| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
|
|
|
|
| |
The crash could happen if a QWidget in the UI got deleted, and
the AT client later tried to access the widget through the cache
(qAccessibleRecentSentEvents()).
Solution: Use a QPointer as a guard.
Task-number: QTBUG-26187
Change-Id: I41eab158989fddfa147309b6bd91ac0cd1fe7b1a
Reviewed-by: Frederik Gladhorn <frederik.gladhorn@digia.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
For moc, rcc and uic, then it's friendly for tools like ccache.
ccache is using md5 to check file modification, but the different
timestamp info will cause different md5 for same meaningful
contents, it will disabled ccache.
Updated the autotest for uic and rcc.
Task-number: QTBUG-26589
Change-Id: I9e72ff022f722809b3f86bd9d7e0ae1e470e8eb2
(back port from qtbase/139f416237c52575b236c3b61e25796c83034567)
Reviewed-by: Olivier Goffart <ogoffart@woboq.com>
|
|
|
|
|
|
|
|
|
| |
... which is important for an empty but non-null authority.
In Qt5 this is already working.
Change-Id: I7d389037f71320c6f06897b220633311c3611eea
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
|
|
|
|
|
|
|
|
|
|
|
| |
This was already fixed for toString() with commit 884f15e, this commit
also fixes the toEncoded() implementation.
In Qt5 this is already working.
Task-number: QTBUG-8701
Change-Id: I0c81801e74d97de31632df45e52a25eb037b4710
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
|
|
|
|
|
|
|
|
|
|
|
| |
Patch-by: Greg Bentz
Task-number: QTBUG-28247
(backport of commit b6250913be2fecedbc61738a47ff147f0a4ea240,
Id Ibd1f959969a8678b828c8e8de8be0a200a632b22)
Change-Id: I5e5b85fb0fce52ba50a1600bf91212f46b2597c8
Reviewed-by: Greg Bentz <gbentz@rim.com>
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@digia.com>
|
|
|
|
|
|
|
|
|
| |
composition_mode is only defined if QT_NO_XRENDER is not defined.
Bug was introduced with 439aa67582e715ddc0ca7e30305b0c9498a71f81
"Fixes ignored composition mode in QPainter::fillRect with X11 engine."
Change-Id: I2dc036ffa277cbf0f9327ede3fead9a407147fc0
Reviewed-by: Samuel Rødal <samuel.rodal@digia.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Adding QCheckBox, QComboBox, QLineEdit etc. into an HBox layout or grid
layout makes the layout look like a snake's trail. Fixed the positioning
of these widgets to make the layouts visually more appealing.
Updated qmacstyle auto-test accordingly.
Task-number: QTBUG-13635
Ammended-by: Gabriel de Dietrich
Ammended-by: Liang Qi
Change-Id: Ic12346f764d8092292b8388f50d184b37d7cef5f
(cherry picked from qtbase/8b6de4677142cab41790ca50ce6c7f1efbab5de8)
Reviewed-by: Gabriel de Dietrich <gabriel.dedietrich@digia.com>
|
|
|
|
|
|
|
|
|
|
| |
Use localized family name and style name when selecting font with
non-English locale
Task-number: QTBUG-27415
Change-Id: Ia10aaff50e1ffdb248c730fb46a8a66088ee64bb
(cherry picked from qtbase/3c09f6bc9aee0c97427fe8da6efdc73b4ac473aa)
Reviewed-by: Liang Qi <liang.qi@digia.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Qt gets the information about locale digits from the <defaultNumberingSystem>
entries in the CLDR data xm files. Current CLDR data in Qt4.8 is 2.0, which
specifies native digits for most of Indic locales and arabic digits for slovak locale.
This doesn't concern Qt5 that uses the most recent version of CLDR (22.1) which specify Latin digits for the locales in question.
This patch updates the defaultNumberingSystem (to latin) and provides an autotest to compare the returned vs expected digits for the following langagues/regions:
- Gujarati, Hindi, Kannada, Konkani, Malayalam, Marathi, Nepali, Punjabi, Tamil, Telugu (to use latin digits instead of native ones)
- Slovak (to use latin digits instead of arabic ones)
Change-Id: I04364cdb80292f7705b1321377623298a01a9c5b
Reviewed-by: Vladimir Minenko <vminenko@rim.com>
Reviewed-by: Konstantin Ritt <ritt.ks@gmail.com>
|
|
|
|
|
|
|
|
| |
In commit Ib9ffcf0e26d3be36a0e158fd12a363b97177dcbf I forgot
about two local fields for QBasicAtomicPointer and QBasicAtomicInt.
