As discussed in QTQAINFRA-6146, there were two potential approaches to
enabling backtraces in Qt Quick tests:
- Either via a thorough refactoring of quicktest.cpp so that each
testcase is a Slot in a programmatically-created QObject-derived class,
which will be executed via QTest::qExec() similar to the documented
QTest way. This way the pre-existing code in QTest::qRun() will setup
the fatal signal handler.
- Or as a quick fix, modify quick_test_main_with_setup() to setup a
FatalSignalHandler class and invoke prepareStackTrace() like it's
already done in QTest::qRun(). This would require exporting these
symbols in the private header.
This patch enables the implementation of the latter, as it has a
fairly light footprint, is easily revertable (should we need to),
and allows us to immediately gain insight into crashes in CI.
Task-number: QTQAINFRA-6146
Change-Id: Iedffc968acb3e570214df34884ab0afcb6b30850
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
In preparation for reusing it in Qt Quick, which currently doesn't
print backtraces upon crashes.
Task-number: QTQAINFRA-6146
Change-Id: Ib0384f514b348a398f53529ff3bcc7d4ac2daba7
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Ahmad Samir <a.samirh78@gmail.com>
Because we can't support it on 32 bit operating systems.
Task-number: QTBUG-69558
Change-Id: I406ecccdf039d7d4f4c24268c92c91e367655cba
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
f1bb9cfbf6 added this value, but it was
only when a test in qtdeclarative tried to use it that it was
discovered that it couldn't be set on 32 bit operating systems (armv7,
AKA imx7) due to overflow as a result of the bit shifting that is done.
Fix it by using an old, deprecated value. If any old codebase using
that older flag tries to build against a newer Qt with this change, it
shouldn't affect it, as setting the flag does nothing in Widgets, and
native menus didn't exist in earlier versions.
Task-number: QTBUG-69558
Change-Id: I520154d02e9ccf007ebd73807685212a19fbee1b
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Draw the glyph directly in the paint() override, and use that
from the scaledPixmap implementation. This avoids a pixmap
creation just for drawing an icon with an existing painter.
Pick-to: 6.7
Change-Id: Iece0083a3a9f3625d843bc6e9d836baf9b5d84f8
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Doris Verria <doris.verria@qt.io>
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
Instead of implementing scaledPixmap() to rasterize the
vector image we get from App/UIKit, and calling that from
paint, implement paint() to draw the vector image directly
through the painter, and use that in scaledPixmap.
Pick-to: 6.7
Change-Id: I2c62826f29406543bc8d8c7fa71199e91586d83b
Reviewed-by: Amr Elsayed <amr.elsayed@qt.io>
Reviewed-by: Doris Verria <doris.verria@qt.io>
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
Although the output of these functions is premultiplied ARGB, we know
the alpha is always going to be full (1.0) because the source is CMYK 32
bit (without alpha). We can therefore avoid calling qPremultiply.
This work has been kindly sponsored by the QGIS project
(https://qgis.org/).
Change-Id: I129b601f5c93a1c444ab06c3325f946d2bcc6efc
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
As part of eacd58d4e7, a mechanism was
added to prevent posting redundant UpdateRequest events to the top
level widget, managed by QWidgetRepaintManager. The mechanism relied
on a boolean that was set when posting an update request event, and
reset when processing the event for the top level widget, as part
of QWidgetRepaintManager::sync().
However, for paint-on-screen widgets, we don't post an update request
to the top level, we post it to the paint-on-screen widget directly.
And when processing that event, we don't paint the widget though the
normal QWidgetRepaintManager machinery, but instead call the paint
event via QWidgetPrivate::paintOnScreen().
As a result, an update() on a paint-on-screen widget would result
in never receiving updates for non-paint-on-screen widgets, as
we assumed there was no reason to send further update requests.
We could fix this by clearing the updateRequestSent flag as part
of the paintOnScreen() code path, but that's incorrect as the flag
represents whether the top level QWidgetRepaintManager needs an
update request event or not, and would lead to missed updates
to normal widgets until the paint-on-screen widget finishes its
update.
Instead, we only set updateRequestSent if we're posting update
requests for non-paint-on-screen widgets, which in practice
means the top level widget.
The paint on screen logic in QWidgetRepaintManager::markDirty
still takes care of rate-limiting the update requests to the
paint-on-screen widget, by comparing the dirty area of the
widget.
There is definite room for improvement here, especially in the
direction of handling update requests via QWindow::requestUpdate
instead of manually posted events, but for now this will have to
do.
Fixes: QTBUG-80167
Pick-to: 6.7 6.6 6.5 6.2 5.15
Change-Id: Ib5685de7ca2fd7cd7883a25bb7bc0255ea242d30
Reviewed-by: Richard Moe Gustavsen <richard.gustavsen@qt.io>
It was introduced in December 2018, which is apparently too recent for
some Linux distributions. So remove the dependency that was there only
to give us MS_RDONLY and revert to the old statfs.h / sys/vfs.h flag.
We still don't include <sys/vfs.h> because it is absent on some old
Android versions.
Amends ea6abe583f.
Pick-to: 6.7
Fixes: QTBUG-123932
Change-Id: If1bf59ecbe014b569ba1fffd17c29cc448d16358
Reviewed-by: Ahmad Samir <a.samirh78@gmail.com>
We can't fix the underlying reported problem, but we can warn that the
user has made a mistake.
[ChangeLog][QtCore][QObject] The class template parameter passed to
QObject::findChild() or findChildren() is now required to have the
Q_OBJECT macro. Forgetting to add it could result in finding children of
the nearest ancestor class that has the macro.
Pick-to: 6.7
Fixes: QTBUG-105023
Change-Id: I5f663c2f9f4149af84fefffd17c008f027241b56
Reviewed-by: Giuseppe D'Angelo <giuseppe.dangelo@kdab.com>
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
We don't use that in the XCB library or plugin anywhere.
Pick-to: 6.7
Change-Id: I5f663c2f9f4149af84fefffd17be8c7f3bd7bd14
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
On systems without the RTLD_NODELETE flag, simply don't call dlclose()
and leak the handle. Amends ef5ab6e006.
Pick-to: 6.7
Change-Id: I01ec3c774d9943adb903fffd17b76673562daa8a
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Ivan Solovev <ivan.solovev@qt.io>
Remove unused code.
This work has been kindly sponsored by the QGIS project
(https://qgis.org/).
Change-Id: I3db7705b3963d7a7669f8c9b284d3d35a2a7da92
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
JPEG part 6 defines CMYK support. This commit adds such support to the
JPEG plugin.
A *very* interesting discovery is the fact that Photoshop inverts the
meaning of the CMYK color channels when saving into JPEG: 0 means "full
ink", and 255 means "no ink". Most other image viewers/editors follow
the same interpretation, I imagine for compatibility.
But others, like Adobe Reader, don't (???) -- a PDF expects a DCT
encoding with 0 meaning "no ink". I am adding a SubType to the image I/O
handler to let the user choose what they want, defaulting to Photoshop
behavior.
Also, turns out that Qt was already loading CMYK files and converting
them to RGB. I don't think we should do automatic, lossy conversions (we
were not taking into account an eventual colorspace...), so I'm changing
that loading to yield a CMYK QImage.
Finally: save the colorspace, even if it's a CMYK image.
QColorSpace doesn't support anything but RGB matrix-based colorspaces.
Yet, it can load an arbitrary ICC profile, and will store it even if
it's unable to use it. We can use this fact to preserve the colorspace
embedded in CMYK images, or let users set an arbitrary ICC profile on
them through Qt APIs, and then saving the result in JPEG.
[ChangeLog][QtGui][JPEG] Added support for loading and saving of JPEG
files in 8-bit CMYK format. When loading a CMYK JPEG file, Qt used to
convert it automatically to a RGB image; now instead it's kept as-is.
This work has been kindly sponsored by the QGIS project
(https://qgis.org/).
Change-Id: Ibdbfa16aa35814f5dba28c2df89577175162b731
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
Input events have a timestamp. When dispatching an event through QPA, a
platform plugin can either provide it, or QPA will use an internal
QElapsedTimer to provide a timestamp.
Windows input messages do come with a timestamp already, so we can use
that instead of the QPA.
The two methods are not equivalent.
For instance: for various reasons, Qt does not honor Windows' "double
clicked" message, but uses the delta between two mouse events to
establish if the second click is actually a double click.
Now suppose that the user double clicks on a widget. On the first click,
the application does something that freezes it for a bit (e.g. some
heavy repainting or whatever). Does the second click register as a
double click or not?
* If we're using Qt's own timer, the answer is NO; the event is pulled
from the WM queue after the freeze, given a timestamp far away from
the last click, and so it will be deemed another single click
* If we use the OS' timestamps, then the second click will be seen as
"close" to the first click, and correctly registered as second click.
This reasoning can be extended to many other QPA events, but looks like
the APIs for some are missing (e.g. enter events), so I'm not tackling
them here.
This commit reverts ade96ff644.
Task-number: QTBUG-109833
Task-number: QTBUG-122226
Change-Id: I5f3fba029b44c618ef622bacdc4bcc3928e04867
Reviewed-by: Oliver Wolff <oliver.wolff@qt.io>
Reviewed-by: Shawn Rutledge <shawn.rutledge@qt.io>
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Yuhang Zhao <yuhangzhao@deepin.org>
Add setDepthResolveTexture(). Should work similarly to the color
attachments' resolveTexture, but for depth or depth-stencil.
However, this is another fragmented feature.
- D3D11/12:
Not supported. AFAICS multisample resolve (ResolveSubresource) is just
not supported for depth or depth-stencil formats.
- Vulkan:
Not supported with Vulkan 1.0.
Supported with Vulkan 1.1 and the two extensions.
(VK_KHR_depth_stencil_resolve which in turn requires
VK_KHR_create_renderpass2 since the 1.0 structs are not extensible, so
now need to use VkRenderPassCreateInfo2 and all the '2' structs)
In Vulkan 1.2 the above are in core, without the KHR suffix, but we
cannot just use that because our main target, the Quest 3 (Android) is
Vulkan 1.1. So 1.2 and up is ignored for now and we always look for
the 1.1 KHR extensions.
The depth resolve filter is forced for SAMPLE_0. AVG seems to be
supported on desktop (NVIDIA) at least, but that's not guaranteed, so
would need physical device support checks. On the Quest 3 it does not
seem to be supported. And in any case, other APIs such as Metal do not
have an AVG filter mode at all, so just use SAMPLE_0 always.
- OpenGL (not ES):
Should work, both when the multisample data is a renderbuffer and a
texture. Relies on glBlitFramebuffer with filter NEAREST. What it does
internally, with regards to the depth/stencil resolve mode, is not under
our control.
- OpenGL ES:
Should work when the multisample buffer is a texture. But it will not
work when a multisample renderbuffer (setDepthStencilBuffer, not
setDepthTexture) is used because the GLES-only multisample extensions
(GL_EXT_multisampled_render_to_texture,
GL_OVR_multiview_multisampled_render_to_texture, which we prefer over
the explicit resolve-based approach) work with textures only.
- Metal:
Should work.
Task-number: QTBUG-122292
Change-Id: Ifa7ca5e1be78227bd6bd7546dde3a62f7fdbc95e
Reviewed-by: Andy Nichols <andy.nichols@qt.io>
"Just in case" -- as the comment says,
QInternalMessageLogContext's constructor is usually inlined, but
we can't be sure if it has been.
Change-Id: I4da2f0875d9fd9f7bd6d79447b4761fda329c7fd
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
This commit adds a CMYK 8bpp format to QImage. The idea is to enable the
transport of CMYK images inside Qt, for instance to be loaded/saved from
files or painted on CMYK capable paint devices (e.g. PDF).
Also, rasterization support *from* a CMYK image is added (on top of a
RGB surface), as well as CMYK image scaling/conversion.
Conversion and rasterization between CMYK and RGB isn't particularly
optimized nor it honors any colorspaces yet. The overall idea is to
match 1:1 the existing behavior of CMYK QColor (which get naively
changed to RGB; there isn't colorspace support in QPainter yet).
There are no plans to add rasterization *towards* CMYK.
Image save/load in native CMYK formats will be added in future commits.
This work has been kindly sponsored by the QGIS project
(https://qgis.org/).
[ChangeLog][QtGui] Support for 8-bit CMYK images has been added.
Change-Id: I4b024cd4c15119c669b6ddd450418a9e425587f8
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
Change the invalid static_cast<std::int16_t>(i)+1
to the correct static_cast<std::int16_t>(i+1).
Change-Id: I5d3e17d29deb2a70fa0d7d7838531a3dc80b4e45
Reviewed-by: Giuseppe D'Angelo <giuseppe.dangelo@kdab.com>
Calling data() for parameters of the type bigint
failed in combination with stored procedures with
output parameters. Cast the result to qlonglong to
fix it.
