This allows a user to efficiently watch for services with a common
domain prefix.
This is exposed in the API via a wildcard character in the service name.
For example creating a watcher on "org.mpris*" will match
"org.mpris.foo" "org.mpris.bar" and "org.mpris" itself. It will not
match org.mprisasdf.
Internally the argument match rules have been expanded from a single
QStringList to a struct containing args and arg0namespace. This was done
so that we can easily use argpath in match rules in the future.
Change-Id: I55882ab603cc6ba478e8c0ea9a6800f6e483a50c
Reviewed-by: Kai Uwe Broulik <kde@privat.broulik.de>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: David Faure <david.faure@kdab.com>
In c2049f67e4, all DBus I/O was
moved into a service thread. The dispatch lock used to protect
DBus I/O in the threading setup before that commit. It is not
needed anymore.
No discernible difference in QtDBus benchmarks, roughly 500
bytes code size reduction on AMD64. The main point is to reduce
confusion from unnecessary code.
Change-Id: Idcbdd2b7e2b317cf6da0b5bfc5ec70afed1f1b48
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Remaining uses of Q_NULLPTR are in:
src/corelib/global/qcompilerdetection.h
(definition and documentation of Q_NULLPTR)
tests/manual/qcursor/qcursorhighdpi/main.cpp
(a test executable compilable both under Qt4 and Qt5)
Change-Id: If6b074d91486e9b784138f4514f5c6d072acda9a
Reviewed-by: Ville Voutilainen <ville.voutilainen@qt.io>
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
This reverts commit daeb334039.
The commit was causing race conditions, and random failures in CI.
Task-number: QTBUG-60792
Change-Id: I6e49b733965632a1a268f0e88809794098465ec0
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
It's possible that some code executes after QDBusConnectionManager is
destroyed and still tries to access QtDBus. Protect against such
crashes.
Change-Id: I87e17314d8b24ae983b1fffd1453c13fbd3cf48e
Reviewed-by: René J.V. Bertin <rjvbertin@gmail.com>
Reviewed-by: Alex Blasche <alexander.blasche@qt.io>
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@qt.io>
This allows us to get rid of two Q_GLOBAL_STATIC in QtDBus, which means
fewer opportunities for screwing up the order of destruction. And since
QDBusConnectionManager now ensures that the types are initialized, we
don't need to re-initialize them everywhere.
The Q_GLOBAL_STATIC for QDBusConnectionManager ensures the proper
thread-safe locking, so we don't need to lock for every type that we're
trying to register. This should make things faster.
But as a side-effect, trying to register a D-Bus metatype will cause the
QDBusConnectionManager thread to start too.
Change-Id: Ifea6e497f11a461db432ffff1449a4e535234485
Reviewed-by: René J.V. Bertin <rjvbertin@gmail.com>
Reviewed-by: Anton Kudryavtsev <antkudr@mail.ru>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
For the windows file system engine, we add an extra macro to use
library loading if configured to do so, but avoid it on WinRT, as
none of the symbols would be found.
We also QT_REQUIRE_CONFIG(library) in the library headers and
exclude the sources from the build if library loading is disabled.
This, in turn, makes it necessary to clean up some header inclusions.
Change-Id: I2b152cb5b47a2658996b6f4702b038536a5704ec
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@qt.io>
5971b88e is not needed in new configure.
This merge also reverts "fix QMAKE_DEFAULT_*DIRS resolution with
apple SDK", 2c9d15d7, because it breaks iOS build with new
configure system.
Conflicts:
mkspecs/features/default_pre.prf
mkspecs/features/mac/toolchain.prf
mkspecs/features/toolchain.prf
src/dbus/qdbusconnection.cpp
src/plugins/sqldrivers/mysql/qsql_mysql.cpp
src/sql/drivers/mysql/qsql_mysql.cpp
src/widgets/widgets/qmenubar.cpp
src/widgets/widgets/qmenubar_p.h
tools/configure/configureapp.cpp
tools/configure/environment.cpp
tools/configure/environment.h
Change-Id: I995533dd334211ebd25912db05b639d6f908aaec
This function sends D-Bus messages directly (in huntAndEmit), so it
should only be called from the QDBusConnectionManager thread. Somehow, I
missed this in the Qt 5.6 refactoring of QtDBus.
