000 (15:15 +0100)]
Merge branch 'dbus-1.4'
Simon McVittie [Wed, 22 Jun 2011 10:59:32 +0000 (11:59 +0100)]
_dbus_connect_tcp_socket_with_nonce: don't create an extra fd (which is then leaked)
This block should have been deleted in 2007, when IPv6 support was added:
previously, the fd allocated at the beginning of the function was used
for connect(), but for IPv6 support, the socket() call has to be inside
the loop over getaddrinfo() results so its address family can be changed.
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=37258
Reviewed-by: Cosimo Alfarano <cosimo.alfarano@collabora.co.uk>
Simon McVittie [Thu, 14 Jul 2011 15:49:02 +0000 (16:49 +0100)]
Merge branch 'dbus-1.4'
Simon McVittie [Thu, 23 Jun 2011 12:22:44 +0000 (13:22 +0100)]
DBusServer: exclusively use atomic operations on the refcount
Same reasoning as for fd.o #38005, commit
50732523a7.
Reviewed-by: Will Thompson <will.thompson@collabora.co.uk>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=38005
Bug-NB: NB#263436
Simon McVittie [Thu, 23 Jun 2011 12:08:14 +0000 (13:08 +0100)]
DBusPendingCall: exclusively use atomic operations on the refcount
Same reasoning as for fd.o #38005, commit
50732523a7.
Reviewed-by: Will Thompson <will.thompson@collabora.co.uk>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=38005
Bug-NB: NB#263436
Cosimo Alfarano [Wed, 13 Jul 2011 17:05:32 +0000 (18:05 +0100)]
Add testsuite for eavesdrop=true DBusMatchRule
[with some typos fixed during commit -smcv]
Reviewed-by: Simon McVittie <simon.mcvittie@collabora.co.uk>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=37890
Bug-NB: NB#269748
Cosimo Alfarano [Tue, 12 Jul 2011 14:56:43 +0000 (15:56 +0100)]
Fix dbus-monitor using eavesdrop=true when no filters are passed
Reviewed-by: Simon McVittie <simon.mcvittie@collabora.co.uk>
Simon McVittie [Fri, 10 Jun 2011 09:12:07 +0000 (10:12 +0100)]
DBusHeader: only store byte-order in the fixed-length header
Reviewed-by: Thiago Macieira <thiago@kde.org>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=38287
Simon McVittie [Fri, 10 Jun 2011 09:10:54 +0000 (10:10 +0100)]
DBusMessage: don't redundantly store byte order, ask the DBusHeader
Reviewed-by: Thiago Macieira <thiago@kde.org>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=38287
Cosimo Alfarano [Thu, 7 Jul 2011 15:10:54 +0000 (16:10 +0100)]
Add "eavesdrop=true" as constant match rule for dbus-monitor
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=37890
Bug-NB: NB#269748
Reviewed-by: Simon McVittie <simon.mcvittie@collabora.co.uk>
Cosimo Alfarano [Mon, 11 Jul 2011 16:03:00 +0000 (17:03 +0100)]
Do not allow eavedropping unless rule owner explicitely declare it
Adds "eavesdrop=true" as a match rule, meaning that the owner
intend to eavedrop.
Otherwise the owner will receive only broadcasted messages and the ones
meant to be delivered to it.
[plus a typo fix in an error message -smcv]
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=37890
Bug-NB: NB#269748
Reviewed-by: Simon McVittie <simon.mcvittie@collabora.co.uk>
Simon McVittie [Mon, 11 Jul 2011 12:23:38 +0000 (13:23 +0100)]
Merge remote-tracking branch 'origin/dbus-1.4'
Jiří Klimeš [Fri, 1 Jul 2011 11:59:48 +0000 (13:59 +0200)]
DBusConnection: simplify registering object paths code
Introduce static generic _dbus_connection_register_object_path()
function to remove duplicate code for registration object paths.
Having *four* almost the same functions is error-prone and hard
to follow as well.
Signed-off-by: Jiří Klimeš <jklimes@redhat.com>
Reviewed-by: Simon McVittie <simon.mcvittie@collabora.co.uk>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=38874
Simon McVittie [Sun, 3 Jul 2011 04:47:39 +0000 (06:47 +0200)]
Cope with platforms whose vsnprintf violates both POSIX and C99 - part 1
Simon McVittie [Tue, 21 Jun 2011 13:09:52 +0000 (14:09 +0100)]
In Windows _dbus_printf_string_upper_bound, don't crash on malloc failure
Ralf Habacker [Fri, 1 Jul 2011 22:50:39 +0000 (00:50 +0200)]
Added DBUS_ENABLE_STATS option to cmake buildsystem to keep in sync with autotools.