Pick-to: 6.7 6.6
Change-Id: I84ef04ed26821b92ef7c5bcdf12b778e91450e0b
Reviewed-by: Giuseppe D'Angelo <giuseppe.dangelo@kdab.com>
Inline namespaces serve the purpose for which the original
-qtnamespace option was added and allow for macro simplification (no
need for any using directives). This makes it possible to use
namespaced builds of Qt also for Qt for Python and similar use cases
which have issues with the additional namespace.
[ChangeLog][QtCore] It is now possible to use an inline namespace for
-qtnamespace (option -qtinlinenamespace).
Task-number: PYSIDE-2590
Change-Id: Ia0cecf041321933a2e02d1fd8ae0e9cda699cd1e
Reviewed-by: Alexey Edelev <alexey.edelev@qt.io>
We're not guaranteed to get into any of the code paths that call
removePopupMonitor() before the app goes away. In a plugin-scenario,
this may cause crashes when our monitor then gets a callback and
we try to access QGuiApplicationPrivate::instance().
Pick-to: 6.7 6.6 6.5 6.2
Fixes: QTBUG-123959
Change-Id: I287b91ff261a8aab74adbbad8c63a042daf944d5
Reviewed-by: Doris Verria <doris.verria@qt.io>
Like the other times it's used in this file
This is causing compilation errors in projects that use -Werror=undef
Fixes: QTBUG-123937
Pick-to: 6.7
Change-Id: I0cdd2910755dc9079890011dd8dbc27a6e64793e
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Apparently it has trouble with multiple attributes between the
class keyword and the class name, looking only for the EXPORT
macros.
As a quick band-aid we can add the #pragma to tell syncqt
to generate the header explicitly.
Pick-to: 6.7
Fixes: QTBUG-123875
Change-Id: If155a5b667b9e71d43dfac04ad19caee0ff23793
Reviewed-by: Alexey Edelev <alexey.edelev@qt.io>
Reviewed-by: Juha Vuolle <juha.vuolle@qt.io>
QMessageBox contains an icon that can be set using icon type. Before
this commit, android.R.attr.alertDialogIcon was used regardless of the
expected icon type. Only if setting failed, the icon from
android.R.drawable was set (depending on the expected icon type).
Previously, usage of android.R.attr.alertDialogIcon was the way to
consider Theme in choosing icon view. Since
31a0d99fa5 commit, getDrawable(id, theme)
is used with second parameter: Resources.Theme.
Because of that we can start to use only icons from android.R.drawable
and remove usage of android.R.attr.alertDialogIcon
Fixes: QTBUG-123334
Pick-to: 6.7 6.6 6.5
Change-Id: I6cfdaf30aea5d132e38ba5d78054089b51cf5f13
Reviewed-by: Assam Boudjelthia <assam.boudjelthia@qt.io>
Instead of deep inside the processing of the message pattern tokens.
This reduces the number of uninlineable functions in release builds to
2. Experimentation shows that even in debug mode our overhead to
backtrace() is just 2, despite the debugger's backtrace showing 3 frames
-- the debugger is using QtCore's debug symbols and does know about the
inlined function.
As a nice side-effect, we capture the backtrace only once per message,
not once per %{backtrace} token in the QT_MESSAGE_PATTERN.
Change-Id: I01ec3c774d9943adb903fffd17b7d6de09167620
Reviewed-by: Giuseppe D'Angelo <giuseppe.dangelo@kdab.com>
QMessageLogContext is a primitive type that may be extended in the
future with more fields (it has been at version 2 since commit
6d166c8822, though that did not extend the
struct's size). This introduces a QInternalMessageLogContext which is
used in before all our calls to qt_message_output().
Currently there's no difference and no way to tell that the internal
version is used.
Change-Id: I01ec3c774d9943adb903fffd17b7d5abc0052207
Reviewed-by: Giuseppe D'Angelo <giuseppe.dangelo@kdab.com>
8cd7a3d472 remembered the current screen
in QMenuPrivate::popupScreen. This QPointer member is not reset, after
QMenu:exec(), which makes a re-used menu remember the wrong screen, if
its next usage happens on another screen.
Reset the member variable at the end of QMenuPrivate::exec().
This amends 8cd7a3d472.
Fixes: QTBUG-118434
Pick-to: 6.7 6.6 6.5
Change-Id: I7457ca72166346f01cf71b2706ebc20ecd71173c
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
Add the note about the way the str size is determined when using
QByteArray::operator=(const char*).
Pick-to: 6.7 6.6 6.5
Change-Id: I39b2d0fc2967832622fbf0c11b3ff6c7ff99b8f2
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Leena Miettinen <riitta-leena.miettinen@qt.io>
The QVkPipelineCacheDataHeader::reserved field wasn't initializaed by
the code, but then memcpy()ed with the struct into the result
QByteArray. At best, this contains random data, at worst, it leaks
information.
Initialize it to zero.
Found by Coverity.
Amends df0e98d408.
Pick-to: 6.7 6.6 6.5 6.2
Coverity-Id: 444147
Change-Id: I398c9a1e99483f2f9887d768319b20ecc11e2c86
Reviewed-by: Giuseppe D'Angelo <giuseppe.dangelo@kdab.com>
What was done:
* Removed headers in src/gui/accessible/windows/apisupport: as of
v13.1.0, MinGW supports most of the definitions in these headers.
Including uiautomation.h should be enough.
* Removed the QWindowsUiaWrapper class: it's not meant to be extended
or itself instantiated, is an "ultra-thin" layer (it even preserves
the "all-caps" Win types of function args), and is in effect only a
MinGW-bound "kludge". Instead of this class, use the UI Automation
API directly, with the assumption that it's available and fully
functional, as specified in the MS docs. Any gaps between this
assumption and what is delivered by MinGW are bridged with specific
(and explicit) temporary "kludges".
* Implemented said specific "kludges" in qwindowsuiautomation. For
Windows builds, the header just includes uiautomation.h, and the
.cpp is empty. For MinGW, the header contains definitions still
missing from uiautomation.h, and the .cpp implements functions
of the UI Automation core library through imports from the
uiautomationcore DLL.
* Windows plugins (and tst_qaccessibility): use the UI Automation API
definitions directly, instead of the "ultra-thin" wrapper.
* Windows plugin builds: use uiautomationcore library, if found.
What's intended:
* Unburden Gui of the Windows UI Automation COM interfaces and other
definitions that are copied in the uia*.h headers.
* Make the Windows plugins independent of MinGW shortcomings.
* Remove the QWindowsUiaWrapper class that essentially only hides these
shortcomings and the "kludge" code needed to overcome them.
* As MinGW adds further support to the UI Automation API over time,
make it noticeable which workarounds are no longer needed. The
current approach of hiding "kludges" in a wrapper class will also
hide the fact that they're no longer needed, if/when that time comes.
Change-Id: I0070636817d5de81d0b106e9179e2d0442362e2a
Reviewed-by: Wladimir Leuschner <wladimir.leuschner@qt.io>
Reviewed-by: Oliver Wolff <oliver.wolff@qt.io>
Commit 87362f3f58 added
qt_linux_find_symbol_sys(). Given the authors and reviewers, I'm
guessing that was used somewhere in qtdeclarative. Indeed, its use was
removed in qtdeclarative/67191c2b3213259c6eaf045154e9370faa085868
("Remove qqmlmemoryprofiler*").
qt_mac_resolve_sys() was never used in the public repository, as far as
I can tell.
Change-Id: I6818d78a57394e37857bfffd17bbf1fae8688b1a
Reviewed-by: Ahmad Samir <a.samirh78@gmail.com>
These options are declared with TYPE enum and a MAPPING that's supposed
to control the feature 'system-zlib' or 'system-sqlite'. Since only
inputs of type boolean control features now, we need to somehow declare
that this non-boolean input controls a feature.
We do this by adding the keyword CONTROLS_FEATURE to
qt_commandline_option. For example,
qt_commandline_option(zlib
CONTROLS_FEATURE
TYPE enum
NAME system-zlib
MAPPING system yes qt no
)
declares
- commandline option "zlib" sets the input "system-zlib",
because of the "NAME system-zlib" argument
- accepted input values are "system" and "qt", because
we have "TYPE enum" and the odd values of MAPPING
- those values are translated to yes/no, because of the
even values of MAPPING
- CONTROLS_FEATURE forces the translated input's type
to boolean, and with that it will set the corresponding
feature 'system-zlib'
Luckily, only qtbase has command line options with MAPPING declared.
Change-Id: I82d06cec43ece3b002c8f5dd414c68dc730909af
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
QColorVector is not forced to 128-bit alignment.
Change-Id: Ifacc171296ddddda742d49745372b47585e40802
Reviewed-by: Giuseppe D'Angelo <giuseppe.dangelo@kdab.com>
To indicate color spaces that can not be used as a target,
but only as a source.
Change-Id: Iae79e3533599c112872d171a2f45178029be89dc
Reviewed-by: Giuseppe D'Angelo <giuseppe.dangelo@kdab.com>
Amends aadf1d447c because size_t on 32-bit
platforms is 32-bit.
qhash.cpp:331:15: error: left shift count >= width of type [-Werror=shift-count-overflow]
Change-Id: I5f663c2f9f4149af84fefffd17c0567b17f832ce
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
And ensure all combination of CBOR types are tested.
Amends e5ebb9022a
Change-Id: I5f663c2f9f4149af84fefffd17c02d352cd41f3f
Reviewed-by: Tatiana Borisova <tatiana.borisova@qt.io>
The name says it's a warning, it's pretty unexpected that it should
lead to a critical message.
One case where this is a problem is QTBUG-48488
where merely printing with "Microsoft Print To PDF" and canceling
the file dialog leads to "QWin32PrintEngine::begin: StartDoc failed"
as a *critical* message. Some Windows applications have a message
handler that shows a msgbox in case of a critical message,
and getting such a msgbox after canceling the file dialog seems very
wrong.
Task-number: QTBUG-48488
Pick-to: 6.7
Change-Id: I1c842340dd2faf2be6e64e0522f9e2b33547d3cf
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Like the QPolygon code. This fixes a mistake in failing to clear the
list before de-marshalling in operator>>.
Updated most of the QDataStream tests to have data in the objects
they're streaming into, to ensure that the stream overwrites everything.
Fixes: QTBUG-122684
Task-number: QTBUG-122704
Pick-to: 5.15 6.5 6.6 6.7
Change-Id: I01ec3c774d9943adb903fffd17b6920c72f5042b
Reviewed-by: Giuseppe D'Angelo <giuseppe.dangelo@kdab.com>
New comparison macros are used for following classes:
- QJsonArray::iterator
- QJsonArray::const_iterator
Replace public operators operator==(), operator!=(), operator!<(), etc
of classes to friend methods comparesEqual(), compareThreeWay();
Use *_helper methods to have an access to protected members of
QCborValueConstRef class from friend functions.
Task-number: QTBUG-120300
Change-Id: I9b41b619107ce69d8b6dab4938232fab841aab51
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
QDbusTrayIcon has a convoluted workaround/hack that consists in putting
an icon file in /tmp/ and then using that path as the icon. Opening the
icon file may fail, so handle it.
Change-Id: I5d1c681e2fe3cfb23e93fd20f6758d4c83fe1578
Reviewed-by: Dmitry Shachnev <mitya57@gmail.com>
Reviewed-by: Shawn Rutledge <shawn.rutledge@qt.io>
The QTextStream(FILE*) constructor may fail to open the filehandle
through QFile (e.g. if the open mode is incompatible). However
QTextStream has no error report mechanism for this and still reports
status Ok. This is consistent for instance with the QIODevice*
constructor, where even passing a closed device sets the status to Ok.
Add a comment and discard the return of open().
Change-Id: I430b96fd26e0ebca15a4d9ee640b09895bdd0b03
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
We had no examples of this type, and didn't catch the table was parsed as
-0.5 to 0.5 instead of 0 to 1.
Change-Id: I904937a50deaeecfc89e271bf918eedc521bc8a2
Reviewed-by: Giuseppe D'Angelo <giuseppe.dangelo@kdab.com>
Coverity complains that QToolBarAreaLayout's
addToolBarBreak(QInternal::DockPosition) could access
QToolBarAreaLayout::docks out of bounds if passed
QInternal::DockCount.
That is correct, but a valid pos seems to be a precondition for this
function, judging from its sister functions, e.g.
addToolBar(DockPosition, .) or insertItem(DockPosition, .), which also
don't validate `pos`. All in-module callers of addToolBarBreak() only
pass valid positions, and use validateToolBarArea() to ensure that. So
it seems that Coverity doesn't grok the pass-by-in/out -parameter used
by that function. That, or it doesn't track back far enough.
Before attempting more drastic measures, first try rewriting the
function to return-by-value instead, and see what Coverity has to say
afterwards.