Being called in the wrong thread means that there's a visible behavior
change compared to Qt 5.5: if the user code sent a method call or method
return/error and then emitted a signal, we'd have two threads racing to
send the D-Bus messages. This was observed in Telepathy-Qt code: certain
signals arrived before a method return, even though they were clearly
emitted by a queued QMetaObject::invokeMethod.
In addition to that, we have has an internal problem (though not
observed): the libdbus-1 timer and socket callbacks would be called in
the wrong thread and we no longer have protection against that.
Unit testing not possible since this is a race condition.
Change-Id: I9e96ecd4f6aa4ff0ae08fffd1471d002142613d6
Reviewed-by: Gustavo Pichorim Boiko <gustavo.boiko@canonical.com>
Reviewed-by: Alex Blasche <alexander.blasche@qt.io>
We must ensure that QDBusServer's newConnection() signal has been
processed by the application, before starting processing messages on it.
Task-number: QTBUG-55087
Change-Id: I595329b2f98788dbf9f40558b8c230c0c0817ef8
Reviewed-by: Timo Jyrinki <timo.jyrinki@iki.fi>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
QDBusConnection::connect() behaves like QObject::connect with a
connection type of Qt::UniqueConnection | Qt::QueuedConnection. So
return false if it's already connected.
[ChangeLog][QtDBus][QDBusConnection] Fixed a bug that would cause
QDBusConnection::connect() to return true if a slot was already
connected to the same D-Bus signal. QtDBus does not support multiple
connections.
Change-Id: I87e17314d8b24ae983b1fffd1453aef5a7c9ad0b
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
Because the moment we disconnect from the object's destroyed() signal,
it may get destroyed in another thread. If the same object appears more
than once in the object tree or in the signal hook table, we could be
accessing a dangling pointer.
Task-number: QTBUG-52988
Change-Id: Ifea6e497f11a461db432ffff14496f0f83889104
Reviewed-by: Weng Xuetian <wengxt@gmail.com>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Replace substring functions that return QString with
corresponding functions that return QStringRef where
it's possible.
Create QString from QStringRef only where necessary.
Add overloaded functions with QStringRef arg
in QDBusUtil:
- isValidUniqueConnectionName()
- isValidMemberName()
- isValidPartOfObjectPath()
Change-Id: I4a24a298702728ba7d3a65c39e25c3a9c759e07f
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>
libdbus will send a local signal if connection gets disconnected. When
this happens, end all pending calls with QDBusError::Disconnected.
Task-number: QTBUG-51649
Change-Id: I5c7d2a468bb5da746d0c0e53e458c1e376f186a9
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
If a QObject is added or passed as receiver to QDBusConnection::connect()
and it is managed by Q_GLOBAL_STATIC or similar mechanism, it is
possible that when that its destructor is called after the dbus daemon
thread ends. In that case, QObject::destroyed connected via
Qt::BlockingQueuedConnection to QDBusConnectionPrivate will cause dead
lock since the thread is no longer processing events.
Task-number: QTBUG-51648
Change-Id: I1a1810a6d6d0234af0269d5f3fc1f54101bf1547
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Whenever a message spy was installed, we failed to actually process
looped-back messages by queueing them for processing by the spy. That
had as a consequence that the caller got an error reply. Worse, since
the message had been queued, QtDBus would attempt to deliver it later.
Since that message had isLocal==true, bad things happened inside the
manager thread.
The correct solution is not to queue the message for the filter. If the
message is local, then simply deliver directly, as we're still in the
user's thread. This used to be the behavior in Qt 5.5.