Change-Id: I10637a4c85319ee16a8ea610840d31f331575267
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
|
|
|
|
|
|
|
|
| |
There aren't that many captured properties, decreasing the
increment will give better performance.
Change-Id: Iddb17c695eda5513f2d038123f4c886e559d8f6b
Reviewed-by: Christopher Adams <chris.adams@jollamobile.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
OpenSSL changed the layout of X509_STORE_CTX between 0.9 and 1.0
So we have to consider this struct as private implementation, and use
the access functions instead.
This bug would cause certificate verification problems if a different
version of openssl is loaded at runtime to the headers Qt was compiled
against.
Task-number: QTBUG-28343
Change-Id: I47fc24336f7d9c80f08f9c8ba6debc51a5591258
Reviewed-by: Richard J. Moore <rich@kde.org>
(cherry picked from commit eb2688c4c4f257d0a4d978ba4bf57d6347b15252)
|
|
|
|
|
|
|
|
|
| |
It's not sufficient to check that the source brush has an alpha, there
are composition modes that depend on the destination instead.
Task-number: QTBUG-28499
Change-Id: I4b267544c607b6a68c4c52a6784d526bec604d75
Reviewed-by: Gunnar Sletta <gunnar.sletta@digia.com>
|
|
|
|
|
|
|
| |
cherry-picked from qt5 stable 675676aba1e71410fa6cfca430ab9bd64ebdcdaa
Change-Id: I94e8ce037de7df50511be2471db842cca726ccef
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@digia.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Commit 94253f3472a13ac4eab3ccfbbba9d14165647f85 made QMAKE_MAC_XARCH
to be set to 'no' for MinGW environments. Due to this change building
for MinGW environments where SSSE3 support is available got broken.
While building the imageformats plugin qjpeg4 the following error occurs:
undefined reference to `_imp___Z32qt_convert_rgb888_to_rgb32_ssse3PjPKhi'
This commit makes building Qt on MinGW environments where SSSE3 support
is available working again
Change-Id: Ib2a4b3cc244af8bdde70d2d2240c70ff0e7a4cc2
Reviewed-by: Andy Shaw <andy.shaw@digia.com>
Reviewed-by: Daniel Molkentin <daniel@molkentin.de>
|
|
|
|
|
|
|
|
| |
Remove a reference to blackberry.json from the Blackberry bearer engine
plugin. This reference only makes sense under Qt5
Change-Id: I1592a17e38cb80b3fb138fc0538f3b4ea33739c8
Reviewed-by: Sean Harmer <sean.harmer@kdab.com>
|
|
|
|
|
|
|
|
|
|
|
|
| |
The BPS version provided by the current version of the Playbook OS does not
support orientation locking. This patch implements this manually.
This patch is not needed in Qt5, since the API has changed.
Change-Id: Ibba6fb3437cb392e3ba3d2b1997a178653226131
Reviewed-by: Nicolas Arnaud-Cormos <nicolas@kdab.com>
Reviewed-by: Bernd Weimer <bweimer@rim.com>
Reviewed-by: Thomas McGuire <thomas.mcguire@kdab.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
QStackedLayout doesn't have support for QLayout, only QWidget, so
the issue doesn't arise there.
(cherry-picked from qtbase/716d33d2a73ade42eb31be3e8ecbaeecdd5ddd21)
(only change: s/reset()/clear()/ in the tests)
Reported-by: Johannes Schaub
Task-number: QTBUG-27420
Change-Id: I7f3c4b1996e954428c00d4dda1095712efa91367
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@digia.com>
|
|
|
|
|
|
|
|
|
|
|
| |
The differences to VS 2010 project files are the
version number (surprise!) and the PlatformToolSet tag which
sets the version of the toolchain.
Change-Id: If26f08fad1a69d7e6cd28cc5e860ff964f19b264
Reviewed-by: Andy Shaw <andy.shaw@digia.com>
Reviewed-by: Joerg Bornemann <joerg.bornemann@digia.com>
(cherry picked from commit 461a01c1df88cdea0e8109948d125fc58a185cf8)
|
|
|
|
|
|
|
|
|
| |
Since graphics effects on widgets are not supported on Mac then the
code is #ifdef'd out so that it does not crash instead.