Ralf Habacker [Fri, 8 Jul 2011 08:53:20 +0000 (10:53 +0200)]
Visual Studio Professional uses a different ide starter - pointed out by Romain Pokrzywka
Ralf Habacker [Sat, 2 Jul 2011 19:02:46 +0000 (21:02 +0200)]
Collected cmake doc into README.cmake.
Ralf Habacker [Fri, 8 Jul 2011 08:53:20 +0000 (10:53 +0200)]
Visual Studio Professional uses a different ide starter - pointed out by Romain Pokrzywka
Ralf Habacker [Fri, 1 Jul 2011 14:28:19 +0000 (16:28 +0200)]
win32 fix: Keep server autolaunch mutex settings according to client detection algorithmus.
Ralf Habacker [Mon, 4 Jul 2011 21:47:45 +0000 (23:47 +0200)]
Fixed cmake buildsystem - option command only supports boolean values
Simon McVittie [Sun, 3 Jul 2011 04:47:39 +0000 (06:47 +0200)]
Cope with platforms whose vsnprintf violates both POSIX and C99 - part 1
Simon McVittie [Tue, 21 Jun 2011 13:09:52 +0000 (14:09 +0100)]
In Windows _dbus_printf_string_upper_bound, don't crash on malloc failure
Ralf Habacker [Sat, 2 Jul 2011 22:54:20 +0000 (00:54 +0200)]
Limit batch file creating to windows
Ralf Habacker [Sun, 3 Jul 2011 00:01:25 +0000 (02:01 +0200)]
Fixed cmake buildsystem on unix - option command only supports boolean values
Ralf Habacker [Sat, 2 Jul 2011 19:02:46 +0000 (21:02 +0200)]
Collected cmake doc into README.cmake.
Ralf Habacker [Fri, 1 Jul 2011 22:50:39 +0000 (00:50 +0200)]
Added DBUS_ENABLE_STATS option to cmake buildsystem to keep in sync with autotools.
Ralf Habacker [Fri, 1 Jul 2011 14:28:19 +0000 (16:28 +0200)]
win32 fix: Keep server autolaunch mutex settings according to client detection algorithmus.
Simon McVittie [Fri, 1 Jul 2011 13:54:42 +0000 (14:54 +0100)]
Merge branch 'dbus-1.4'
Jiří Klimeš [Fri, 1 Jul 2011 10:58:11 +0000 (12:58 +0200)]
DBusConnection: use DBUS_ERROR_OBJECT_PATH_IN_USE instead of DBUS_ERROR_ADDRESS_IN_USE
While registering an object path the possible error is
DBUS_ERROR_OBJECT_PATH_IN_USE, not DBUS_ERROR_ADDRESS_IN_USE (the error
is set by _dbus_object_tree_register()).
Signed-off-by: Jiří Klimeš <jklimes@redhat.com>
Reviewed-by: Simon McVittie <simon.mcvittie@collabora.co.uk>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=38874
Simon McVittie [Tue, 8 Feb 2011 15:48:12 +0000 (15:48 +0000)]
Include size of link cache in per-connection statistics
Reviewed-by: Cosimo Alfarano <cosimo.alfarano@collabora.co.uk>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=34040
Simon McVittie [Fri, 11 Feb 2011 17:16:09 +0000 (17:16 +0000)]
Also record peak values for queued bytes/fds in connection stats
Reviewed-by: Cosimo Alfarano <cosimo.alfarano@collabora.co.uk>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=34040
Simon McVittie [Tue, 8 Feb 2011 15:41:54 +0000 (15:41 +0000)]
Add an initial round of stats to the Stats interface
Reviewed-by: Cosimo Alfarano <cosimo.alfarano@collabora.co.uk>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=34040
Simon McVittie [Tue, 8 Feb 2011 12:51:33 +0000 (12:51 +0000)]
DBusConnection, DBusTransport: add queue statistics
Reviewed-by: Cosimo Alfarano <cosimo.alfarano@collabora.co.uk>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=34040
Simon McVittie [Tue, 8 Feb 2011 12:31:08 +0000 (12:31 +0000)]
BusConnections: add usage stats for well-known names, match rules
Reviewed-by: Cosimo Alfarano <cosimo.alfarano@collabora.co.uk>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=34040
Simon McVittie [Tue, 8 Feb 2011 12:30:06 +0000 (12:30 +0000)]
DBusList: add usage stats
Reviewed-by: Cosimo Alfarano <cosimo.alfarano@collabora.co.uk>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=34040
Simon McVittie [Tue, 8 Feb 2011 12:29:52 +0000 (12:29 +0000)]
DBusMemPool: add usage stats
Reviewed-by: Cosimo Alfarano <cosimo.alfarano@collabora.co.uk>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=34040
Simon McVittie [Tue, 8 Feb 2011 15:33:35 +0000 (15:33 +0000)]
Add a stub .Debug.Stats interface if --enable-stats
There are no actual statistics yet, just a count of how many times the
method has been called, and (for the per-connection stats) the unique name.