Task-number: QTBUG-51676
Change-Id: I1dc112894cde7121e8ce302ae51b438ade1ff612
Reviewed-by: David Faure <david.faure@kdab.com>
Reviewed-by: Dmitry Shachnev <mitya57@gmail.com>
Reviewed-by: Jan Kundrát <jkt@kde.org>
The printf-style version of QDebug expands to a lot less code than the
std::ostream-style version. Of course, you pay in type safety (but
compilers warn about it these days), you cannot stream complex Qt
types and streaming QStrings is awkward, but in many cases you
actually improve on readability.
But the main reason is that something that's not supposed to be
executed under normal operation has no business bloating executable
code size.
This is not an attempt at converting all qWarnings() to printf-style,
only the low-hanging fruit.
In this first part, replace
qWarning() << ""
with
qWarning("...").
Had to fix broken qImDebug() definition. Instead of defining it as
a nullary macro in the QT_NO_DEBUG case and as a variadic macro in
the other, define it in both cases, as is customary, as a non-function
macro so that overload selection works without requiring variadic
macro support of the compiler.
Saves e.g. ~250b in text size in QtPrintSupport on optimized GCC 5.3
AMD64 builds.
Change-Id: Ie30fe2f7942115d5dbf99fff1750ae0d477c379f
Reviewed-by: Kai Koehne <kai.koehne@theqtcompany.com>
This also reverts commit 018e670a26.
The change was introduced in 5.6. After the refactoring, 14960f52,
in 5.7 branch and a merge, it is not needed any more.
Conflicts:
.qmake.conf
src/corelib/io/qstandardpaths_mac.mm
src/corelib/tools/qsharedpointer_impl.h
tests/auto/widgets/itemviews/qlistview/tst_qlistview.cpp
Change-Id: If4fdff0ebf2b9b5df9f9db93ea0022d5ee3da2a4
interface is a define under wince. This define is included even
with standard header includes already. It needs to be undefined
for using it.
Task-number: QTBUG-50853
Change-Id: Ie44681f03709848e9747a8aec11835c8d62aa409
Reviewed-by: Lars Knoll <lars.knoll@theqtcompany.com>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
qfeatures.txt says that it depends on PROPERTIES, so this condition will
never happen.
Change-Id: I42e7ef1a481840699a8dffff1409e2e63c6857fa
Reviewed-by: Alex Blasche <alexander.blasche@theqtcompany.com>
Whenever there are spies installed, we call out to the main thread to
call to the kded/kiod message spies. This allows the spy code to do just
about anything, where previously it was restricted in what it could do
to avoid deadlocking or triggering assertions if it recursed back into
QDBusConnection code in the manager thread. After the spies are done,
the message is re-inserted into the QDBusConnection processing pipeline.
This commit moves the spy handling to after the check for disabled
dispatching, as doing otherwise would mean the message could get
postponed again for no good reason. It's also possible that the main
thread isn't done installing the hooks, so waiting until the dispatching
is enabled is a good idea.
For simplicity, this commit also restricts spying to method calls
only. Signals are no longer spyable.
Change-Id: I3d11545be52c43119f0fffff142b0e9d447415c2
Reviewed-by: David Faure <david.faure@kdab.com>
From Qt 5.7 -> tools & applications are lisenced under GPL v3 with some
exceptions, see
http://blog.qt.io/blog/2016/01/13/new-agreement-with-the-kde-free-qt-foundation/
Updated license headers to use new GPL-EXCEPT header instead of LGPL21 one
(in those files which will be under GPL 3 with exceptions)
Change-Id: I42a473ddc97101492a60b9287d90979d9eb35ae1
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Lars Knoll <lars.knoll@theqtcompany.com>
From Qt 5.7 -> LGPL v2.1 isn't an option anymore, see
http://blog.qt.io/blog/2016/01/13/new-agreement-with-the-kde-free-qt-foundation/
Updated license headers to use new LGPL header instead of LGPL21 one
(in those files which will be under LGPL v3)
Change-Id: I046ec3e47b1876cd7b4b0353a576b352e3a946d9
Reviewed-by: Lars Knoll <lars.knoll@theqtcompany.com>
The reply serial is displayed for method call returns and errors,
while the serial is displayed for all message types.
To see a message serial it is required to dump messages after
sending, not before.