Task-number: QTBUG-24792
Change-Id: I4d81675b54816fa60cf66696b1a0a7b0206474e3
Reviewed-by: Morten Johan Sørvig <morten.sorvig@digia.com>
|
|
|
|
|
| |
Change-Id: I7e991d4b1f7f3b5a51170e6ff09e293c4285ea26
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
|
|
|
|
|
|
|
|
|
| |
Fix for QTBUG-27607. Now it handles the setMaxWindowRect for fullscreen
widgets. It is called when the resolution changes with QWS.
Change-Id: I0fdbb39c55f727b6507fbb14369943dd8b7c2d7e
Reviewed-by: Samuli Piippo <samuli.piippo@digia.com>
Reviewed-by: Paul Olav Tvete <paul.tvete@digia.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Fix probably-benign uninitialized memory reads
of QOpenGLPaintEnginePrivate's member has_brush
(if first call to updatePen() is before first call to updateBrush())
or has_pen (if the opposite).
This makes the following error reported by valgrind 3.7.0 disappear:
% valgrind --track-origins=yes --leak-check=no demos/composition/composition
(press the "Use OpenGL" button)
...
==30137== Conditional jump or move depends on uninitialised value(s)
==30137== at 0x4C72CBD: QOpenGLPaintEnginePrivate::updateUseEmulation() (qpaintengine_opengl.cpp:2135)
==30137== by 0x4C76F2A: QOpenGLPaintEngine::updatePen(QPen const&) (qpaintengine_opengl.cpp:2145)
==30137== by 0x4C7708D: QOpenGLPaintEngine::penChanged() (qpaintengine_opengl.cpp:5416)
==30137== by 0x4C7C960: QOpenGLPaintEngine::begin(QPaintDevice*) (qpaintengine_opengl.cpp:1421)
==30137== by 0x5376BE7: QPainter::begin(QPaintDevice*) (qpainter.cpp:1881)
==30137== by 0x41B06F: ArthurFrame::paintEvent(QPaintEvent*) (arthurwidgets.cpp:149)
==30137== by 0x5223315: QWidget::event(QEvent*) (qwidget.cpp:8532)
==30137== Uninitialised value was created by a heap allocation
==30137== at 0x4A082CF: operator new(unsigned long) (vg_replace_malloc.c:287)
==30137== by 0x4C7736C: QOpenGLPaintEngine::QOpenGLPaintEngine() (qpaintengine_opengl.cpp:1234)
==30137== by 0x4C5171C: QGLEngineThreadStorage<QOpenGLPaintEngine>::engine() (qgl_p.h:939)
==30137== by 0x4C48E32: qt_qgl_paint_engine() (qgl.cpp:5333)
==30137== by 0x4C48E4E: QGLWidget::paintEngine() const (qgl.cpp:5345)
==30137== by 0x5376351: QPainter::begin(QPaintDevice*) (qpainter.cpp:1785)
==30137== by 0x41B06F: ArthurFrame::paintEvent(QPaintEvent*) (arthurwidgets.cpp:149)
==30137== by 0x5223315: QWidget::event(QEvent*) (qwidget.cpp:8532)
==30137==
...
These members don't appear at all in the qt5 source,
so I'm assuming this isn't an issue there.
Therefore I'm submitting this fix directly, and only,
to the qt 4.8 source, rather than putting it in qt5 first.
(I wasn't able to get the composition demo program to exercise
the opengl drawing code at all on qt5, so I can't say for sure
whether qt5 is free of similar issues at this point.)
Change-Id: Ic69beddd6e2c50b827140cb8790b40b9336c8f4e
Reviewed-by: Jonas Rabbe <jonas.rabbe@gmail.com>
|
|
|
|
|
|
|
|
|
|
| |
that is, .file and .subdir "structs".
(cherry picked from commit 3460395ebbf1e9eb05419854c0d5de28b89b9b0a)
Change-Id: Ifdeedbbc0bccc9bc7a21707152750d7a2e839a5c
Reviewed-by: Liang Qi <liang.qi@digia.com>
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@digia.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
A focus scope has effective focus if one of its children is the focus item,
clearFocus() should remove effective focus from an item and its children
not just from the focus item.
