Latest QNX versions support _readdir64_r. It will be used by
QFileSystemIterator now, when available to support large file
systems.
Change-Id: I40135f405ad72e1ba85c6a4cd282144dd8171bc6
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Thomas McGuire <thomas.mcguire@kdab.com>
Reviewed-by: Rafael Roquetto <rafael.roquetto@kdab.com>
There is a known bug where duplicate textChanged signals are
triggered on backspace/delete. The bug has been seen on Windows,
Mac, and Linux. Gabi showed that this duplicate signal is caused by
two calls to finishEdit(), one direct and one nested,
in QTextCursorPrivate::remove().
To attempt a fix, I removed the direct call and do not change the
nested call. This change only affects text buffers when they
receive remove commands.
This seems to fix the problem, shown by the test project
uploaded to the bug tracker and also in countTextChangedOnRemove
in tests/auto/widgets/widgets/qtextedit/tst_qtextedit.cpp.
Further analysis of finishEdit() and QTextCursorPrivate::remove():
finishEdit() calls signals contentsChanged() as long as a valid block
is being edited. Remove() calls finishEdit for all non-table edits,
so finishEdit will be called for most cases.
Methods in the public QTextCursor all seem to set adjusted_anchor
and anchor, none reading it directly, so I haven't found publicly
observable consequences to "adjusted_anchor = anchor = position;".
Task-number: QTBUG-15003
Change-Id: Ic35f25ee81c4007867b47cd8be03c146a673f86d
Reviewed-by: Graham Coleman <ravelite@gmail.com>
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>
constexpr requires inline functions. We convert more member functions
to inline functions while retaining a version of those functions
inside QtGui. This prevents any BC breakage.
Change-Id: I325b78da6b41611bda3994869f474c700b6fb306
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
Be more rigorous about checking if we're in a session before invoking IMF functions. In
particular the code would call initEvent if a key was received while no text field had focus
leading to a crash.
Also closeSession is now always called when a session is terminated rather than just
setting sInputSession to 0 and hasSession is always called to check for a valid session.
Change-Id: I22ce1c9bcdf8dd130f05ae319c4cc1e01afc4456
Reviewed-by: Kevin Krammer <kevin.krammer@kdab.com>
Reviewed-by: Fabian Bumberger <fbumberger@rim.com>
Reviewed-by: Bernd Weimer <bweimer@blackberry.com>
Reviewed-by: Rafael Roquetto <rafael.roquetto@kdab.com>
Starting with MSVC2013, a separate set of libs for Windows Store apps is
supplied, so make sure it is in the LIBPATH (and before the desktop libs).
Change-Id: I74f3f385c2db749010fbfe7e2d4c3d1228e4e603
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@digia.com>
getenv isn't available on this platform, and has been stripped from the
lib starting with MSVC2013.
Change-Id: I5a69055b7c3ca85dc6d5a74c5b2842e7886b13bd
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@digia.com>
Make it possible to match messages by a pattern.
Change-Id: I713312e86db5471755459f1ecc43e8f1ac7a95fb
Reviewed-by: Jason McDonald <macadder1@gmail.com>
Enable use of ANGLE on lower-end hardware, such as Surface RT and
Windows Phone 8.
Based on https://codereview.appspot.com/12917046/
Change-Id: Ice536802e4eedc1d264abd0dd65960638fce59e4
Reviewed-by: Kai Koehne <kai.koehne@digia.com>
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@digia.com>
This turns QVector2D(1,1) into compile-time expressions when using
a C++11 compiler.
Change-Id: Ifea2d1cdfd5481300f9a5a0c7bbe8bfe9115c636
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>
Reviewed-by: Sean Harmer <sean.harmer@kdab.com>
The xcb QPA plugin uses a separate thread for getting events from the
X11 server in a race-free manner. This thread has to be stopped when
shutting down. This is done by causing the X11 server to send a special
event that causes the thread to exit when it sees it. (Also, the thread
exits if the xcb connection goes into an error state, but that does not
matter here)
So this event listener window is only needed when Qt is sending events
to itself. This means that it does not have to be kept around and
instead can be created only when needed.