As a drive-by, make validateToolBarArea() constexpr.
Pick-to: 6.7 6.6 6.5 6.2 5.15
Coverity-Id: 444141
Coverity-Id: 444135
Change-Id: I5fcc664c3cea608429036cad75c37f5c38059733
Reviewed-by: Richard Moe Gustavsen <richard.gustavsen@qt.io>
A literal backslash needs to be doubled so that the parser doesn't treat
it as escaping the following character when the markdown is read back.
In ca4774131b we tried to limit it to
backslashes that were not already escaped. In case someone really needs
a longer series of backslashes, it's more correct to escape them all;
but this comes with the risk that if they do not get un-escaped by the
markdown parser in some scenario, repeated round-trip saving and loading
could multiply them excessively. So we also add a lot of tests to try
to verify that this is safe.
Task-number: QTBUG-96051
Fixes: QTBUG-122083
Pick-to: 6.7
Change-Id: I64f610d24e99f67ebdc30d5ab5c6cf3985aec5ec
Reviewed-by: Axel Spoerl <axel.spoerl@qt.io>
Coverity complained about a use of the moved-from (in the first line
of the loop) `info` object in subsequent lines.
This specific instance is harmless, because the field being accesssed
is a scalar and the move SMFs are the default ones, so the field isn't
actually changed when the struct is moved from.
Still, to silence Coverity and to guide other attentive readers of the
code, take a copy of the field before moving from the struct, and use
the copy's value after the move.
Amends ddc39eb3a4.
Amends 3e330a79ec.
Pick-to: 6.7
Coverity-Id: 444199
Change-Id: I26ea8669f27124fb2567b16d803d47ab439f1e41
Reviewed-by: Ahmad Samir <a.samirh78@gmail.com>
Replace public operators operator==(), operator!=() of
QJsonObject to friend methods comparesEqual().
Use QT_CORE_REMOVED_SINCE and removed_api.cpp to get rid of current
comparison methods and replace them with a friend.
Add friend method comparesEqual(QJsonObject, QJsonValue)
to the QJsonObject class, to support comparison between QJsonObject
and QJsonValue elements, see test-case valueEquals().
Task-number: QTBUG-120300
Change-Id: Ibab0b4b39966205447e31c41e94e7e1a4e31e553
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Ivan Solovev <ivan.solovev@qt.io>
The QProcess default wait time of 30s may be too short
in e.g. CI environments where processes may be blocked
for a longer time waiting for CPU or IO.
Task-number: QTBUG-117598
Change-Id: I27dbe83ddbe811ae4ff28767de67cb0ceaab267e
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
A QMenuBar may span multiple screens. Its menus will
then open on the screen which the bottom middle of
the action rect is at, and will snap to that screen's
geometry if needed.
However it can happen that the bottomLeft() of the action
rect is on a different screen from the selected popup
screen, in which case mapping this point to global
coordinates can give incorrect coordinates if the screens
have different scale factors.
The x value will be corrected by the screen snapping
if needed. Use the y from the screen test point, which
will be correct for the selected screen.
Task-number: QTBUG-73231
Change-Id: If9a39f6b832a64f2f701868f2be0d3a6468fe553
Reviewed-by: Axel Spoerl <axel.spoerl@qt.io>
Coverity correctly complains that we're copying the QDebug object when
calling print*Container() when we could have moved it into the helper.
So do move it.
Pick-to: 6.7 6.6 6.5
Coverity-Id: 406803
Coverity-Id: 407406
Coverity-Id: 408523
Coverity-Id: 408562
Coverity-Id: 418431
Coverity-Id: 424788
Coverity-Id: 425106
Coverity-Id: 426537
Coverity-Id: 427163
Coverity-Id: 428925
Coverity-Id: 444463
Change-Id: Ic80247f315a09fffe9363577dff1d1c781859304
Reviewed-by: Giuseppe D'Angelo <giuseppe.dangelo@kdab.com>
Either the code in the try {} block
- doesn't throw, then the return within the try {} block exits the
function, or it
- throws fileystem_error, then we rethrow in the catch, or it
- throws any other exception, then we leave the function by exception
In no case can control reach the trailing 'return {}'.
Found by Coverity.
Amends 7aecb189d5.
Pick-to: 6.7 6.6 6.5
Coverity-Id: 444466
Change-Id: I1c1bf752453076724c2fa9367ea5309e741d84ac
Reviewed-by: Alexey Edelev <alexey.edelev@qt.io>
Fixes failure to find qHash functions when building Squish. Here, QMap
is used with custom types that do not have any associated qHash
functions. It should therefore not require a qHash function to be
usable with QMap.
There is no indication in the build output where the instantiation came from. There's nothing in the output about the function it should have
called, likely because no qHash functions exists.
Amends: e1f45ad818
Fixes: QTBUG-123310
Change-Id: Idb12fb6ffab06ce242c43f2c42ea7a105e8fa0f4
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Marc Mutz <marc.mutz@qt.io>
Replace public operators operator==(), operator!=() of
QJsonValue/QJsonValueConstRef/QJsonValueRef classes to friend
methods comparesEqual().
Use QT_CORE_REMOVED_SINCE and removed_api.cpp to get rid of current
comparison methods and replace them with a friend.
Delete non-exported public operators operator==(), operator!=()
for QJsonValueConstRef/QJsonValueRef classes.
Add comparison check to auto-test.
Task-number: QTBUG-120300
Change-Id: I01434af4ce5a7589733db4a9b14f54ad42e852ed
Reviewed-by: Ivan Solovev <ivan.solovev@qt.io>
So we have a single central location for all generated files.
[ChangeLog][Build System] Generated resource files (and supporting
files) will now be placed into the .qt/rcc subdirectory of a project
build dir. The location is an implementation detail that might still
change in the future, so it should not be relied upon.
Pick-to: 6.7
Change-Id: Id21df22cac832b618e98c25e0e134f4cf70ed9bd
Reviewed-by: Alexey Edelev <alexey.edelev@qt.io>
content URI obtain permissions when selected by the Android file
provider, and the Android APIs will treat them as different URIs if
a slash is added and thus such paths will be unaccessible the same way.
Amends d89c32140a.
Change-Id: I8107c7d482dee75f4637e13400b8844b3d3ff804
Reviewed-by: Ahmad Samir <a.samirh78@gmail.com>
In case a custom TLS plugin is provided, the order can be unclear
(next after OpenSSL is either Schannel on Windows, or Secure Transport
on Darwin, then a custom plugin, if any, and the last one is 'cert-only').
Pick-to: 6.7 6.6 6.5 6.2
Fixes: QTBUG-123092
Change-Id: I02bcc1fa5448f64846d561a72b2522af3286c66c
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
QtInstallPaths needs to be loaded at early stages
Change-Id: Ie275ad2a8855b7555b110c35814ebadafe1817c6
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
First, realize that we don't need the isObjectValid() call, because
that's done by verify(QObject*,QMetaMethod) later.
That leaves said fromSignal() and verify(QObject*, QMetaMethod) calls,
which we can just inline into the (QObject*, Func) ctor, thus making
said constructor SCARY, having extracted all template-independent code
into other functions/ctors.
Task-number: QTBUG-123544
Change-Id: I6b8afc541f75936045e2d28cfde51a34f98a1fdd
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
This reverts commit f4bac3ca17.
It broke builds with Xcode 15
Change-Id: Iee232658ede3dfb09d65f3f6a95410c069941421
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
If a signal was invalid, the code still called messageSignature(),
which returns a null QByteArray, and passed its constData() to
qWarning("%s"). That probably worked in our implementation, because it
falls back to QString::asprintf(), but it raised eyebrows, so avoid
calling messageSignature() on invalid QMetaMethod.
This changes the warning output, so adjust the test.
Task-number: QTBUG-123544
Change-Id: I41bc6650de091f61354ff91ee45659668f0e0223
Reviewed-by: Giuseppe D'Angelo <giuseppe.dangelo@kdab.com>
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Following the verify() Extract Method refactorings, this function has
only one caller left, so it doesn't pull its weight anymore,
esp. considering that it'll be exported soon.
Task-number: QTBUG-123544
Change-Id: I1690b4b6e5a0e0c56fcc9c34544fca3b34e2f9a6
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
... where it belongs.
Amends e68edd6a07.
Task-number: QTBUG-123544
Change-Id: Ic0e5128555465485b579607a61925cefa5f4716d
Reviewed-by: Giuseppe D'Angelo <giuseppe.dangelo@kdab.com>
And try to not call Q_ASSERT() to avoid crash.
Fixes: QTBUG-123554
Pick-to: 6.7 6.6 6.5
Change-Id: I9443c5f0ab4ca7a858df9b328f517b48ab8f122d
Reviewed-by: Axel Spoerl <axel.spoerl@qt.io>
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
Replace public operators operator==(), operator!=() of
QJsonDocument to friend methods comparesEqual().
Use QT_CORE_REMOVED_SINCE and removed_api.cpp to get rid of current
comparison methods and replace them with a friend.
Task-number: QTBUG-120300
Change-Id: I7b61765c34406b7a9fb7dd8b1fc554c87af6a3f3
Reviewed-by: Ivan Solovev <ivan.solovev@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
There is a limit in WebGL that OpenGL context sharing is not supported.
There is a proposal from 2013
(https://www.khronos.org/webgl/wiki/SharedResouces), but it seems that
it never was implemented.
There is an additional limit in that there is exactly one WebGL context
for each canvas (i.e. window).
A part of the problem here is that the identifier for an OpenGL context
is essentially a surface-thread-context triplet. And the thread part
might be more difficult to handle in a javascript setting.
Regardless of why, we need to have an opinion about what the
consequences are, and what we support to what extent.
As such this commit:
1) Adds a comment on the QOpenGLContext describing the limitations
2) Adds a qWarning() on the first activation of a shared context.
The second item is not complete. We will have problems with multiple
individual contexts also, It is just not possible to warn for these
cases. The second item covers, maybe, the most common case.
Change-Id: I51550a6acb0a7f6f6fa5e9e2c3da080a1d2b498f
Reviewed-by: Morten Johan Sørvig <morten.sorvig@qt.io>
Types marked with [[clang::trivial_abi]] are considered to be trivially
relocatable for Clang. This is ABI compatible, since in Qt 6 we can
change the value of QTypeInfo::IsRelocatable "after the fact" -- it
simply means that code that doesn't get recompiled is pessimized.
Change-Id: I32e52bfb212c7919b2ebcf7832ede4404358330f
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Add per-ctor verify() functions that take care of the warnings for
invalid inputs, and return a struct { QObject*; QMetaMethod; } to feed
into a delegatee constructor.
This solves the problem that the varying parts between the ctors are
at the beginning, not the end, using another level of indirection, and
will eventually allow to make the `args` and `sig` members const, and
therefore remove the need to protect either with the mutex.
This patch tries to keep the diff minimal. I'm planning to de-inline
most of the class in a future commit, so it doesn't matter that
private and public sections are interleaved, that will be cleaned up
when the code is moved to a .cpp file later.
Task-number: QTBUG-123544
Change-Id: Idc628c927736880a8fd580089ed5177361c94ed9
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
src/gui/painting/qpaintengine_raster.cpp:3811:42: error: ‘void* calloc(size_t, size_t)’ sizes specified with ‘sizeof’ in the earlier argument and not in the later argument [-Werror=calloc-transposed-args]
3811 | m_clipLines = (ClipLine *)calloc(sizeof(ClipLine), clipSpanHeight);
| ^~~~~~~~~~~~~~~~
src/gui/painting/qpaintengine_raster.cpp:3811:42: note: earlier argument should specify number of elements, later size of each element
Pick-to: 6.7 6.6 6.5 6.2 5.15
Change-Id: I41ec3dd5c439e5cd51dd917741125ce50659500e
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Ahmad Samir <a.samirh78@gmail.com>
Reviewed-by: Giuseppe D'Angelo <giuseppe.dangelo@kdab.com>
On macOS, entries that do not match the filter are shown in the
directory listing.
Do not accept the dialog when these entries are double-clicked
(activated).
Pick-to: 6.7 6.6 6.5
Fixes: QTBUG-120768
Change-Id: If8ff6c56f1d21861b4e30051c212c9497042ed0f
Reviewed-by: Axel Spoerl <axel.spoerl@qt.io>
Currently if qt6_generate_deploy_app_script is called on an executable
target that does not have the MACOSX_BUNDLE property set, it will
error out with a message about not supporting non-bundle apps.
This error is shown even if NO_UNSUPPORTED_PLATFORM_ERROR is passed
as an option which looks like an oversight.
Change the code not to show the error if NO_UNSUPPORTED_PLATFORM_ERROR
is passed. This means user projects can call
qt6_generate_deploy_app_script for non-bundle apps without having to
wrap the code in a if(NOT APPLE) check, and deployment will simply not
run for that target on macOS.