Task-number: QTBUG-28328
Reviewed-by: Martin Jones <martin.jones@jollamobile.com>
(cherry picked from commit ab97a44b6ae5cc826460f3704669383afd398594)
Change-Id: Ib6f43b40091cb1bb1a0d5b8b9ac71e44f35b037b
Reviewed-by: Martin Jones <martin.jones@jollamobile.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* relro makes some sections read-only before giving control to the
program;
* now resolves all symbols when the program is started and allows for
the GOT being marked read-only to prevent it from being overwritten.
These settings have been used internally for device builds already.
Change-Id: I03576d4fd1585ad566865e2a7a0341b3e13c3b06
(cherry picked from commit c8c8e1990ce74590739bd52874e38e03e3d64825)
Reviewed-by: Sean Harmer <sean.harmer@kdab.com>
|
|
|
|
|
|
|
|
|
| |
akin to qt5/qtbase 31468563892a184d0a45cb0fdd7a91fa0da681e7, but without
feature auto detection due to Qt4 configure.exe nature.
Change-Id: I481e9ee01692dab9bd267a7b14722b8471719d1a
Reviewed-by: Thomas McGuire <thomas.mcguire@kdab.com>
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@digia.com>
|
|
|
|
|
|
|
|
|
|
|
|
| |
The buffer count check should take place right after buffers are created. For
some reason, the buffer count value inside libscreen may become incosistent during
the course of the program.
cherry-picked from qt5/qtbase 1b95d3331cb43fda36f92189ddce4e4c45bd47ea
Change-Id: Id542b3f8633ce118b0c125d47f3eda63ca31fe5e
Reviewed-by: Kevin Krammer <kevin.krammer@kdab.com>
Reviewed-by: Thomas McGuire <thomas.mcguire@kdab.com>
|
|
|
|
|
|
|
| |
This is a back port from: 5e59215ac649190a5b86f9af283d0cf41b88b3bf
Change-Id: I889f8b69a0c81a5e0ed71d28a65279fe6e238e2f
Reviewed-by: Peter Hartmann <phartmann@rim.com>
|
|
|
|
|
|
|
| |
Change-Id: Ibf1cfd47ef1e1694491d448f7dd73a338aad7043
Reviewed-by: Akseli Salovaara <akseli.salovaara@digia.com>
Reviewed-by: Sergio Ahumada <sergio.ahumada@digia.com>
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
|
|
|
|
|
|
|
|
|
|
|
|
| |
This assert has become invalid since commit 40fef403. After that commit, the
'origin' mentioned by the assert is not necessarily the first item of
cachedItemsUnderMouse, since it origin can be reassigned at line 5839.
cherry-picked from qt5 f4b80c1210e444ad7454abfb80ba166ce2e2a5f2
Change-Id: Iae264f5b301b4bf76b59c7262b06edb24c594ee5
Reviewed-by: Bernd Weimer <bweimer@rim.com>
Reviewed-by: Denis Dzyubenko <denis@ddenis.info>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
So that it can skip loading state when the underlying
QNetworkReply is already finished.
Parts-of-the-patch-by: Jeremy Nicholl
Task-number: QTBUG-27723
backport of e5783b79887299d094e6976630373a4899bd7074 from qtquick1
Change-Id: I8f5ee61a754ddec81ec70f82eb39e727534a6049
Reviewed-by: Alan Alpert <aalpert@rim.com>
|
|
|
|
|
|
|
|
|
|
|
| |
Under QNX, QProcess was not inheriting the parent's file descriptors. This
patch fills in the fd_map array, containing the file descriptors to be
inherited, which is passed to spawn(), accordingly.
cherry-picked from qt5 226f245c71df5673b5114615fbd9ad5c285b8d3a
Change-Id: Id2957c6278bc21c89234a84b364763b601ae08a1
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@digia.com>
|
|
|
|
|
|
|
|
|
| |
Ensure that dock widgets in Qt Creator can be resized.