Since this window does not stay around for long, it does not need a
name. However, it does need the list of screens to find the correct root
window, so destroy screens later.
Change-Id: Ib4104cbc3f372fa5d1391bcf5a633a158409862f
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@digia.com>
Reviewed-by: Shawn Rutledge <shawn.rutledge@digia.com>
Compared to other platforms there is no concept of a console
application in WinRT. Hence all applications need to be UI
applications and use winmain.
Furthermore winmain takes care of launch arguments to be
properly converted to arguments passed to user's main().
There is a chicken and egg problem with config.tests as
compilation needs to have an existing entry point which is not
available at configure time.
Hence hardcode the entry point to main for configuring to WinRT.
Those tests are pure compile tests, so the logic of the test
does not change.
Change-Id: I4d3186691a8440845c24b2529cc9646e86dfd8da
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@digia.com>
The updated sqlite3 lacks a forward declaration of localtime.
Depending on the CE version that forward declaration
was sometimes available.
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@digia.com>
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@digia.com>
(forward-ported from commit fb1649d30b)
Change-Id: Ief6d8ed1cad51fa92a333a20c5dfe781d19761eb
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@digia.com>
(forward-ported from commit ea70ec8711)
Reviewed-by: Björn Breitmeyer <bjoern.breitmeyer@kdab.com>
Reviewed-by: Mark Brand <mabrand@mabrand.nl>
After profiling drawing with RGBA8888 images most of the time appears
to be spend in converting to and from ARGB32PM.
This patch adds four small converters that are 3-4 times faster than the
generic converter.
Change-Id: I3c7498756f440ca3ea9c1417b26dd8e1953b9d06
Reviewed-by: Gunnar Sletta <gunnar.sletta@digia.com>
Implement swap interval support for EGL, GLX and WGL.
The environment variable QT_QPA_EGLFS_SWAPINTERVAL is renamed to
QT_QPA_EGL_SWAPINTERVAL and can be used to override the applications'
setting of the swap interval.
Task-number: QTBUG-31939
Change-Id: I644325d5d3306b7604bffd7efccda3c00ed37d36
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@digia.com>
Reviewed-by: Giuseppe D'Angelo <giuseppe.dangelo@kdab.com>
Reviewed-by: Gunnar Sletta <gunnar.sletta@digia.com>
need to leave some items out like media folder access, as this is not
available by default and also requires certain capabilities to use
those.
Furthermore updated the tests for sandboxing as well as skip cmd.exe
related tests as that does not exist on WinRT.
Change-Id: I992b1e195b79615bea0be4f84f56cfb8f0d902bf
Reviewed-by: Oliver Wolff <oliver.wolff@digia.com>
FILE_FULL_DIR_INFO failed under certain circumstances on WinRT and
provides far more information than required. Hence prefer the basic
version in every case and standard for files with size description.
Change-Id: I63f1365f83cdd5d69f81278411f822dbd361fa92
Reviewed-by: Oliver Wolff <oliver.wolff@digia.com>
Currently network is only required to gather the host info for one
test case. That does not justify to disable all other tests, which can
provide useful information on the state when doing a port. Hence
disable that testcase if no network is available.
Change-Id: I202ef49b3e07ae69ec85ee0432ae0a771a90e816
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@digia.com>
Enums are movable even if they have not been tagged as such, and
therefore can safely make use of the internal storage.
Change-Id: I3bec8556bfeb4a07cb30d918f8d1dd2565d6d94a
Reviewed-by: Jędrzej Nowacki <jedrzej.nowacki@digia.com>
While exception handling methods are not available on WinRT, the debug
mode ones are. This way we can suppress error dialogs from popping up
like on desktop windows.
Change-Id: Id6e9666af9274d373e7d46e3f9e44892a2bd0dbc
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@digia.com>
Fixes several things
* setOption(options) has the wrong name (singular)
* setOption(options) doesn't set the current options to the argument, but
ORs the current options with the argument and sets the result
* testOption(options) has the wrong name
The old methods get deprecated, and new methods and overloads get
introduced here. Old code behavior is thereby preserved.