[ChangeLog][Build System] The qt6_generate_deploy_app_script
NO_UNSUPPORTED_PLATFORM_ERROR option will now have an effect when
calling the API on non-bundle macOS executable targets.
Pick-to: 6.7
Change-Id: I932d6bfa2d3c7e2aaf8be967fea1f682eacf0112
Reviewed-by: Alexey Edelev <alexey.edelev@qt.io>
Replace public operators operator==(), operator!=() of
QJsonArray to friend methods comparesEqual().
Use QT_CORE_REMOVED_SINCE and removed_api.cpp to get rid of current
comparison methods and replace them with a friend.
Add friend method comparesEqual(QJsonArray, QJsonValue)
to the QJsonArray class, to support comparison between QJsonArray
and QJsonValue elements, see test-case fromToVariantConversions()
Task-number: QTBUG-120300
Change-Id: I8440ca0761bede8551ff792bfa7f22e47b56fa79
Reviewed-by: Ivan Solovev <ivan.solovev@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Qt's openssl plugin is dependent on openssl libraries. Users should
make a conscious decision about deployment of the plugin and the
corresponing libraries so deployment of that functionality is now
opt in.
[ChangeLog][Tools][Windeployqt] Deployment of openssl plugins is now
optional. For proper deployment of openssl related functionality pass
in --openssl-root to make sure that openssl libraries are also deployed.
Fixes: QTBUG-117394
Change-Id: Iad43c7666b9af491f7695783b4e23a811256515b
Reviewed-by: Miguel Costa <miguel.costa@qt.io>
So that they are moved.
Pick-to: 6.7 6.6 6.5
Change-Id: I3d1767944200962be90d3eb8695c92a766c4c744
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
This is probably a remnant from when QAbstractFileEngine was public API
since it's been changed to private API, just use QFile::FileTime.
Pick-to: 6.7
Change-Id: I60d3d4ff811f95434b81d5ca115f5d43cfff8b15
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
This makes the ownership of the returned pointer clearer. It also
matches reality, some call sites were already storing the pointer in a
unique_ptr.
Also shorten the function name to "createLegacyEngine", you have to read
its docs anyway to figure out what it does.
Drive-by changes: less magic numbers; use sliced(); return nullptr
instead of `0`.
Change-Id: I637759b4160b28b15adf5f6548de336887338dab
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Assam Boudjelthia <assam.boudjelthia@qt.io>
The Qt::WA_ShowWithoutActivating flag was not respected
Added test in the part of the code that calls
requestActivateWindow
Added selenium focus test
Fixes: QTBUG-122776
Change-Id: I1a248ed4352f86376d615a4cb7022e7ea095d4e7
Reviewed-by: Piotr Wierciński <piotr.wiercinski@qt.io>
Ideally, this wouldn't be a named function, but a delegatee
constructor, but the current structure of the three constructors
doesn't, yet, lend itself to extracting a delegatee constructor (the
tail is copied, not the head). To get there, we need more work (coming
up in follow-up commits).
Task-number: QTBUG-123544
Change-Id: I46dd030e314d67c2ab624279d669db76e58bc569
Reviewed-by: David Faure <david.faure@kdab.com>
Use the QT_CMAKE_EXPORT_NAMESPACE prefixed INSTALL_LIBS variable when
resolving the cmake path of the ABI-specific cmake directory.
Pick-to: 6.7 6.6 6.5
Change-Id: Ibc4a1b152135d840de104c15a183b13fca0739ea
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Use QVLA instead of QQueue to avoid allocations
and use QSpan to emulate queue behavior to avoid N^2
Change-Id: I333bc1af98a596fc041765996867a3d6449fcfde
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Using the rvalue overload of append() is more efficient, since we skip
the alias check and the appendee's atomic ref-count ping-pong inside
lvalue append().
Pick-to: 6.7
Task-number: QTBUG-123544
Change-Id: Ia76fdf28cba13d524fbbe894658a86a45a1ebe79
Reviewed-by: David Faure <david.faure@kdab.com>
Clazy complains that QMetaMethod should be passed by value, so do
that.
Change-Id: I42afda5af6910eefc8783b1feac00fd11e1f017e
Reviewed-by: David Faure <david.faure@kdab.com>
Due to https://gitlab.kitware.com/cmake/cmake/-/issues/25770, the
presence of the Q_OBJECT macro here triggers automoc rule generation,
followed by moc's "No relevant classes found. No output generated."
warning.
Comment the macro out until the issue is fixed, leaving a comment to
avoid a helpful soul commenting the macro back in again.
Amends 32a1151245.
Task-number: QTBUG-123229
Pick-to: 6.7
Change-Id: I5387128f8fcd2793f0d4b9ff73a551999fea84b0
Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
Reviewed-by: Ahmad Samir <a.samirh78@gmail.com>
The static keyword prevents the linker from reusing the executable
code of one TU's instantiation of this function in other TUs.
There's no reason for this restriction, so remove it.
Amends 292cb12e02.
Pick-to: 6.7
Change-Id: Ide60ce0bf4b7118295fad98a25bd438fc1da2039
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
While a signal with more than 2Gi arguments is only a theoretical
possibility, still use the correct index variable for the indexed loop
over this QList<int>.
Pick-to: 6.7 6.6 6.5
Change-Id: I2ed33238c2cd9d2d1c39cd29c988a2adfd821897
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
A recent change in that area triggered Clang-SA to rescan the code and
complain about this pre-existing bug: If receiver == nullptr, then the
`previous` member was never initialized.
Fix by null'ing it using NSDMI. This is a trivial type, the compiler
will be able to avoid the redundant write.
Amends ab92b9e40025dcf08c14232de762a268201a78b4(!).
Pick-to: 6.7 6.6 6.5 6.2 5.15
Change-Id: Ideed71f0f36d5f896fb6a4614f233757c1371ee3
Reviewed-by: David Faure <david.faure@kdab.com>
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
And remove hasNext/next() methods. This remodels QAFEI to be like
QFileSystemIterator. This better fits the logic in the newly added
QDirListing class (which uses STL-style iterators).
QFSFileEngineIterator:
Initialize the internal nativeIterator in the constructor; also replace
the advance() private method with an override for the advance() method
inherited from the base class.
QResourceFileEngineIterator:
Override currentFileInfo(), with a QResouces the QFileInfo is created
on demand if/when this method is called.
This is the backend/private API, and QDirListing is the public API that
can be used in a ranged-for to iterate over directory entries.
Change-Id: I93eb7bdd64823ac01eea2dcaaa6bcc8ad868b2c4
Reviewed-by: Assam Boudjelthia <assam.boudjelthia@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
New comparison macros are used for following classes:
-QXmlStreamAttribute
-QXmlStreamNamespaceDeclaration
-QXmlStreamNotationDeclaration
-QXmlStreamEntityDeclaration
Replace public operators operator==(), operator!=() of
classes to friend methods comparesEqual();
Use QT_CORE_REMOVED_SINCE to get rid of current comparison methods
and replace them with a friend.
Add checkStreamNotationDeclarations()/checkStreamEntityDeclarations()
test-cases to test change.
Task-number: QTBUG-120300
Change-Id: I0b5642b2e23cc21ede7bc4888f0a9bddd6c08d07
Reviewed-by: Ivan Solovev <ivan.solovev@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Drop the \c command when in front of \l commands as that doesn't work.
Pick-to: 6.7 6.6
Change-Id: I0aa092461807e068e9c2368f5d6f04e77b56c910
Reviewed-by: Andreas Eliasson <andreas.eliasson@qt.io>
Reviewed-by: James DeLisle <james.delisle@qt.io>
Extract duplicated logic to own method, and call where previously used.
Also, make `QDomNodeListPrivate::[list|timestamp]` mutable, to allow
`maybeCreateList()` and `createList()` to be `const` methods, and avoid
`const_cast` in `QDomNodeListPrivate::length()`.
Task-number: QTBUG-115076
Change-Id: I4f3a27315da28082a12cc4f5653567039b4cb376
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Reviewed-by: Marc Mutz <marc.mutz@qt.io>
Replace public operators operator==(), operator!=(), operator<() of
QCborMap to friend methods comparesEqual() / compareThreeWay().
Use QT_CORE_REMOVED_SINCE to get rid of current comparison methods
and replace them with a friend.
Delete #if 0 && __has_include(<compare>) blocks,
since they are not required anymore.
Add friend methods comparesEqual(QCborMap, QCborValue) and
compareThreeWay(QCborMap, QCborValue) to the QCborMap
class, to support comparison between QCborMap
and QCborValue elements, see test-case mapSelfAssign() ->
QT_TEST_EQUALITY_OPS(it.key(), QCborMap({{0, v}}), true);
Task-number: QTBUG-120300
Change-Id: I9e33df255d16484efd3124cf0632db859408fb5d
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Ivan Solovev <ivan.solovev@qt.io>
There is no concept of a screen on visionOS, and hence no concept of
screen-relative system UI that we should keep top level windows away
from.
As UIWindow and UIScreen on visionOS report the exact same bounds, we
can't use the same code path we use for iOS/iPadOS to detect whether
to apply the safe area insets, so skip the logic entirely.
We still report the safe area insets on a QWindow level, so applications
can avoid rendering content close to UI elements such as the window
move and window close buttons.
Task-number: QTBUG-121781
Change-Id: I20ad02390564972a3715c27f351689b79ad579d8
Reviewed-by: Amr Elsayed <amr.elsayed@qt.io>
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
The API has been deprecated for a long time, and doesn't have the same
semantics as we want for modern iOS apps that may run in windowed mode
on iPad, macOS, or visionOS.
Instead we base the availableGeometry exclusively on the safeAreaInsets
of the screen's UIWindow. But we only do so if the UIWindow bounds match
the bounds of the UIScreen in each dimension. This means that iOS apps
that take up the entire screen, or run in Split View on iPad, will get
a smaller availableGeometry, but apps that run in Slide Over on iPad,
or windowed in Stage Manager on iPad, will get an availableGeometry that
matches the entire screen.
It's then up to the QWindow to take the safeAreaMargins into account
when rendering, to stay out of any UI elements that the OS may add.
We do this for QWidgets with layouts automatically, and the plan is
to do this for Qt Quick Controls as well. As the current situation
is quite broken anyways, we take the first step now, and then follow
up with the Quick changes later on.
Task-number: QTBUG-121781
Change-Id: I7728e117969474654fcccdb19a6c954bb0eea3f9
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
In 4c21f72837 we did this for
QImageReader. So, to support the same for QMovie, we just need
to remove the custom (and very inefficient) scaling for frames.
This should be safe since the underlying QImageReader must ensure
that the returned image matches the scaled size.
[ChangeLog][QtGui][QMovie] Allow only one dimension (width
or height) to be set for the scaled size. In this case, the other
will be calculated automatically based on the original movie size
and maintaining the aspect ratio.
Fixes: QTBUG-115039
Change-Id: I50979a75970c79647dbb8c8b4b121266ba033a63
Reviewed-by: Eirik Aavitsland <eirik.aavitsland@qt.io>
12203e94f5 exposes the issue in xcb
qpa plugin.
We should not recreate xcb window via the call of focusWindow->winId().
Fixes: QTBUG-123032
Pick-to: 6.7 6.6 6.5 6.2 5.15
Change-Id: I6da4f3e64a9d7a92a2aab714591986c5d128fbd4
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
In modern iPad apps, windows can be moved between screens, just like
regular macOS apps, but we don't reflect this as changes to the QWindow's
screen, meaning the screen() is not reliable input when computing
a window's full screen or maximized geometry.
In addition, when running iOS apps on macOS as "Designed for iPad",
the system will scale the UI down by 77%, to better match metrics
(fonts, sizes) to the macOS environment. Classes like UIView and
UIWindow are oblivious to this scaling, and will think they are
larger than what they really are on the screen. However, UIScreen,
for some reason, reflects the actual screen size, instead of taking
part in the "inflated" view of the world.
As a result, even if screen() would reflect the correct screen the
window is on, we can't use the screen geometry for clamping the
window geometry, as the two have separate coordinate systems.
We could scale up the QScreen geometry accordingly to work around
this, but we don't know if the UIScreen behavior is a bug or not,
so instead we skip the screen() as input and use the UIWindow
bounds directly.
Task-number: QTBUG-121781
Fixes: QTBUG-106709
Pick-to: 6.7 6.6 6.5
Change-Id: Ie734fc477b37a7e39e6fb16d7738d3f69731a975
Reviewed-by: Amr Elsayed <amr.elsayed@qt.io>
Reviewed-by: Doris Verria <doris.verria@qt.io>
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
It says:
/home/qt/qt6dev-src/qtbase/src/corelib/tools/qcontainertools_impl.h: In function ‘auto QtPrivate::sequential_erase_with_copy(Container&, const T&) [with Container = QList<void (*)()>; T = void (*)()]’:
/home/qt/qt6dev-src/qtbase/src/corelib/tools/qcontainertools_impl.h:383:14: error: ‘D.282000’ is used uninitialized [-Werror=uninitialized]
383 | const T &tCopy = CopyProxy(t);
| ^~~~~
cc1plus: all warnings being treated as errors
We can avoid storing the value into a const ref to silence this. Storing
a non-const pointer into a const reference is quite confusing anyway.