Task-number: QTCREATORBUG-8284
Change-Id: If6aa55b2dde8fce53e1527db678f518b3912907b
Reviewed-by: Jarek Kobus <jaroslaw.kobus@digia.com>
(cherry picked from qttools/5e4bc7e3556cb5afcd82731277bdef60ae2cbe33)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
After bc3491c1b85ca36486c9472ecf7ba82f46699e8a, throwing an exception
from a slot will cause the application to crash (segfault). This patch
wraps the offending callFunction call in a try/catch block when
QT_NO_EXCEPTIONS is defined, allowing the appropriate cleanup to occur
and hence preventing the crash.
Task-number: QTBUG-26825
Task-number: QTBUG-27548
Change-Id: Ia3a02398b0308b2216ad17f8f643745bd013fd50
Reviewed-by: Olivier Goffart <ogoffart@woboq.com>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Andy Shaw <andy.shaw@digia.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
If QMainWindow::restoreState() then QWidget::setStylesheet() were
called before the QMainWindow is shown, the size of QDockWidget can not
be restored.
QWidget::setStylesheet() will generate QEvent::StyleChange event, which
will cause the function QDockAreaLayout::fitLayout() to be called
before the layout of MainWindow is activated. Although the state info
has been stored in a QMainWindowLayoutState variable by
QMainWindow::restoreState(), but QMainWindowLayout::setGeometry() still
isn't called at present. So QDockAreaLayout::fitLayout() will force the
size of dockwidgets and centralwidget to be calculated using the wrong
geometry, which will break the state restored by
QMainWindow::restoreState().
This is a side effect of 692e9103ebb85b90e79377206d5d03b704d43d42.
This patch is a backport of d2f65aa470fe30849a01380e4a50e8a4ebbce07e.
Task-number: QTBUG-15080
Change-Id: I5b912e80e99b1f8b7a6f2a9a3e541fac3455c3db
Reviewed-by: Olivier Goffart <ogoffart@woboq.com>
Reviewed-by: Andy Shaw <andy.shaw@digia.com>
Reviewed-by: Debao Zhang <hello@debao.me>
|
|
|
|
|
| |
Change-Id: Ia79f5dd4a55d5260032781eab2f74737d65729f2
Reviewed-by: Jerome Pasion <jerome.pasion@digia.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
When the floating-point behavior in MSVC is set to "precise" (default),
assigning nan numbers to a float causes the bit pattern to be altered
(only affects 32bit builds). We should therefore not assign the swapped
value back to a float and use it.
Backported from Qt 5 (3f936e9094f3a6e4d76791c1eff7ae92f91b61ae)
Task-number: QTBUG-25950
Change-Id: I39725557ab89c978eb1c59a4c79df8a27ed70ecf
Reviewed-by: Jędrzej Nowacki <jedrzej.nowacki@digia.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Should become more stable due to two changes:
1. When QAccessibleTextWidget::characterRect() calculates the height,
make sure the leading is not included.
2. Accept a small error (1) on height and width.
It there is an error on the height or width, dump diagnostics
This makes it pass with all Latin fonts on my system
(roughly ~150 fonts), where it didn't before.
Change-Id: I964207bad1f3ceb0103129501bdc857aff5885d2
Reviewed-by: Rafael Roquetto <rafael.roquetto@kdab.com>
Reviewed-by: Frederik Gladhorn <frederik.gladhorn@digia.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
libscreen will always provide the same physical screen dimensions
regardless of the initial orientation. Width and hight will now be
swapped appropriately (to make sure that the DPI value is correct).
The Qt5 implementation is different (the qnx plugin takes the
initial orientation into account).
Change-Id: I983d1db7ae334b442597b62dde64d41bb991690c
Reviewed-by: Sean Harmer <sean.harmer@kdab.com>
Reviewed-by: Kevin Krammer <kevin.krammer@kdab.com>
|
|
|
|
|
|
|
|
|
|
|
|
| |
This change ensures that BPS events are delivered to the proper
channel even if the application code would change the channel.
Backport from qtbase: 131b86347384a97a7044942acd723c4ef5669636
Change-Id: Ib2bdd3d973bc51f5ce85eb6ef9adcb3e57093a54
Reviewed-by: Fabian Bumberger <fbumberger@rim.com>
Reviewed-by: Kevin Krammer <kevin.krammer@kdab.com>
Reviewed-by: Thomas McGuire <thomas.mcguire@kdab.com>
|
|
|
|
|
|
|
|
|
| |
This defines Q_OS_BLACKBERRY_TABLET. This define is not necessary on Qt5.
