The problem was caused by commit 0feeb6f6d2
which fixed QTBUG-58036. It reparented widget that used to be skiped.
In particular, floating widgets are skiped. But seting the parent of a
floating widget docks it. And so as a result it would not be skiped anymore.
This has two side effect: This breaks the animation (as the widget is
docked too early; and cause crash when QDockWidgetGroupWindow get
reparented as this breaks invariant that these are always floating.
So restore the skip from before commit 0feeb6f6d2,
and explicitly set the parent in all cases when the animation finishes.
Change-Id: I0e3e29ad22d6ffe4d62242d48a18dadb916fc14f
Reviewed-by: Sérgio Martins <sergio.martins@kdab.com>
The 'rules' vector is made up of all the individual {env,config,...}Rules
vectors under mutex protection whenever init() is called (only from the
QCoreApplication ctor) or, at any time, by a call to QLoggingCategory::
setFilterRules().
Yet, the writes to the individual *Rules vectors were never protected by
registryMutex, racing against the reads of the same vectors in the
updateRules() function.
Fix by protecting all access of all member variables with registryMutex.
Add some strategic comments to make analysis easier for the next guy.
Change-Id: If68d15a553ec7038693574a34f10a39f4cd480e8
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
In this case we can safely return the same QSize which we accept since
it is already contains the right size for the given menu and this size
will be bounded to screen geometry before QMenu will be displayed
anyway. We also get rid of one dependency on HITheme.
Change-Id: I7502a96d180fc4a41ce3dfabe8a200b886016348
Reviewed-by: Gabriel de Dietrich <gabriel.dedietrich@qt.io>
AİİA.pdf takes 8 QChars, but after lowercasing it takes 10, so the code cannot
assume the length to be the same.
Task-number: QTBUG-58822
Change-Id: Id6fbb99f6afd08ee420099cd66372732d7598d9e
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
An Acquire barrier in QReadWriteLocker::unlock was missing to synchronize
with the testAndSetOrdered on d_ptr in the lock functions.
The race is between the write of d->writerCount in tryLockForWrite,
and the read in unlock. The acquire on d->mutex is not enough because
it is not on the same object. While that race could be fixed by taking
the newly-allocate()ed d's mutex before publishing it with testAndSet,
there's another race, on 'recursive', between a newly-minted Private*
with recursive == false in tryLockForWrite(), and the read of 'recursive'
in unlock().
Task-number: QTBUG-58917
Change-Id: I10ba36573c0e57468d11e9b77d85045711feaea1
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>
see 5a1b4832a2 for more detail
Task-number: QTBUG-58142
Change-Id: I51851ea9b4fe7b8eeadc452bc3dbb1ea00026d29
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
The _next variable need the acquire and release fence in next()
to synchronize with the equivalent operations in release() (which
already have the them)
The ordering on the _v[block] is not enough as this does not
synchronize the same object.
Task-number: QTBUG-58917
Change-Id: I17cc39e6791433348b6227363dbea92bcf03700d
Reviewed-by: David Faure <david.faure@kdab.com>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Like the lvalue QVector::append() overload, when we reallocate,
we need to take a copy of the function's argument because the
reference will get stale upon reallocation.
Add a test.
[ChangeLog][QtCore][QVarLengthArray] Fixed a bug involving
appending an item already in the container to the container
again.
Change-Id: I06eeed6cb383dd5924e47a302bb3d1666d04c8e8
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
In QCocoaDrag::dragPixmap, it treats QDrag::source as a QWindow, but it
is not - it's just a generic QObject* of some kind (which QQuickDrag
sets to the originating QQuickItem, and the widgets stack sets to a
QWidget). This failure means that dpr stayed at 1.0.
Unfortunately it’s not possible to receive a pointer on QWindow directly
from QQuickItem because QtWidgets and QtQuick do not share the sources,
but we can use the same dpr as current focused window has because drag
can only start from active window - press on a window which is not
focused should activate it first.