Task-number: QTBUG-44490
Change-Id: I859f50d739ed059d5b2dfe1a2efdf04b906891a7
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
To retain a bit compatibility with applications developed in the last 9
years that expect that QDBusConnections won't process their events until
the event loop runs, we now suspend the handling of incoming messages
in the two default buses (and only in them) and resume when the event
loop starts. This is required because the new threaded QtDBus would
otherwise process incoming messages that the application didn't expect
it to.
For example, if the application first acquires names on the bus and only
after that registers objects with QtDBus, there's a small window in
which the name is acquired and visible to other applications, but no
objects are registered yet. Calls to those objects may be received,
would then be processed in the QDBusConnectionManager thread and fail.
The work around is to disable the actual handling of method calls and
signals in QDBusConnectionPrivate::handleMessage. Instead, those
messages are queued until later.
Due to the way that libdbus-1 works, outgoing method calls that are
waiting for replies are not affected, since their processing does not
happen in handleMessage().
[ChangeLog][Important Behavior Changes] QtDBus now uses threads to
implement processing of incoming and outgoing messages. This solves a
number of thread safety issues and fixes an architectural problem that
would cause all processing to stop if a particular thread (usually the
main thread) were blocked in any operation. On the flip side, application
developers need to know that modifications to a QDBusConnection may be
visible immediately on the connection, so they should be done in an
order that won't allow for incomplete states to be observed (for
example, first register all objects, then acquire service names).
Change-Id: I39cc61d0d59846ab8c23ffff1423c6d555f6ee0a
Reviewed-by: David Faure <david.faure@kdab.com>
Calling QVector::erase(it) in a loop consitutes quadratic
behavior (O(N) function called O(N) times).
Fix by using std::remove_if(), which is linear.
Change-Id: I39c11231d604bc2d9506427bc3411b71d71b5569
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
If, after checking a condition, we issue a qFatal()
or a qCritical(), by definition that check is
unlikely to be true.
Tell the compiler so it can move the error handling
code out of the normal code path to increase the
effective icache size.
Moved conditional code around where possible so that
we could always use Q_UNLIKELY, instead of having to
revert to Q_LIKELY here and there.
In some cases, simplified the expressions newly wrapped
in Q_UNLIKELY as a drive-by.
Change-Id: I67537d62b04bc6977d69254690c5ebbdf98bfd6d
Reviewed-by: Konstantin Ritt <ritt.ks@gmail.com>
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
Any connect requiring a lambda to be ported or function casts were not touched
Change-Id: I1718121986ba6632b5754efa631f7b599358e186
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Normally, disconnectNotify() is called at the end of QObject::disconnect
and all the locks have been dropped. That is not the case for the
QObject destructor, so we need to deal with the fact that it there may
be some locks held.
I didn't catch this issue during testing because it depends on the
pointer addresses of the object being destroyed and that of the
QDBusAbstractInterface sender object, as we use one global, non-
recursive mutex pool. For the same reason, this patch is not testable.
The fix is simple: we don't need to remove the relay rules immediately.
It's ok for them to happen later, since the worst case scenario is that
we'll receive a few more signals than we have objects to deliver them
to. If that happens, we'll do a little more work than we have to. But in
the normal case, the amount of work is the same and we get the benefit
of returning more quickly from the destructor. What's more, if the
QDBusAbstractInterface object also gets destroyed, the events are
deleted and QDBusConnectionPrivate will clean everything up.
Task-number: QTBUG-48410
Change-Id: I42e7ef1a481840699a8dffff1406b789ba5217b3
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
QDBusConnectionPrivate can only be a client or a server, not both, so
the DBusServer and DBusConnection pointers can be shared, like the
QDBusConnectionInterface and QDBusServer pointers in the other anonymous
union.
Change-Id: I9a75ad8521ae4e5cbbe5ffff13d1baa8ab83c42f
Reviewed-by: Albert Astals Cid <aacid@kde.org>
Reviewed-by: Alex Blasche <alexander.blasche@theqtcompany.com>
This commit moves the code that finishes the signal-slot connection into
the QtDBus auxiliary thread. That is necessary because we're holding the
lock for writing while making blocking calls. The auxiliary thread might
be waiting for us to release that lock while processing some previous
message.