Fixes: QTBUG-123486
Pick-to: 6.5 6.6 6.7
Change-Id: I77fcd871724ce7f81b9567603dc5b4cb31f121c5
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
The atomic pointer "connections" is always populated under
mutex in QObjectPrivate::ensureConnectionData() but isn't necessarily
read under mutex protection (e.g. in maybeSignalConnected()).
This caused a data race, fixed by using storeRelease and loadAcquired.
Task-number: QTBUG-100336
Pick-to: 6.7 6.6 6.5
Change-Id: Ifd605e65122248eb08f49e036fdda6e6564226bc
Reviewed-by: Marc Mutz <marc.mutz@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
The static inline was only a workaround for maintaining BC within
minor releases while backporting the fix. Since we don't promise BC for
QtTest between minor releases, we can make the mutex a proper member
variable for Qt 6.8.
Amends c837cd7593.
Change-Id: I0d6353bdd6a11daa4f927139abf9a867d8c9f95f
Reviewed-by: Marc Mutz <marc.mutz@qt.io>
The warning for the new-style signal constructor was copied from the
old-style signal constructor, but not adjusted to its new home. The
signal pointer passed here is not the signal "name", but a signal
"pointer" (-to-member-function, but no need to go into that much
detail).
Amends 6fc7d76e73, but not picking to
very strict LTS branches, just in case someone has a
QTest::ignoreMsg() installed on it.
Pick-to: 6.7 6.6 6.5
Change-Id: Ia1f6b7001f38202ac72f9945c4a822d81562cdbf
Reviewed-by: David Faure <david.faure@kdab.com>
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
If only one dimension (width or height) of the scaled size is set
by an user, let's try to calculate the second one, based on the
image original size and maintaining the aspect ratio.
[ChangeLog][QtGui][QImageReader] Allow only one dimension (width
or height) to be set for the scaled size. In this case, the other
will be calculated automatically based on the original image size
and maintaining the aspect ratio.
Task-number: QTBUG-115039
Change-Id: If1b13b1ead3cf788915c08bdb3ba8becd8bf8042
Reviewed-by: Eirik Aavitsland <eirik.aavitsland@qt.io>
We would like to add deployment api calls to all our examples.
Doing that in the CI, where we build all examples, would mean running
the deployment tools for each example. This takes time and space.
Add the option to skip running deployment code by setting the
QT_INTERNAL_SKIP_DEPLOYMENT variable.
This can be set per-directory or at the root directory.
In that case we will generate an install script that does nothing.
With this option, we can ensure we run deployment for a few examples,
but skip it for the rest.
Also skip running the deployment api for non-prefix builds, because it
only partially works. In such a case, show a warning. The warning
can be hidden by setting the QT_INTERNAL_HIDE_NO_OP_DEPLOYMENT_WARNING
variable. This might be revisited in the future.
The newly introduced helper functions are also meant to be used in
qtdeclarative.
qtbase only uses one of them, because currently it does not contain
POST_BUILD deployment logic.
Pick-to: 6.7
Task-number: QTBUG-90820
Task-number: QTBUG-96232
Task-number: QTBUG-102057
Change-Id: If5a4102137e9dfc4a8bfde7b26d511ed1700340e
Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
Since d08fa86e63 we add extra deploy
files with target file information.
The file paths contain a $<CONFIG> genex to handle multi-config
builds. These were not evaluated in the DeploySupport file content.
Add a TARGET_GENEX_EVAL genex around the TARGET_PROPERTY genex, to
ensure they do get evaluated.
This avoids an error like:
include could not find requested file:
C:/Users/qt/work/qt/qtbase_build/.qt/QtDeployTargets-$.cmake
Amends d08fa86e63
Pick-to: 6.7
Task-number: QTBUG-117948
Task-number: QTBUG-118153
Change-Id: Ia3898c31d80b82bdb25fe733fc73b45c6d689ed0
Reviewed-by: Alexey Edelev <alexey.edelev@qt.io>
Replace public operators operator==(), operator!=(), operator<() of
QCborArray to friend methods comparesEqual() / compareThreeWay().
Use QT_CORE_REMOVED_SINCE to get rid of current comparison methods
and replace them with a friend.
Delete #if 0 && __has_include(<compare>) blocks,
since they are not required anymore.
Add friend methods comparesEqual(QCborArray, QCborValueConstRef)
and compareThreeWay(QCborArray, QCborValueConstRef) to QCborArray
to support comparison between QCborArray and
QCborValueRef/QCborValueConstRef, see test-case mapMutation().
Add QT_TEST_EQUALITY_OPS/QT_TEST_ALL_COMPARISON_OPS tests for QCborArray
test-cases.
Task-number: QTBUG-120300
Change-Id: Ifad1a04c61363618e8bba73cf7c87757552d722a
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Simplifies maintenance of code paths that rely on Metal.
Change-Id: I1d1f705fffc14dbafde346eeb555b43be6d5be54
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
The Find pasteboard has been unavailable on iOS since iOS 10.
Change-Id: I05d5ae49b3a733144ac5c35970ff4a3c62923650
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
We have implemented the same functionality recently for POST, in this
patch implement it for PUT.
Task-number: QTBUG-108309
Change-Id: I34c41538054fec836d0d1d1dbb44fabab9bc0e9a
Reviewed-by: Juha Vuolle <juha.vuolle@qt.io>
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Actually this has already worked if a nullptr was casted as a
QIODevice*. Add an overload with a nullptr_t type, that does
this behind the scenes.
Fixes: QTBUG-108309
Change-Id: I2d4b17ae94cf4de2c42257d471ef901c8994fee5
Reviewed-by: Juha Vuolle <juha.vuolle@qt.io>
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Remove minimum window size restriction. User should be able to change
minimum window size if needed.
Set default size to 160x160 to match other platforms.
Change-Id: Ic199fc34982021ba38d631476fbb1c51370b2e8e
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
You're not supposed to mess with the refcount directly. That's what the
addRef() and deref() methods are for.
Change-Id: I5cae946cef7ac72dd99b247ade95590d142c269e
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
bb6d68703b67e042e2a7254c2ca6a004a1441cc5 fixed warnings in the
Universal style by using a faster alternative. It's possible that users
will run into these warnings too, and they should be provided with
information to make a more informed choice about which approach they
can use.
Fixes: QTBUG-123360
Pick-to: 6.5 6.6 6.7
Change-Id: I4170e9ade40c4b54dbc2bd73d124b2ade4d8c939
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@qt.io>
The API is available on all iOS versions we support nowadays.
Pick-to: 6.7 6.6 6.5
Change-Id: Ia58c5ad1649e7e6b22f9c56a809e2455586a8e5a
Reviewed-by: Amr Elsayed <amr.elsayed@qt.io>
Reviewed-by: Doris Verria <doris.verria@qt.io>
This amends a34e81ab8b .
The previous range comparison doesn't work since XKB_KEY_KP_9 is 0xffb9
while XKB_KEY_KP_Equal is 0xffbd. Change to an explicit switch.
Pick-to: 6.7 6.6 6.5 6.2 5.15
Change-Id: I3a340bac61fb074eef505ef9b06300a6468877f1
Reviewed-by: Giuseppe D'Angelo <giuseppe.dangelo@kdab.com>
Reviewed-by: Liang Qi <liang.qi@qt.io>
Detected by TSAN in tst_QThread::terminateAndPrematureDestruction()
but better have a dedicated unittest, with values emitted by the signal
and recorded in the spy.
Pick-to: 6.7 6.6 6.5
Change-Id: I141d47b0ea0b63188f8a4f9d056f72df3457bda5
Reviewed-by: Marc Mutz <marc.mutz@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Replace public operators operator==() and operator!=() of
QCborValueConstRef and QCborValueRef classes
to friend methods comparesEqual().
Replace public operator<() of QCborValueConstRef and QCborValueRef
classes to friend methods compareThreeWay() respectively.
Use QT_CORE_REMOVED_SINCE to get rid of current comparison methods
and replace them with a friend.
Delete #if 0 && __has_include(<compare>) blocks,
since they are not required anymore.
Add comparison() test-case for QCborValueConstRef/QCborValueRef
testing.
Add QCborValue::operator==()/QCborValue::operator!=()/
QCborValue::operator<() and QCborValueRef::operator==()/
QCborValueRef::operator!=()/QCborValueRef::operator<() operators
to the removed_api file.
Task-number: QTBUG-120300
Change-Id: I2e8e4e32b7b5b49da321364cc12986e9c64b5f37
Reviewed-by: Ivan Solovev <ivan.solovev@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
For QMap on a std::pair, QMap::iterator{}.operator->() would result in a
type that has "first" and "second" members: std::pair itself. But it
would be wrong to marshall and demarshall the first and second elements
thereof as the key and value, so give preference to the .key() and
.value() selection, which would store the std::pair as the type.
Fixes: QTBUG-123401
Pick-to: 6.5 6.6 6.7
Change-Id: I6818d78a57394e37857bfffd17bd69bb692dd03b
Reviewed-by: Alexey Edelev <alexey.edelev@qt.io>
Qt supports some complex foreground brushes which we cannot
express using normal CSS, so we introduce a Qt-specific property
for this. We already had some support for background gradients
in widget style sheets, but this expands support to foreground
brushes of text when converting a QTextDocument from and to HTML.
It also adds an optional "coordinatemode" attribute to the
gradient functions so that this can be faithfully restored from HTML.
Task-number: QTBUG-123357
Change-Id: I3d6dd828f68272995c8525bec5a7b421fdbed670
Reviewed-by: Eirik Aavitsland <eirik.aavitsland@qt.io>
CSS does not have text outline properties, instead different
browsers have custom properties for this. That currently means
that you can have a QTextDocument where you applied a stroke to
text and textEdit.setHtml(textEdit.toHtml()) will remove it.
Since a primary goal of the HTML support in QTextDocument is that
it can be used to save and faithfully restore its contents, we
implement qt specific properties for stroke.
Task-number: QTBUG-123357
Change-Id: Id9cf63abfabe2109ffb6fd74f9cb013304763ccb
Reviewed-by: Eirik Aavitsland <eirik.aavitsland@qt.io>
It shows up in our headersclean mode.
Fixes: QTBUG-123374
Pick-to: 6.6 6.7
Change-Id: I6818d78a57394e37857bfffd17bcf9e350dc493c
Reviewed-by: Simo Fält <simo.falt@qt.io>
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
Send connection error type PROTOCOL_ERROR when receivng CONTINUATION
frames without any data received from previous HEADERS or PUSH_PROMISE
frames, instead of assert.
Task-number: QTBUG-122375
Pick-to: 6.7
Change-Id: Ib14e4610692dc4832b1f3e99dca497d9baf3d9d3
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
The SDK is relevant for all Apple systems, including macOS, iOS, tvOS,
watchOS, and visionOS.
We still pick up -DQT_UIKIT_SDK for iOS for compatibility.
[ChangeLog][CMake] The -sdk configure argument now maps
to the QT_APPLE_SDK CMake variable. QT_UIKIT_SDK is still
supported for iOS builds for compatibility.
Change-Id: I983a2f23c2414eb73cd35bb83738088defb45cbd
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
Do not remove the existing documentation for now, because the new
commands do not contain any links to explain what the types of ordering
actually mean.
Fixes: QTBUG-108805
Change-Id: I1297c435d700d02968f0d7d3a922af78a61d7e45
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
We were looking at all active streams, but that also includes promised
streams.
By the RFC the limitation that our peer specifies only applies to the
number of streams _we_ create, not the total amount of active streams.
More importantly, for the qhttp2protocolhandler it could mean that we
could end up having a few promised streams push the active stream count
over the limit, which could lead us to start more streams than intended
(then bounded by the number of queued requests).
The worst case in this scenario is that a **non-compliant** server
doesn't track how many connections we open and the user has queued
a ton of requests, so we open a ton of streams.
But on the upside: server-push is not widely used.
Pick-to: 6.7
Change-Id: I2a533472eb9127fd176bb99e9db0518f05c3fffe
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Reviewed-by: Alexey Edelev <alexey.edelev@qt.io>
The settings frame with the max streams might be received late
or be revised later, so we cannot assert something on the
relation with the max streams allowed.
Amends 22c99cf498
Pick-to: 6.7
Change-Id: I973dfcf91541becc8c3d6363f9065bb1b9183062
Reviewed-by: Øystein Heskestad <oystein.heskestad@qt.io>
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
This has to be done from all places where we deref it. Otherwise we leak
memory.