Task-number: QTBUG-57942
Change-Id: Id358c181d03d519188caaa83fb4226033b8ed1ea
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
Reviewed-by: Robin Burchell <robin.burchell@crimson.no>
Automount filesystems like /home on many operating systems (QNX and
OpenIndiana, at least) don't like if you try to mkdir in them, even if
the file path already exists. OpenIndiana even gives you an ENOSYS
error.
So instead, let's try to mkdir our target, if we fail because of ENOENT,
we try to create the parent, then try again.
Task-number: QTBUG-58390
Change-Id: Ibe5b1b60c6ea47e19612fffd149cce81589b0acd
Reviewed-by: James McDonnell <jmcdonnell@blackberry.com>
Reviewed-by: David Faure <david.faure@kdab.com>
QNetworkSession's ctor has a race condition: it 1) connects signals/slots
and also 2) registers meta-types (but after these connects). Our users apparently
have a lot of per-thread QNAMs in multiple threads (and implicitly
many QNetworkSessions on different threads too). From error logs it appears
that while one thread tries to connect signals/slots and evaluates the types
of signal and slot arguments, another thread can register this type.
If the first thread extracted signal argument types, then the second
registered this type, we can end up in a 'type mismatch' error on the first
thread with seemingly the same types in a debug message (something like
"type mismatch A::Some <-> A::Some") - they have the same name, but one
has type() == 0 and another - some non-zero type().
Now we call qRegisterMetaType before connect.
Task-number: QTBUG-50901
Change-Id: Idbb9515573e174bbc5e5bf55dc3a7911a81646ea
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
The library feature already covers this. As library depends on the
dlopen compile check, we can assume dlopen to be available, also on
vxworks.
Change-Id: Idcdb07ab4688c6158651d9a5ad5e2ba126bd7d9e
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@qt.io>
It is never written after initialization, and accessed concurrently
outside mutex protection, so make sure that it never changes by
marking it as const.
Change-Id: Ib28b2e8b3067d596a95d1699165a6ebfb15fe674
Reviewed-by: David Faure <david.faure@kdab.com>
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
Adding 0 each time will obviously not produce a new identifier each
time...
Also use static initialization for QBasicAtomicInt.
A default-constructed static QBasicAtomicInt at function scope
will be dynamically initialized. It will still be zero-initialized,
but at least GCC adds guard variables for such objects.
When using aggregate initialization, the guard disappears.
Amends 265db5ad9b.
Change-Id: Ia71290cf26c486dcbcc74381f12cd0c4712d6019
Reviewed-by: David Faure <david.faure@kdab.com>
Use of XInput 2 may result in regressions. For example we call
XISelectEvents for each window every time a device plugs/unplugs.
It causes significant delays when there are many native windows.
Change-Id: I60b799bb667d0e4bca1f9c52cdaa07b04bcc749f
Task-number: QTBUG-57013
Reviewed-by: Gatis Paeglis <gatis.paeglis@qt.io>
QDeadlineTimer will eventually support CLOCK_MONOTONIC_COARSE on Linux,
so let's be safe before that change appears.
Change-Id: I9b423f1b71194205b27afffd14a383a7e9f66a9d
Reviewed-by: Stéphane Fabry, Cutesoft <stephane.fabry@gmail.com>
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
It was unnecessary, since we only cached the static types, which are all
generated by the macro anyway. The way it was implemented, this produced
data races that are strictly-speaking UB, even if all the threads were
writing the same values to the same data locations.
This commit changes a little the code to simplify, since we're changing
those lines anyway.
Task-number: QTBUG-58851
Change-Id: Idc5061f7145940f987dffffd14a30047846e3113
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
When the QFileDialog receives a language change event, and native
dialogs are used, this code path dereference a nullptr in the qFileDialogUi member.
Protect by checking the return value of QFileDialogPrivate::usingWidgets() as
done in other places deferencing qFileDialogUi.
Fixes a crash in GammaRay
Change-Id: I3e9eb81c75d9fc983519a58c5ec06e316ac80263
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>
If the QSemaphore::tryAcquire() call times out, we mustn't
touch *res, because there was no happens-before relation
established between *res = result in the lambda and our
returning *res;
Fix by returning a default-constructed hash in that case.