Change-Id: I2e2553f3b33d896df9e219c12e738251eed8a231
Reviewed-by: Robin Burchell <robin+qt@viroteck.net>
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@digia.com>
Reviewed-by: Nicolas Arnaud-Cormos <nicolas@kdab.com>
|
|
|
|
|
|
|
|
| |
Task-number: QTBUG-27992
Change-Id: I4ed5661cfe0d0a10143224b73fbad6fac4927a37
Reviewed-by: Joerg Bornemann <joerg.bornemann@digia.com>
(cherry picked from qtbase/5223e7eb27dc15f8428aeb66dd8c57098dedce58)
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
... and not only for the C compiler. Apparently the logic has changed
between Qt 4 and Qt 5.
This commit is not needed for Qt 5, because there we set the CXXFLAGS
after parsing the BlackBerry mkspec file.
Change-Id: I225d636586d19fa0e2b0d5a644071202774a385c
Reviewed-by: Kevin Krammer <kevin.krammer@kdab.com>
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@digia.com>
Reviewed-by: Sean Harmer <sean.harmer@kdab.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
QGLTextureCache has a read write locker to protect texture lookups.
However, even calling QCache::object() might modify the cache, causing
race conditions, since it modifies the priority order of the objects in
the cache. Therefore, we need to protect it with a QWriteLocker instead
of a QReadLocker.
Task-number: QTBUG-22560
Change-Id: I46fffc624ace27e25fb100f865e1df0a19b9093d
Reviewed-by: Gabriel de Dietrich <gabriel.dedietrich@digia.com>
(cherry picked from commit 6633dbbac49f4ce8a89b67090fdffaa8c8126eb5)
Reviewed-by: Gunnar Sletta <gunnar.sletta@digia.com>
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
|
|
|
|
|
|
|
|
| |
Task-number: QTBUG-21863
Change-Id: I70b661fa5fcaba1293a80c971c506798826f5b23
Reviewed-by: Joerg Bornemann <joerg.bornemann@digia.com>
(cherry picked from qtbase/1022c3ce9d445fd482a62309b9a6c52495c04a92)
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
When sifting through a file to find Q_OBJECT and friends, qmake
was locking it for reading and writing (_SH_DENYRW).
This breaks parallel builds where multiple qmake instances might want
to scan the same file.
Changed the sharing constant to _SH_DENYNO, which doesn't lock the
file at all. This is consistent with the _sopen_s call in findDeps.
Change-Id: I9c0a06db7f580f411e79d9d96dd36a6d705679aa
(cherry picked from commit 9cd7eb29f3b216bdac215fb7633b4b51d534d220)
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@digia.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
On X11 if you have multiple tablet input devices, like a Wacom Cintiq and
a Wacom Intuos tablet connected at the same time, one of the tablets will
generate incorrect high resolution coordinates.
The reason seems to be that when translating the XEvent in
qapplication_X11.cpp the code searches for the first tablet device it
finds that supports the type of the current event. Since both tablets
will support this event type, we wind up always finding only one of the
tablets. Specifically, the first tablet that was found when enumerating
tablets during qt_init. Then all calls to QTabletDeviceData::scaleCoord
are made on the same device, so the size of one of the devices is never
used. That device will wind up with incorrect high resolution
coordinates.
When not running under IRIX the XEvent will have a deviceid which we can
use to find the correct device to use for scaling. This change
implements that fix.
Change-Id: I667a52a729273a9b12880592d0a4cd7ce7f16106
Reviewed-by: Gunnar Sletta <gunnar.sletta@digia.com>
Reviewed-by: Konstantin Ritt <ritt.ks@gmail.com>
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
|
|
|
|
|
|
|
|
|
| |
The test is now passing on all CI platforms.
Change-Id: I205d8d6951c64f5eca3b2f7ea4d6c362ab046fda
Reviewed-by: Caroline Chao <caroline.chao@digia.com>
Reviewed-by: Simo Fält <simo.falt@digia.com>
Reviewed-by: Janne Anttila <janne.anttila@digia.com>
|
|
|
|
|
|
|
|
|
| |
tst_QHostInfo::cache()
tst_QHostInfo::abortHostLookup()
Task-number: QTBUG-28079
Change-Id: I6f524cdee86bc77583852327e18c6d4a79595323
Reviewed-by: Caroline Chao <caroline.chao@digia.com>
|