Pick-to: 6.7 6.6 6.5
Fixes: QTBUG-116086
Change-Id: I57307ac746205578a920d2bb1e159b66ebd9b2cc
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
Check that all string-like types implement compareThreeWay() as a
(hidden) friend function.
This test revealed some problems: because QByteArrayView is implicitly
constructible from QLatin1StringView, the qCompareThreeWay() call
sometimes picked the compareThreeWay() overloads with QByteArrayView
instead of picking the "reversed" overloads that use QLatin1StringView.
This was leading to wrong results, because QByteArrayView is
interpreted as utf-8 data, not latin-1.
Explicitly add the missing compareThreeWay() and comparesEqual()
overloads.
Note that relational operators were never affected by this problem,
because in C++17 mode we explicitly generate the reversed versions,
and in C++20 mode the compiler does that for us.
Task-number: QTBUG-117661
Change-Id: Ia96af59c60ebf2fae6cf2a49231d6b6f401aceaa
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Commit fb50ab7006 replaced relational
operators in QByteArrayView with comparison helper macros.
That commit had to implement the helper methods as static members
instead of hidden friends, because otherwise it broke the
construction of QByteArrayView from QLatin1StringView.
The reason for that is the conditional noexcept on the relational
operators. The noexcept(noexcept(comparesEqual)) check inside the
class declaration expands before the class is complete, and as a
result discards the QByteArrayView constructor, which is restricted
by the QtPrivate::IsContainerCompatibleWithQByteArrayView trait.
Back then I could not find a proper solution, so just decided to
declare the helper functions for QByteArrayView as private static
members.
Such approach has two drawbacks:
- all new helper functions for QBAV vs other type T relational
operators should also be static member functions, because
the compiler always prefers member functions over friend
functions, even if the argument types do not match.
- QBAV helper functions cannot be used in generic code. Also,
qCompareThreeWay() would not work for QBAV.
To fix the issue this patch explicitly adds a
QByteArrayView(QLatin1StringView) constructor which is not
restricted by any trait. To keep the constructor constexpr, we
have to move the actual definition into qlatin1stringview.h
header file.
Integrity compiler also complains about
QByteArrayView(QUtf8StringView) constructors, so the same approach
is used to enable them.
This allows to convert all the private static member helpers in
QByteArrayView to hidden friends.
The extended tests will be added in a follow-up commit, because
they require some more changes.
Amends fb50ab7006 and
a08bafc920.
This patch removes some exported methods and adds new ones, but
it is not BiC, because the aforementioned patches are only merged
into dev branch at this point.
Task-number: QTBUG-108805
Change-Id: Iee6526a71d859c4fcb2e95bf20fe84ddead6dfb0
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Focus abstraction in QWidgetPrivate makes direct access to
QWidget::focus_next and focus_prev an antipattern.
Remove usage.
Task-number: QTBUG-121478
Change-Id: I741e6875e686a9cfb4e6a113e7575c911a38e80c
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
Focus abstraction in QWidgetPrivate makes direct access to
QWidget::focus_next and focus_prev an antipattern.
Remove usage.
Task-number: QTBUG-121478
Change-Id: Iaab97024c20b97a0d9850dce43816a432d4bc7a1
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
When high-dpi scaling is enabled, we default to HintNone
currently, since otherwise text layouts (and widget layouts
and anything depending on text size) will have to be updated
when the window is dragged onto a screen with a different dpr.
The check for whether scaling is enabled was based on
QHighDpiScaling::isActive(), which is technically incorrect
since this does not return whether there is a scale on the
painting, rather whether the coordinate system in Qt matches
the one of the platform.
Now that we support agreeing on fractional scale factors with
the Wayland compositor, this issue has become visible, since
QHighDpiScaling::isActive() will now return false for these
compositors, even for fractional scales. For integer scales,
the issue existed before as well, but the kerning issues are
less noticeable in that case.
Pick-to: 6.6 6.7
Fixes: QTBUG-122910
Change-Id: Ic82b07d57a06a351255619f9227dd60396903ade
Reviewed-by: David Edmundson <davidedmundson@kde.org>
Reviewed-by: Kai Uwe Broulik <kde@privat.broulik.de>
Palette and font changes made by a style sheet are propagated to
existing widgets and their children at change time.
Palette and font changes made by setPalette() and setFont() are
inherited by all existing and future children of the widget to which the
call was made.
Clarify this in the documentation.
Fixes: QTBUG-122588
Pick-to: 6.7 6.6 6.5 6.2 5.15
Change-Id: Ic40d96fc1e5e4507f84a33138303b7193948d3fe
Reviewed-by: Richard Moe Gustavsen <richard.gustavsen@qt.io>
Amends commit 41f84f3ddb - QDate and
QDateTime declare fromString() methods only when feature datestring is
enabled. So their inline implementations should also be conditioned on
that feature.
Pick-to: 6.7 6.7.0
Change-Id: I84fc877001d3fc97c6ca149864e4ad5a2dbabe87
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
The feature remains disabled for now since its implementation is
incomplete. This facilitates introducing the various parts of it in
reviewable chunks, prior to actually plumbing it into the code that'll
end up using it and enabling it. Includes stubs of the files that
shall contain its source code.
Task-number: QTBUG-115158
Change-Id: Ie49251c0aaf5feaf1ff522e469479020e4c73065
Reviewed-by: Alexey Edelev <alexey.edelev@qt.io>
CreateStream returns a new error code, StreamIdsExhausted, when next
stream id counter exceeds max stream id instead of assert.
Task-number: QTBUG-122375
Pick-to: 6.7
Change-Id: I653b20c24c1429fe88d476beb1ca952aa1bbb320
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
This reverts commit fc33fea999.
The change caused issues with system-wide PUA fallbacks on
platforms where this is supported. It needs to be replaced by
an approach which still falls back, but only for fonts which
are explicitly categorized as PUA fallbacks.
Pick-to: 6.5 6.6 6.7
Task-number: QTBUG-110502
Change-Id: I985a1f8076645593c50e81759872b4227d0fcd0d
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
Change beginEntryList() to take a path parameter, which it passes on
to the QAFEIterator constructor; setting the path at construction
makes more sense, because typically the path isn't supposed to change
during iteration, and this simplifies the code at the call site.
Remove setPath(), the last usage in Qt repos was in QtCreator, and that
has been ported away from it.
Change-Id: I01baa688e0f9b582aacb63d7d98a794276e58034
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
according to DRY. Also don't use erase after unique call,
just use new past-the-end iterator.
Change-Id: I5c033b6433105842e72eca9a7a2d5ea9ec0032ec
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
The postgresql server by default returns the datetime in it's local
timezone. This works as long as this is the same as on the client. If
they are different, the parsing is going wrong.. Therefore let the
server return the datetime in UTC. Also do not convert the datetime into
local time to be in sync with the MySQL plugin.
[ChangeLog][SQL][PostgreSQL] Fixed a bug where a wrong QDateTime might
be returned when the PostgreSQL server and the Qt client had different
time zones configured.
Fixes: QTBUG-115960
Change-Id: I1a6dda69359a34b99ef399b2a54f35c8ba041326
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
We only activated our fake context sharing logic for offscreen
surfaces. This commit removed that restriction this allows the
rendercontrol example to run
Fixes: QTBUG-117410
Change-Id: Id86617f57c6a98a920f5dd82332dcd8277f103e4
Reviewed-by: Lorn Potter <lorn.potter@gmail.com>
We link object files with the same names into the BundledLibjpeg
static archive.
This caused warnings when running dsymutil on shared libraries using
that archive:
skipping debug map object with duplicate name and timestamp
could not find object file symbol for symbol _jpeg_start_compress
Avoid that by creating copies of the source files with different
names, so that all object files are unique.
Pick-to: 6.2 6.5 6.6 6.7
Fixes: QTBUG-123324
Change-Id: I1d4ebdd111b4172cde793671fbe059957f102871
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Alexey Edelev <alexey.edelev@qt.io>
At the time of bb1f616ff1 the
Flickable.flickDeceleration property still applied to both touch
flicking and mouse wheel scrolling. In qtdeclarative
b1766d9d629f61d824146e69f1f3b319cbee3d11 we decoupled them. Switching
from the traditional 1500 logical pixels/sec² to 5000 was not enough to
satisfy those who complained about the mouse wheel "not being linear"
and at the same time made touch-flicking feel too sluggish. So let's
restore the traditional default deceleration value.
The flickDeceleration property is still adjustable, and the default
can still be overridden in any QPlatformTheme subclass.
[ChangeLog][QPA] The default value for the platform FlickDeceleration
hint is reverted to 1500 (rather than 5000 as in Qt 6.6). This sets
the default value for Flickable.flickDeceleration, which can be
overridden directly; and the default can also be overridden in any
QPlatformTheme subclass.
Pick-to: 6.7
Task-number: QTBUG-35608
Task-number: QTBUG-35609
Task-number: QTBUG-52643
Task-number: QTBUG-97055
Fixes: QTBUG-121500
Change-Id: If52b61dfcd0c08a7c6e753f39dbe01f417e94bf4
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
Reviewed-by: Richard Moe Gustavsen <richard.gustavsen@qt.io>
Even in the original patch where the comment was added the
callouts were made with the lock held.
Change-Id: Id8d122010d2195d83ddd4fdaf638f7fc4ac8163d
Reviewed-by: Ivan Solovev <ivan.solovev@qt.io>
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
A qt installation might not contain the non-versioned the qtpaths.exe
installed, but keep the versioned qtpaths6.exe.
Try to use the versioned version before the non-versioned one.
If none exists, show a warning at deployment time.
Amends 571201603a
Pick-to: 6.5 6.6 6.7
Fixes: QTBUG-122664
Task-number: QTBUG-119619
Change-Id: I23caf9ed3c7928fbab9657b0c0c64517dfc7d68e
Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
Change 997fd3b88e fixed integer overflows
with huge texts. This was done by using qsizetype for size calculations
instead of int. However, that change introduced a serious regression
due to an itermediate imultiplication result being "promoted" to unsigned,
and therefore a negative value being converted to a large positive.
The solution is to make sure all values in the expression are signed.
Fixes: QTBUG-123339
Task-number: QTBUG-119611
Pick-to: 6.7
Change-Id: I3f9189f77b383c6103cf5b35981cdb607b065f6f
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@qt.io>
to reuse existing buffer of QString and save some allocations
Change-Id: I31810c2fd3f0f70b19c19a530600e8cee5d6631a
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Add a template argument a la std::set and a ctor taking a Compare
object (thus supporting stateful comparators, too), and storing it in
QtPrivate::CompactStorage. Rewrite lookup() to use the stored
comparator instead of std::less, carefully avoiding a copy to match
what std::map implementations do, even though key_comp() itself is
specified to return by value.
Change-Id: I4f74aaf5eda0a4ed276e99f73f407042c2767828
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
... moving it away from between a comment and the C++ code the comment
pertains to.
Pick-to: 6.7
Change-Id: I56b3ded01d1800bae19121e4b9340d0c43f1da85
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Reviewed-by: Ahmad Samir <a.samirh78@gmail.com>
There's no reason to check whether the timer has expired after
rounding the timer's native nanoseconds up to milliseconds, so delay
the ceil<ms> operation to after the timer expiry check.
As a drive-by, protect the std::min call from Windows macros with the
parentheses trick and employ C++17 if-with-initializer.
Remove the break-if-expired, as it now obviously duplicates the while
exit condition.
Pick-to: 6.7
Change-Id: If30421012143640c75a9a44fe711d0c1c7cd23b9
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Some outputs weren't guarded with the optVerboseLevel which caused them
to occur even with the --list option. Add a guard to prevent for that.
Fixes: QTBUG-122257
Pick-to: 6.7 6.6
Change-Id: Ide060cda4ac6f9b4470ca608120e2b8aa4819de5
Reviewed-by: Oliver Wolff <oliver.wolff@qt.io>
When connecting a QFutureWatcher to the QFuture it will connect to the
output interface, which will queue up events to notify about the current
state. This happens in the thread of the QFutureWatcher.
Since 07d6d31a4c unfortunately the sending
of those events was done outside the lock, meaning the worker-thread
could _also_ send events at the same time, leading to a race on which
events would be sent first.
To fix this we move the emission of the events back into the lock
and because it is now inside the lock again anyway, we will revert
back to posting the callout events immediately, so this patch also
partially reverts 07d6d31a4c
Fixes: QTBUG-119169
Pick-to: 6.7 6.6
Change-Id: If29ab6712a82e7948c0ea4866340b6fac5aba5ef
Reviewed-by: Arno Rehn <a.rehn@menlosystems.com>
Reviewed-by: Jarek Kobus <jaroslaw.kobus@qt.io>
Creating a QString from a QDate/QTime works even when
QT_CONFIG(datestring) is not defined, so no need to ifdef it out.