Reviewed-by: Cosimo Alfarano <cosimo.alfarano@collabora.co.uk>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=34040
Simon McVittie [Tue, 1 Mar 2011 15:14:02 +0000 (15:14 +0000)]
configure.ac: add --enable-stats
Reviewed-by: Cosimo Alfarano <cosimo.alfarano@collabora.co.uk>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=34040
Simon McVittie [Mon, 31 Jan 2011 15:55:53 +0000 (15:55 +0000)]
match_rule_to_string: don't reinvent dbus_message_type_to_string
We don't allow match rules with unknown message types, so losing the
"type='%d'" case isn't a great loss.
Reviewed-by: Cosimo Alfarano <cosimo.alfarano@collabora.co.uk>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=34040
Simon McVittie [Wed, 22 Jun 2011 13:57:56 +0000 (14:57 +0100)]
Remove maximum length field from DBusString
The source code says it's "a historical artifact from a feature that
turned out to be dumb". Respond accordingly!
This reduces sizeof (DBusString) by 20% on ILP32 architectures, which
can't hurt. (No reduction on LP64 architectures that align pointers
naturally, unfortunately.)
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=38570
Reviewed-by: Will Thompson <will.thompson@collabora.co.uk>
Simon McVittie [Wed, 22 Jun 2011 13:56:27 +0000 (14:56 +0100)]
Remove unused _dbus_string_copy_data_len, _dbus_string_steal_data_len
They're only compiled when building tests.
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=38570
Reviewed-by: Will Thompson <will.thompson@collabora.co.uk>
Simon McVittie [Wed, 22 Jun 2011 13:44:28 +0000 (14:44 +0100)]
Split DBUS_GENERIC_STRING_PREAMBLE into sensible lines
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=38570
Reviewed-by: Will Thompson <will.thompson@collabora.co.uk>
Ralf Habacker [Mon, 20 Jun 2011 07:29:12 +0000 (09:29 +0200)]
CMake fix: Do not use FORCE in variables which should be changeable by the user.
Ralf Habacker [Tue, 14 Jun 2011 16:45:46 +0000 (18:45 +0200)]
fixed cmake error: doc says CACHE needs type *and* doc string
Simon McVittie [Mon, 13 Jun 2011 15:54:00 +0000 (16:54 +0100)]
NEWS
Simon McVittie [Thu, 7 Apr 2011 16:36:54 +0000 (17:36 +0100)]
DBusLoop: store watches in a hash table keyed by fd
This means we don't need to build up the watches_for_fds array, because
it's quick to find the watches based on the fd.
This also fixes fd.o #23194, because we only put each fd in the
array of DBusPollFD once, with the union of all its watches' flags; we
can no longer get more entries in the array than our number of file
descriptors, which is capped at our RLIMIT_NOFILE, so we can't get
EINVAL from poll(2).
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=23194
Reviewed-by: Thiago Macieira <thiago@kde.org>
Simon McVittie [Mon, 24 Jan 2011 15:30:02 +0000 (15:30 +0000)]
DBusLoop: fold Callback into TimeoutCallback
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=33342
Reviewed-by: Thiago Macieira <thiago@kde.org>
Simon McVittie [Tue, 25 Jan 2011 12:55:27 +0000 (12:55 +0000)]
Drop WatchCallback entirely, just use a list of DBusWatch in the main loop
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=33342
Reviewed-by: Thiago Macieira <thiago@kde.org>
Simon McVittie [Mon, 24 Jan 2011 14:38:13 +0000 (14:38 +0000)]
DBusLoop: inline add_callback, remove_callback into their callers
The watch and timeout code paths will diverge completely when we change
WatchCallback * to just be a DBusWatch *, removing the benefit of having
common code.
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=33342
Reviewed-by: Thiago Macieira <thiago@kde.org>
Simon McVittie [Tue, 25 Jan 2011 12:37:01 +0000 (12:37 +0000)]
DBusLoop: move OOM flag in watches inside the DBusWatch
This will eventually let us maintain a DBusPollFD[] of just the active
watches, between several iterations. The more immediate benefit is that
WatchCallback can go away, because it only contains a refcount, a
now-useless type, and a watch that already has its own refcount.