Add a strategic std::move().
The same problem exists in runOnAndroidThreadSync(), but
I have no idea how to solve it, because there the shared
object is the runnable itself.
Change-Id: I9a2c431144c169fbd545763555d96153143a11bf
Reviewed-by: BogDan Vatra <bogdan@kdab.com>
The code obtained an iterator into a QHash under mutex protection,
then dropped the lock, dereferenced the iterator several times and
only retook the lock to erase the element from the QHash.
This is very smelly. QHash provides no official iterator validity
guarantees, and the container isn't const, either (which would imply
thread-safety). In particular, the dereference into the container
outside the critical section is cause for concerns.
Simplify the code, removing any doubts about its race-freedom, by
taking the payload item out of the hash before dropping the lock, and
using only the local strong reference in the remainder of the
function.
The only other references to g_pendingPermissionRequests are
insertions with unique-by-construction keys in QtAndroidPrivate's
requestPermissions(), so there was no reason to keep the item in the
hash for the whole duration of the sendRequestPermissionsResult()
call.
Change-Id: I39fe0803b13b3046d1f0fd9c8e96c531406d57da
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
Reviewed-by: BogDan Vatra <bogdan@kdab.com>
Update the QSyntaxHighlighter examples to use the new QRegularExpression
class in place of QRegExp.
Fix typos. Remove duplicated snippet. Replace lengthy section of
duplicate text in highlight(..) with a note to see the detailed
description.
Task-number: QTBUG-58494
Change-Id: Id8d94bddbed52e6e52feac107f6fc84e2fe4518a
Reviewed-by: Samuel Gaist <samuel.gaist@edeltech.ch>
Reviewed-by: Sze Howe Koh <szehowe.koh@gmail.com>
Solves a data race found by TSan.
Since thread and threadId are QAtomicPointer, I've removed the explicit
initialization in the QThreadData constructor
Task-number: QTBUG-58855
Change-Id: I4139d5f93dcb4b429ae9fffd14a34082f2683f76
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>
1. Do not use Q_GLOBAL_STATIC to hold QAtomicInt or QMutex, use
file-static QBasicAtomicInt and QBasicMutex instead. They are
zero-initialized PODs.
2. Use only QMutexLocker to lock mutexes.
Also wrap the atomic counter into a next...() function, as done
elsewhere.
Change-Id: I4b14ac0de9d4cb6780b1f1372c2b5fc88e918e4c
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
$ git grep -we futexFlagSupport
src/corelib/thread/qmutex_linux.cpp:static QBasicAtomicInt futexFlagSupport = Q_BASIC_ATOMIC_INITIALIZER(-1);
The last user of this variable was removed in 9ef59b5.
Change-Id: I818a13a481ad25baa5ff7d389a737b8801adcfcc
Reviewed-by: David Faure <david.faure@kdab.com>
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
NVIDIA 370 introduced new calls needed to work with eglstreams in the
wayland window system
Task-number: QTBUG-58299
Change-Id: I606b143d3016f365b0d5ca4bc163b99289afbea1
Reviewed-by: Laszlo Agocs <laszlo.agocs@qt.io>
Add proper support for 'so' and 'bundle' suffixes. Qt wrongly assumes
.so libraries are not versioned on Apple platforms, which is
wrong. Also, the shared library .bundle which is what Apple recommends
instead of .so, are also versioned (not to be confound with the
different Core Foundation bundles, which are directory hierarchy).
For more info, see http://docstore.mik.ua/orelly/unix3/mac/ch05_03.htm.
Especially the part that reads:
"Loadable modules, called bundles in Mac OS X, have the file type
MH_BUNDLE. Most Unix-based software ports usually produce bundles with a
.so extension, for the sake of consistency across platforms. Although
Apple recommends giving bundles a .bundle extension, it isn't
mandatory."
Task-number: QTBUG-50446
Change-Id: Iacd5136397a12d65d83821434f332eb602550b4b
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
And if __AES__ or __PCLMUL__ is defined, then we expect wmmintrin.h to
exist.