Pick-to: 6.7
Change-Id: Ib3594036f309393b612d3fbf21f51be9c36a9391
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Use view types more to avoid needless allocations
Change-Id: Ifdf5c8f6fecc54a0583444fbf3fe151c2c20002e
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Also add some categorized logging output to make it easier
to check if native menus are being created.
[ChangeLog][Qt Core] Added AA_DontUseNativeMenuWindows
application attribute. Menu popup windows (e.g. context menus,
combo box menus, and non-native menubar menus) created while this
attribute is set to true will not be represented as native top level
windows, unless required by the implementation.
Task-number: QTBUG-69558
Change-Id: Iba11b89c67d942ce6c5a28a7c57a86e63c020618
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
The MYSQL_TIME structure doesn't support per-datum timezone and in any
case the server would not store it: the TIMESTAMP type is always stored
in UTC. So instead let's configure the session time zone to UTC and use
QDateTime to convert to/from it.
Fixes https://bugs.kde.org/show_bug.cgi?id=483060
[ChangeLog][SQL][MySQL] Fixed a bug in passing QDateTime to be passed as
local time to the server, regardless of the QDateTime's time zone
setting. This would cause certain timestamps to be rejected by the
server, such as a UTC time stamp whose time numerically matched the
local timezone's spring forward gap in the transition into Daylight
Savings Time.
Change-Id: I6818d78a57394e37857bfffd17bbce4ae43e823c
Reviewed-by: Christian Ehrlicher <ch.ehrlicher@gmx.de>
It was the only place that called it. Makes the code slightly uglier,
but removes a function that returned a raw pointer. More importantly, it
gets the actual type from QVariant, without relying on it internally
converting from QDateTime to QDate and QTime, or failing to do so in
some cases. This is going to be needed for the next commit.
Pick-to: 6.5 6.6 6.7
Change-Id: I6818d78a57394e37857bfffd17bbcd3f5057eadc
Reviewed-by: Christian Ehrlicher <ch.ehrlicher@gmx.de>
Replace all divisions by constant with multiplications of its inverse
Change-Id: I05aa0631e8117e7d42da0eaa30077cd230caa919
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
When declaring a constructor, you must use the injected name, not a
template.
qfutureinterface.h:472:37: error: template-id not allowed for constructor in C++20 [-Werror=template-id-cdtor]
Pick-to: 6.6 6.7
Change-Id: I6818d78a57394e37857bfffd17bbbf2313001cbf
Reviewed-by: Ahmad Samir <a.samirh78@gmail.com>
Allows to use it without Qt Widgets or Qt Graphics Views.
Since anyone having imported this through Qt Widget will also be
linking Qt Gui this should be compatible.
qt_halfScaled is also unexported for lack of known users.
Change-Id: I39406dfd4839ed46f8cbfd4651577ab6ece9932c
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
When a QDeadLineTimer::isForever() then its remainingTime() is
chrono::nanoseconds::max(), which is exactly representable as
chrono::milliseconds, and greater than 10ms, so the following code
would have done the right thing already.
Let it.
This also removes the duplicate mentioning of the 10ms sleeping
timeslice.
Amends fa296ee1dc.
Pick-to: 6.7
Change-Id: Ibc32d6069b78cd4583df07d0707d98645440b36c
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
...and adjust it to its presence in the new QDeadlineTimer overload,
which sports nanoseconds granularity.
Pick-to: 6.7
Change-Id: Ifa9658ca32c5dc4bef5cf36dec2e452174eebe1c
Reviewed-by: Ahmad Samir <a.samirh78@gmail.com>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
The current approach doesn't work correctly because external project
steps that we run on per-ABI build directories look as
'cmake --build <abi/build/dir> --target <target_name>'. When the
project has a single APK this works perfectly fine. But when building
more than one apk this leads to the simultanous build in the same
'abi/build/dir' which causes undefined behavior and concurrent access
to the build artifacts. This is especially sensible when APK targets
have the common dependencies.
The solution is split for two usecases:
- Ninja-like generators, that support job pools.
- Other generator.
For Ninja-like generators we now create job pools per-ABI with job number 1,
this convinces ninja to run only one 'cmake --build' command in single ABI
scope.
For other generators the solution is not that good. We create the dependency
chain between all APK targets and this leads to the build of the unwanted
dependencies. For example if project has apkA and apkB targets, then
apkB_x86 will depend on apkA_x86(assuming x86 is not the main ABI). This is
the only way we may ensure that 'cmake --build' commands won't be running
simultanously.
Fixes: QTBUG-122838
Pick-to: 6.7
Change-Id: I6f48fae57047a29129836168c28e14cde4eaa958
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
Currently qtloader.js fetches and copies the files manually. By doing
so we are missing some preproccessing by Emscripten preload plugins.
Use Emscripten API to preload files, so preload plugin for .so can
download, compile and resolve dependencies of imported shared libraries.
This makes looking for dependencies in preload_qml_import.py no longer
needed. Remove redundant code.
Fixes: QTBUG-121817
Change-Id: Idd35f25d5f54123910f813a636407eea23e157cb
Reviewed-by: Piotr Wierciński <piotr.wiercinski@qt.io>
...in particular when doing multisampling with multiview.
With this the results are now identical with multiview and
multiview+MSAA on the Quest 3. (previously the depth buffer
was clearly broken when doing multiview+MSAA)
Change-Id: Iec3c6af66510ab76cb0591eb8d002aa5855a399e
Reviewed-by: Andy Nichols <andy.nichols@qt.io>
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Also implement this for OpenGL ES since it can be relevant with tiled
architectures wrt performance.
Task-number: QTBUG-122669
Change-Id: I90dcfe4f5f9edbb8dfb51189d46b89ef2c7a7c06
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Andy Nichols <andy.nichols@qt.io>
One is not enough, we have to have three code paths for three sets of
APIs/extensions.
Task-number: QTBUG-122669
Change-Id: Id10e19dfcfe16a8c21b6c932965f09539f061549
Reviewed-by: Andy Nichols <andy.nichols@qt.io>
When a QWidget was reparented, we would take care to reparent its
backing QWidgetWindow as well, into the nearest QWindow of the
new QWidget parent.
However we would only do this for the reparented widget itself,
and not any of its child widgets. In the case where the widget
has native children with their own QWindows, the widget itself
may not (yet) be native, e.g. if it hasn't been shown yet, or
if the user has set Qt::WA_DontCreateNativeAncestors.
In these scenarios, we would be left with dangling QWindows,
still hanging off their original QWindow parents, which
would eventually lead to crashes.
We now reparent both the QWindow of the reparented widget (as
long as it's not about to be destroyed), and any QQWindow
children we can reach. For each child hierarchy we can stop
once we reach a QWindow, as the QWindow children of that
window will follow along once we reparent the QWindow.
QWindowContainer widgets don't usually have their own
windowHandle(), but still manage a QWindow inside their
parent widget hierarchy. These will not be reparented
during QWidgetPrivate::setParent_sys(), but instead
do their own reparenting later in QWidget::setParent
via QWindowContainer::parentWasChanged(). The only
exception to this is when the top level is about to
be destroyed, in which case we let the window container
know during QWidgetPrivate::setParent_sys().
Finally, although there should not be any leftover
QWindows in the reparented widget once we have done
the QWidgetWindow and QWindowContainer reparenting,
we still do a pass over any remaining QWindows and
reparent those too, since the original code included
this as a possibility.
We could make further improvements in this areas, such
as moving the QWindowContainer::parentWasChanged() call,
but the goal was to keep this change as minimal as possible
so we can back-port it.
Fixes: QTBUG-122747
Pick-to: 6.7.0 6.7 6.6 6.5
Change-Id: I4d1217fce4c3c48cf5f7bfbe9d561ab408ceebb2
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
If the dialog is closed by pressing a button, the button will be
reflected via clickedButton(), and the result() will reflect either
the QMessageBox::StandardButton value, or an opaque value for custom
buttons.
Depending on the role if the buttons, the accepted or rejected
signals are emitted.
If the user called accept() or rejecct() on a dialog that had
already been closed by a button, we would as a result of
1f70c073d4 emit a signal based
on the original button that was clicked, instead of respecting
the newly triggered function.
It's a questionable use-case, as the clickedButton() is still
the original button e.g., but we should still avoid regressing
this, so we now emit the signals based on the newly stored
result code instead of using the clicked button as the source
of truth.
To allow this we had to change the opaque result() value for
custom buttons to stay out of the QDialog::DialogCode enum,
but this should be fine as the documentation explicitly says
that this is an opaque value.
Fixes: QTBUG-118226
Pick-to: 6.7 6.6 6.5
Change-Id: Ia2966cecc6694efce66493c401854402658332b4
Reviewed-by: Axel Spoerl <axel.spoerl@qt.io>
If a custom application delegate is installed prior to creating the
Qt application delegate we will forward callbacks to the delegate,
but this has to be done manually for any callback we implement.
Fixes: QTBUG-122996
Pick-to: 6.7 6.6 6.5
Change-Id: Ia25e2c4b8cac37130d604c772c875c5d76c66764
Reviewed-by: Morten Johan Sørvig <morten.sorvig@qt.io>
QMessageBox::setTextFormat can set the text format for its text label,
but the informative label will still use its default text format. This
change allow the setTextFormat also apply to the informative label, thus
user can use their preferred format like Markdown in the informative label.
Task-number: QTBUG-122712
Change-Id: I6f7487b29712ca07cee27171453312ff1707eea3
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
QRestReply is no longer a QObject, so remove the includemocs line from
the .cpp file.
Amends 9ba5c7ff6a.
Pick-to: 6.7
Change-Id: I6c0ba6b9e3b82f84f3b509755e7da5b33e607776
Reviewed-by: Juha Vuolle <juha.vuolle@qt.io>
When fa296ee1dc ported this function
from int timeout to QDeadlineTimer, the need to keep this variable
outside the do-while so it could be checked in the loop exit condition
fell away.
Moving the definition of the variable to the first (and only)
assignment makes the code clearer and the variable a constant.
Amends fa296ee1dc.
Pick-to: 6.7
Change-Id: I7a0fe01dc68ff140beeb0e76b141c84d4bd28458
Reviewed-by: Ahmad Samir <a.samirh78@gmail.com>
qiterable.cpp is even a comment-only source.
Change-Id: I01ec3c774d9943adb903fffd17b79d567a435594
Reviewed-by: Ahmad Samir <a.samirh78@gmail.com>
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Only QAccessibleWidget was using them, and this was changed some time
ago. No point in keeping dead code.
Change-Id: I14bc40e6d87df234987e82385ce13433c2b82744
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
The multipleDirs warning was added to avoid surprises related to where
windeployqt would deploy files when using binaries from different paths.
To do this properly, make the warning message more meaningful, and
suppress the warning when the --dir option is specified, i.e. when the
user is already explicitly choosing where to deploy.
Pick-to: 6.7 6.6
Change-Id: Ie2984f4af740776c568610370d49ad4ff85ffff0
Reviewed-by: David Faure <david.faure@kdab.com>
Reviewed-by: Oliver Wolff <oliver.wolff@qt.io>
Unlike most of everything else in the Bootstrap lib, this is code that
couldn't be eliminated by the linker because they were referenced in one
static array. Maybe an exceptionally smart whole-program analysis could
do it, but GCC and Clang LTO modes don't do that now.
I removed the code that performed detection from HTML and from data too.
I could have left the detection of UTF-8 and "other" but this code
wasn't necessary. In particular, QTextStream couldn't benefit from it
because it already defaults to UTF-8, so the detection code would never
determine anything different from the input.
Drive-by removed QStringConverter::availableCodecs() too because it was
in the middle of functions #ifdef'ed out to.
This reduced the size of release-mode moc
text data bss dec hex filename
1079858 5440 640 1085938 1091f2 original/moc
1074386 5200 640 1080226 107ba2 updated/moc
-5472 -240 0 -5712 difference
Change-Id: I01ec3c774d9943adb903fffd17b7f114c42874ac
Reviewed-by: Lars Knoll <lars@knoll.priv.no>
That is, use the inline functions that refer to <numeric_limits> and
<cmath> directly, instead of the out-of-line wrappers. Someone should
verify if the hacks for QNX's <math.h> are still required.
Change-Id: I01ec3c774d9943adb903fffd17b7ee560b4b71b9
Reviewed-by: Ahmad Samir <a.samirh78@gmail.com>
Bootstrapped tools don't usually need to list directories; they should
operate on file lists passed to it by the build system instead.
This may deserve a QT_FEATURE.
Change-Id: I01ec3c774d9943adb903fffd17b7ecfba2702fc5
Reviewed-by: Ahmad Samir <a.samirh78@gmail.com>
Just so we don't accidentally use different precisions. Amends commit
bfc7535a10.