Change-Id: I51bba49f22810c80e6b4980892600d616503af6b
Reviewed-by: Gunnar Sletta <gunnar.sletta@digia.com>
Updated dnslookup example to take an optional nameserver argument
for doing DNS lookup against a specific nameserver.
Task-number: QTBUG-30166
Change-Id: I9f46f9f766b56f770d2c8372e3bfad5c71023c73
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
otherwise we are risking a lock by freeing resources from other
threads when calling CoUnitialized()
Change-Id: I1fe09d7d3b1674a00b44ababfc90da79580278b3
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@digia.com>
The actual feature comes from a merge from the stable branch.
Change-Id: Id067fb2b26cb3c6b78c7705b77bd741e3bd8f090
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Specify the root directory to be the package root. Only plugins
inside the root can be opened (actually also only files). Furthermore
current defaults to the package root now, which in most cases is
identical to previous behavior.
When attempting to load a plugin the path can either be specified in
host format "C:/..." or as plugin absolute "/platforms/...". Check for
both, with preference of latter case, like when qt.conf is used with
/ being used as plugin path.
Change-Id: I7e3da293362488b62a3357c4882ebf5e048dcf95
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@digia.com>
Reviewed-by: Andrew Knight <andrew.knight@digia.com>
Reviewed-by: Oliver Wolff <oliver.wolff@digia.com>
...when the buffer's size is smaller than the value size.
Change-Id: Id345982c9fc4ceed6505d0c192680c47c554fcb4
Reviewed-by: Oliver Wolff <oliver.wolff@digia.com>
Winmd is not used, so there is no reason to embed it.
Change-Id: I0820256aecd9c3c71b0b0c8afa53941b03f97363
Reviewed-by: Oliver Wolff <oliver.wolff@digia.com>
WinRT doesn't use the Windows Timezone backend, so don't build the
Windows test.
Change-Id: I32620546de3ad1f19402cc1359f8038200c915ec
Reviewed-by: Oliver Wolff <oliver.wolff@digia.com>
This should be using qputenv; putenv is deprecated on MSVC.
Change-Id: I7c27cf5f7955624fa3553b7a34ab11c6fae462b8
Reviewed-by: Oliver Wolff <oliver.wolff@digia.com>
Visual Studio debuggers send an argument to specify the debugger port.
This argument needs to be skipped while parsing and not to be interpreted
as an option or test function.
Change-Id: I24efb52fbd668a7bc3388c876f5ea0d950de1d5b
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@digia.com>
QByteArray split does create one item even if the string is empty.
Hence check if the launch arguments string needs to be parsed at all.
Change-Id: I0a355212aaa7254fe0f417c61a59c30223311915
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@digia.com>
The current synchronization mechanism was racy: decrementing waitingThreads
and then hoping that the wakeOne will wake a thread before its expiry
timeout happens. In other words, on timeout, a just-assigned task would
never run. And then no other task would run, if maxThreadCount is reached.
Fixed by using a queue of waiting threads (rather than just a count), and by
moving the wait condition into the thread itself, so we know precisely
which one we're waking up, and we can remove it from the set of waiting threads
before waking it up, and therefore it can determine on wakeup whether it
has work to do (caller removed it from the queue) or it expired (it's still
in the queue). This is reliable, whereas the return value from QWaitCondition::wait
isn't reliable, when the main thread has already decided that this thread
has work to do.
Task-number: QTBUG-3786
Change-Id: I1eac5d6c309daed7f483ac7a8074297bfda6ee32
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
This change changes the point where signals are unblocked again in
QDockAreaLayoutInfo::updateTabBar() to include the final tabBar->count()
call. This should not be a problem, as count() is a const function and
thus shouldn't emit any signals.
Change-Id: I6f3dc5696a9c31db51fbe4cdee4b9d83ddeaf61f
Reviewed-by: Olivier Goffart <ogoffart@woboq.com>
This change changes the point where signals are unblocked again in
QAbstractSpinBoxPrivate::updateEdit() to include the final update()
call. This should not be a problem, as update() merely posts an
event and shouldn't emit any signals.
Change-Id: I4b2ae109f057792b573ad6ea168ca77c18c773d0
Reviewed-by: Olivier Goffart <ogoffart@woboq.com>