Change-Id: I445bb15619f6401494e8fffd149db77dc513e071
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
Since 5.7, QIODevice::peek() implementation is based on transaction
mechanism. While technically it's correct, seeking backward on a
buffered random-access device clears the internal buffer that affects
the performance of reading.
To solve the problem, this patch implements peek mode directly inside
the reading procedure.
Task-number: QTBUG-56032
Change-Id: Ic5269f76e44c491a0309e13aba87fa7cf7b9259f
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Move device-dependent part of the code into the private function for
further reusing by peek() procedure.
Task-number: QTBUG-56032
Change-Id: Iedceafe4b0bab109ca5c64ad274d779efe87c27b
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
the file contains no code.
this avoids complaints from ar/ranlib in static/bootstrapped builds.
Change-Id: Iee22ffc61a5f9ea8c25f5455b7e8e017ac521624
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
... instead of merely defaulting it to off on android and windows.
this reflects actual reality.
Change-Id: I880254138bedd07124aa00096a06dd6e1803feb9
Reviewed-by: Laszlo Agocs <laszlo.agocs@qt.io>
Since they are all used in a typical application, this reduces the
number of memory allocations (thus, the overhead) as well as the
state-keeping in the libc atexit() functions.
Change-Id: Ifaee7464122d402991b6fffd14a0e59457ad9cb7
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
If these lists weren't created in the first place, then they are empty.
We don't need to create it in order to conclude that. Unlike most
Q_GLOBAL_STATICS, these are almost never used and yet they were
always created due to where they were checked.
Since we're calling exists() before, there are two consequences: first,
since the list already exists, we're not allocating memory so it cannot
throw std::bad_alloc when being accessed. Second, since we've just
checked it exists, we can use QGlobalStatic's operator*(), which is
slightly faster than operator()(). The weird &(*list) syntax is only to
avoid changing the rest of the code that used a pointer
Change-Id: Ifaee7464122d402991b6fffd14a0e44f533dc3d9
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>
exporting the freetype dependency caused qtwayland to break with
-qt-freetype, as the helper libraries' module pri files are not
installed (for good reasons)
after f9a80e06a, no actual user of FontDatabaseSupport needs access to
the transitive dependencies anyway (one of the headers has a fontconfig
dependency, but it's not used outside the module itself), so hiding them
again is just fine.
this partially reverts ec774500f.
Task-number: QTBUG-56666
Change-Id: I9e68a7e0725a92833b856c9ffdbec61c8aa5fed2
Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
that's where the files are actually used nowadays.
also removes an obsolete (and bogus) gui export.
Change-Id: I4551aad798acb6ce8c0abe43a2fcb8e5ac64a2d4
Reviewed-by: Konstantin Ritt <ritt.ks@gmail.com>
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@qt.io>
I didn't even try to understand what the old code was trying to do;
once you're told by a user that the code is wrong, you see that it is.
Fixed by just using the row as passed to takeRow() instead of trying
to do some storage-index calculations. The m_matrix indexing operator
does it all for us.
Added a test that checks that the expected field widget gets
returned. Fixed expected test data that was wrong, and just checking
that the implementation behaves as implemented, instead of as
documented.
Amends change 8fbae648db.
[ChangeLog][QtWidget][QFormLayout] The functions takeRow() and
removeRow(), new in 5.8.0, now take and remove the correct row.
Task-number: QTBUG-58693
Task-number: QTBUG-15990
Change-Id: I7185ccbc6c03e2579741cad5c0c821d3ed165474
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
Olivier Goffart
Marc Mutz
Oleg Yadrov
David Faure
Shawn Rutledge
Eric Lemanissier
Florian Bruhin
Thiago Macieira
Timur Pocheptsov
Ulf Hermann
Alexander Volkov
Dmitry Shachnev
Samuel Gaist
Kevin Funk
Frederik Schwarzer
Aaron Linville
Heikki Haveri
Jesus Fernandez
Nico Vertriest
Friedemann Kleint
Oswald Buddenhagen
Dominik Holland
Filipe Azevedo
Alex Trotsenko