This doesn't take any more memory for DBusWatch when not using DBusLoop
(e.g. in client/service code or bindings), because we're just using more
bits in an existing bitfield.
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=33342
Reviewed-by: Thiago Macieira <thiago@kde.org>
Simon McVittie [Fri, 21 Jan 2011 18:37:26 +0000 (18:37 +0000)]
DBusLoop: keep separate lists of watches and timeouts
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=33342
Reviewed-by: Thiago Macieira <thiago@kde.org>
Simon McVittie [Fri, 21 Jan 2011 18:29:05 +0000 (18:29 +0000)]
DBusLoop: factor out watch_flags_to_poll_events, watch_flags_from_poll_revents
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=33342
Reviewed-by: Thiago Macieira <thiago@kde.org>
Simon McVittie [Tue, 25 Jan 2011 12:28:32 +0000 (12:28 +0000)]
Remove _dbus_loop_add_watch_full
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=33342
Reviewed-by: Thiago Macieira <thiago@kde.org>
Simon McVittie [Fri, 21 Jan 2011 18:24:19 +0000 (18:24 +0000)]
bus-activation: separate the "finished" callback from the watch callback
This has been marked as broken since 2003...
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=33342
Reviewed-by: Thiago Macieira <thiago@kde.org>
Simon McVittie [Fri, 21 Jan 2011 18:54:33 +0000 (18:54 +0000)]
DBusLoop: remove second layer of watch callbacks where possible
Similar to the previous commit, almost every use of DBusWatch can just
have the main loop call dbus_watch_handle.
The one exception is the bus activation code; it's had a comment
explaining why it's wrong since 2003. We should fix that one day, but for
now, just migrate it to a new _dbus_loop_add_watch_full which preserves
the second-layer callback.
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=33342
Reviewed-by: Thiago Macieira <thiago@kde.org>
Simon McVittie [Fri, 21 Jan 2011 18:54:09 +0000 (18:54 +0000)]
DBusLoop: remove a layer of pointless abstraction around timeouts
Instead of supplying 8 tiny wrapper functions around dbus_timeout_handle,
each with a user_data parameter that's a potentially unsafe borrowed
pointer but isn't actually used, we can call dbus_timeout_handle directly
and save a lot of trouble.
One of the wrappers previously called dbus_timeout_handle repeatedly
if it returned FALSE to indicate OOM, but that timeout's handler never
actually returned FALSE, so there was no practical effect. The rest just
ignore the return, which is documented as OK to do.
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=33342
Reviewed-by: Thiago Macieira <thiago@kde.org>
Simon McVittie [Thu, 20 Jan 2011 15:48:07 +0000 (15:48 +0000)]
DBusConnection: ref the connection in the timeout handler
client_timeout_callback in bus/test.c refs the connection across the
timeout invocation, which looks suspiciously like a workaround. If we
make the timeout handler itself ref the connection, we won't need that,
and can simplify timeout handling drastically.
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=33342
Reviewed-by: Thiago Macieira <thiago@kde.org>
Simon McVittie [Mon, 13 Jun 2011 14:59:14 +0000 (15:59 +0100)]
bus signal_handler: don't use _dbus_warn, and don't pretend to be portable
_dbus_warn isn't async-signal-safe, so that's out. We can use write()
instead; it's POSIX but not ISO C, but then again, so are signals.
Accordingly, 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.
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=33336
Reviewed-by: Thiago Macieira <thiago@kde.org>
Simon McVittie [Mon, 13 Jun 2011 14:58:45 +0000 (15:58 +0100)]
bus: signal_handler: ignore failure to write, and explain why
See the comment in the source code for rationale.
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=33336
Reviewed-by: Thiago Macieira <thiago@kde.org>
Simon McVittie [Fri, 21 Jan 2011 16:49:14 +0000 (16:49 +0000)]
Check that watches are removed, then invalidated, then unreffed
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=33336
Reviewed-by: Will Thompson <will.thompson@collabora.co.uk>
Reviewed-by: Thiago Macieira <thiago@kde.org>
Simon McVittie [Fri, 21 Jan 2011 18:51:27 +0000 (18:51 +0000)]
Always remove, invalidate and free watches before closing watched sockets
This should mean we don't get invalid fds in the main loop.
The BSD (kqueue) and Windows code paths are untested, but follow the same
patterns as the tested Linux/generic Unix versions.
DBusTransportSocket was already OK (it called free_watches() before
_dbus_close_socket, and that did the remove, invalidate, unref dance).
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=33336
Reviewed-by: Will Thompson <will.thompson@collabora.co.uk>
Reviewed-by: Thiago Macieira <thiago@kde.org>