Change-Id: Iee8cbc07c4434ce9b560ffff13d0521b94a51833
Reviewed-by: Albert Astals Cid <aacid@kde.org>
Reviewed-by: Alex Blasche <alexander.blasche@theqtcompany.com>
This is an optimization but is required. Instead of going through the
entire (dis)connectSignal() stack to add/remove matching rules for the
NameOwnerChanged bus signal and call into our serviceOwnerChangedNoLock
slot, create a static hook that will match the signal and simply add/
remove the rules as needed.
The required part is that this avoids a recursion into connectSignal().
The next commit will move this code to the QtDBus manager thread and we
won't be able to call connectSignal() from there (it would create a
deadlock).
Change-Id: Iee8cbc07c4434ce9b560ffff13d074ce90ad02d4
Reviewed-by: Albert Astals Cid <aacid@kde.org>
Reviewed-by: Alex Blasche <alexander.blasche@theqtcompany.com>
In two commits, we will attempt to call this function from the manager
thread, so we need to be sure this function works from there. Right now,
it would deadlock in QDBusPendingCallPrivate::waitForFinished(), inside
QDBusConnectionPrivate::sendWithReply().
The solution is simple: expand sendWithReply to the sendWithReplyAsync
function it calls anyway, but tell the internal DBusPendingCall to
finish before we call waitForFinished().
Change-Id: Iee8cbc07c4434ce9b560ffff13d0749013d771ab
Reviewed-by: Albert Astals Cid <aacid@kde.org>
Reviewed-by: Alex Blasche <alexander.blasche@theqtcompany.com>
With kdbus, we won't have a regular signal, but instead a special
message. So keep the logic of what to do in QDBusConnectionPrivate.
The #ifdef is to make sure the bootstrapped qdbuscpp2xml continues to
build in cross-compilation environments.
Change-Id: Iee8cbc07c4434ce9b560ffff13d06f0d9904cb6d
Reviewed-by: Alex Blasche <alexander.blasche@theqtcompany.com>
They were used when we called the libdbus-1 message-sending functions
from any thread, which meant that the callbacks could be triggered on
any thread. Since we moved the message-sending to one thread only (the
manager's thread), there's no need for the event fallback anymore.
Since they're also always[*] running on an aux thread, there's no point
in checking for the presence of a QCoreApplication instance anymore. I
don't think there has been a need for this for many years, as the event
dispatcher has been decoupled from QCoreApplication for a long time.
[*] exception: the callbacks are first called in the thread that invoked
QDBusConnection::connectTo{Bus,Peer}, before we've done the
moveToThread.
Task-number: QTBUG-43585
Change-Id: Ic5d393bfd36e48a193fcffff13b73758c798d6b0
Reviewed-by: Albert Astals Cid <aacid@kde.org>
Reviewed-by: Alex Blasche <alexander.blasche@theqtcompany.com>
The two global statics for the session and system buses aren't necessary
if they can't outlive the global static for QDBusConnectionManager
anyway. So merge them there. The extra mutex is necessary because the
QDBusConnection::connectToBus function will lock the regular mutex.
This solves a potential memory leak at exit as a side-effect. Before
this change, the session and system QDBusConnection object got destroyed
in the main thread during global destruction, so it had to post an event
to the QDBusConnectionManager thread to finish the destruction of the
private. However, QCoreApplication is already gone by this point, so the
QEvent::DeferredDelete event never got delivered.
After this commit, there's no global static to destroy the
QDBusConnection (there is no QDBusConnection holding a reference), so
the object gets destroyed in QDBusConnectionManager::run()'s cleanup
code.
Change-Id: I9a75ad8521ae4e5cbbe5ffff13d1b967ee1a7a7e
Reviewed-by: Albert Astals Cid <aacid@kde.org>
Reviewed-by: Alex Blasche <alexander.blasche@theqtcompany.com>