Change-Id: I83dda2d36c904517b3c0fffd17b5258e88dd194e
Reviewed-by: Ahmad Samir <a.samirh78@gmail.com>
They're all ported in one go because all the changes are the same and
they all rely on QTimerInfoList. The changes are:
- use Qt::TimerId to uniquely identify timer IDs
- use Duration (nanoseconds) to specify the timer interval
- rename registeredTimers() to timersForObject(), which is const
Change-Id: I83dda2d36c904517b3c0fffd17b52958767d8a68
Reviewed-by: Ahmad Samir <a.samirh78@gmail.com>
No change in behavior, other than the ability for the precise timers to
be even more precise. There's a reduction in the stored interval range
from 292 million years to 292 years, but there's no change in behavior
because the timeout is stored as a steady_clock::time_point, which is
nanoseconds with libstdc++ and libc++. Now, if there's an overflow,
it'll happen on the call to registerTimer() instead of inside the
calculation of the wake-up.
Change-Id: I83dda2d36c904517b3c0fffd17b3d1f2d9505c1c
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Ahmad Samir <a.samirh78@gmail.com>
This way, we can begin using the V2 methods now, regardless of whether
the concrete dispatcher class has been ported or not.
Change-Id: I83dda2d36c904517b3c0fffd17b52a6256b083af
Reviewed-by: Ahmad Samir <a.samirh78@gmail.com>
This class is a temporary hack to enable transition to an API based on
std::chrono for the Qt event dispatcher. In Qt 7, it will be merged with
QAbstractEventDispatcher, replacing the pure virtuals there with the
ones defined here.
The new API differs from V1 in the following ways:
- uses Qt::TimerId instead of int to identify timer IDs, so we can't
accidentally confuse them with something else
- uses Duration (nanoseconds) to specify the interval, instead of a mix
of int and qint64
- add the missing const to remainingTime()
- rename registeredTimers() to timersForObject() (I'd have kept the
original name but we can't overload the name if the parameters are
exactly the same; we could have used QT6_DECL_NEW_OVERLOAD_TAIL, but
I think the new name is actually better)
Because the old API was mixing int and qint64, we didn't officially
support any timer for more than 2^31 ms (~24.85 days). This should
extend the valid range to 292 years once the dispatchers are ported
over.
Change-Id: I83dda2d36c904517b3c0fffd17b3a7e0afef4b59
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Reviewed-by: Ahmad Samir <a.samirh78@gmail.com>
Replace public operators operator==() and operator!=() of QCborValue
to friend method comparesEqual().
Replace public operator<() of QCborValue to friend method
compareThreeWay().
Use QT_CORE_REMOVED_SINCE to get rid of current comparison methods
and replace them with a friend.
Delete #if 0 && __has_include(<compare>) blocks,
since they are not required anymore.
Task-number: QTBUG-120300
Change-Id: I884ff6ce2a71618b0e3eaa907f0852f93c2a073c
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Ivan Solovev <ivan.solovev@qt.io>
QOperatingSystemVersion intentionally does not define operator==() and
operator!=() since ae072cd9c4.
It means that we cannot use comparison helper macros.
Still, we can manually define four relational operators or
operator<=>() in C++20 mode, and give the class a partial ordering.
We choose partial ordering, because versions of different OS types are
incomparable.
Implement the operators in terms of helper function compareThreeWay(),
which potentially allows to use this class in some templated code.
As a drive-by: make the static compare() function noexcept, because
it really is.
Fixes: QTBUG-120360
Change-Id: Id4c9ce740e42baa719ca0ee84146d087b21675c6
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
QTypeRevision consists of two quint8 values: major and minor version.
Each of the versions can be unknown.
The rules for comparing with the unknown version are as follows:
zero version < unknown version < non-zero version
At the same time, two unknown versions are considered equal.
This makes the comparison a bit tricky, but it still fits into the
category of strong ordering.
Replace the existing friend relational operators with helper functions
and comparison helper macros, making this type strongly ordered.
Update the unit-tests to use the new comparison helper test functions.
As the test functions check the reversed arguments as well, we can
reduce the number of rows for the data-driven comparison test.
Fixes: QTBUG-120359
Change-Id: Ib6f1037fc7b5fed148e35ee48b56b05dcd36b3b4
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Done by harmonizing the use on the QT_CONFIG(temporaryfile) macro and
fixing one test that was missing. We can't remove the older macro
because it is marked PBULIC) but we don't need to use it ourselves.
Change-Id: I01ec3c774d9943adb903fffd17b7eb4dd1a4e63f
Reviewed-by: Ahmad Samir <a.samirh78@gmail.com>
If the Android View is resized, the QWindow instantiated by it
should be resized accordingly.
Task-number: QTBUG-122626
Pick-to: 6.7
Change-Id: I7bfbca149f927718d1e28cdabfa8759afbd06039
Reviewed-by: Assam Boudjelthia <assam.boudjelthia@qt.io>
Qt window loading is initiated either when the QtView is attached
to its Android window, or when the Android QPA plugin has been loaded
and is ready, depending on the order. Since the window attachment
happens in the Android UI thread, and the Android QPA plugin callback
happens in Qt thread, add synchronized block to make sure the execution
stays ordered.
Fixes: QTBUG-122626
Pick-to: 6.7
Change-Id: Id476032f02aa8990432a02f62b6bf6237a17e7ac
Reviewed-by: Assam Boudjelthia <assam.boudjelthia@qt.io>
The SPDX database lists the license as 'Unicode-3.0', and 'Unicode
License v3'. Now, the SPDX standard actually says that
License identifiers (including license exception identifiers) used
in SPDX documents or source code files should be matched in a case-
insensitive manner.
But the website at https://spdx.org/licenses/ doesn't treat it this way,
so the link we generate out of the identifier actually gives a 404. So
it's just easier to use the 'original' capitalization.
Amends 063026cc50
Pick-to: 6.5 6.6 6.7
Change-Id: I826077a914721b7b9499ad62c08fdf20be94e88d
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Some libs don't necessarily have the lib prefix in their names,
3rd party libs and Qt for Python might have that, so no need to
always add that prefix to loaded libs is the lib name already
contains a .so suffix.
Fixes: QTBUG-123286
Pick-to: 6.7 6.7.0
Change-Id: Ib65215d9b4410c5c9e00aa0642f48ab45c92fe03
Reviewed-by: Shyamnath Premnadh <Shyamnath.Premnadh@qt.io>
Reviewed-by: Tinja Paavoseppä <tinja.paavoseppa@qt.io>
In the \fn commands for a limited number of methods in the
documentation for Testlib and Widgets, `= 0` is passed as default
argument instead of `= Qt::KeyboardModifiers()`. Until QDoc with Clang
17, inclusive, QDoc generated the correct signature. However, with
Clang 18, QDoc outputs `= 0` in the documentation. While strictly
speaking still correct, this change impacts the documentation
negatively in terms of readability.
Dropping the default argument from the \fn command ensures that QDoc
generates the right signature with both Clang 17 and Clang 18.
Task-number: QTBUG-123130
Pick-to: 6.7
Change-Id: I94ccec2f2c9a02241095fb5b18feb74aa55f97e1
Reviewed-by: Andreas Eliasson <andreas.eliasson@qt.io>
Reviewed-by: Topi Reiniö <topi.reinio@qt.io>
We want to be able to pass as an argument whether the texture grabbed to
an FBO is supposed to have a flipped y-axis or not. This is required for
screen capture on the EGLFS platform.
Task-number: QTBUG-121835
Pick-to: 6.7 6.6 6.5
Change-Id: I6dddc879a4be7ff2c2c189747193423644be55a0
Reviewed-by: Laszlo Agocs <laszlo.agocs@qt.io>
Reviewed-by: Artem Dyomin <artem.dyomin@qt.io>
We already have fine-grained macros to individually disable APIs
that we consider "suboptimal" or "dangerous". This commit adds
a shortcut for the user to set all such macros in one go.
QT_ENABLE_STRICT_MODE_UP_TO is versioned, just like
QT_DISABLE_DEPRECATED_UP_TO; the idea is that users should set it
to the minimum Qt version they want to support.
Also, if QT_DISABLE_DEPRECATED_UP_TO is not set, then
QT_ENABLE_STRICT_MODE_UP_TO will set it as well, to the same value.
[ChangeLog][QtCore][QtGlobal] Added the QT_ENABLE_STRICT_MODE_UP_TO
macro.
Change-Id: I5466465986104e047a6a86369928be9294f24ab7
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
Give it a name in line with the function name:
qAsConst -> QT_NO_QASCONST, as already done for qExchange.
We can do this because we never documented the macro itself.
So, while at it: also document the macro.
Change-Id: I6eb0834df438e4f4e818ef2cf8e702ed156dc253
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
The EventPointMap QPointingDevicePrivate::activePoints is actually a
QVarLengthFlatMap<., ., 20>, not 16, so copy the value instead of just
hard-coding something.
Amends 296ede3aab.
Change-Id: Ic8e83f4095f57be74f7708d5cec6a19971772b76
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
Before the qWaitFor() port from int to QDeadlineTimer, the
implementation constructed a QDeadlineTimer internally, passing int
timeout and Qt::PreciseTimer. The int overload that was retained for
source-compatibility, however, constructs the QDeadlineTimer without
the PreciseTimer flag, which is a behavior change.
Restore the Qt 6.6 behavior and pass Qt::PreciseTimer.
Amends fa296ee1dc.
Pick-to: 6.7 6.7.0
Change-Id: Ib8e5b912c74b70d32a77195edb0d2a30cd7c241d
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
Wrap dlerror() in QLatin1StringView(), like we already do in
unload_sys().
Amends a6a5681470.
Pick-to: 6.7
Change-Id: Ia8c91d6962c74d5916c47b2abbb3920f051c8e5e
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
The bootstrapped tools really mustn't produce random output (they must
always be reproducible exactly). Therefore, ensure we don't need this
file.
Change-Id: I01ec3c774d9943adb903fffd17b7eb94dbd4be89
Reviewed-by: Alexey Edelev <alexey.edelev@qt.io>
Reviewed-by: Ahmad Samir <a.samirh78@gmail.com>
Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
I added QT_NO_VARIANT to qconfig-bootstrapped.h to be clearer on what
the #ifs are, but there's no testing of that feature outside of
QT_BOOTSTRAPPED.
Change-Id: I01ec3c774d9943adb903fffd17b7e8ac4340fb89
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
It was only used by the cmake_automoc_parser so it would write a 64-bit
in big-endian format. So bypass QDataStream and write it native
endianness.
Change-Id: I01ec3c774d9943adb903fffd17b79c78e56db4cf
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Alexey Edelev <alexey.edelev@qt.io>
Commit 90bc0ad41f ("QProcess/Unix: add
failChildProcessModifier()") added this line that set childStartedPipe
so that the failChildProcess() callback had something to write to. But
we left it set on exit from QProcessPrivate::startDetached(), which
caused the QProcess destructor to try and close it.
Noticed when debugging the issue for QTBUG-123083.
Pick-to: 6.7 6.7.0
Task-number: QTBUG-123083
Change-Id: I6818d78a57394e37857bfffd17bbc41c8400270f
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@gmx.de>
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
Currently unused. I've tested that GCC 13 and Clang 17 do compile this
and do output reasonable assembly, with and without AVX512FP16 (I also
tested AVX10.1 with GCC 14).
Clang is unable to allocate a "v" (vector) register when T is qfloat16,
so I could only implement the conversion using _Float16. Strictly
speaking, Clang >= 16 always knows about _Float16, but I didn't want to
have a different detection from qtypes.h.
Change-Id: I6818d78a57394e37857bfffd17bae860f8055324
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
The x86 SSE instructions do what we want and set a flag in the MXCSR to
indicate whether we should return false, so we can improve the codegen.
GCC generates for QStringView::toFloat():
call _ZNK11QStringView8toDoubleEPb
movl $8064, 12(%rsp)
movl $1, %eax
#APP
# 330 "/home/tjmaciei/src/qt/qt6/qtbase/src/corelib/global/qnumeric_p.h" 1
vldmxcsr 12(%rsp)
vcvtsd2ss %xmm0, %xmm0, %xmm0
vstmxcsr 12(%rsp)
# 0 "" 2
#NO_APP
movl 12(%rsp), %edx
testb $24, %dl
je .L2120
... handling of the under/overflow ...
The MXCSR instructions do need to read and write from memory, but the
stack is usually already in L1 and CPUs have special optimizations for
it.
There are two alternative implementations to the implementation chosen:
first, we could confirm there was no underflow or overflow using an
expression like:
(v == 0) == (*value == 0) && qt_is_finite(v) == qt_is_finite(*value);
But that is still very costly, with 4 UCOMISx instructions and several
memory loads.
Second, we could use the VFPCLASSSD and VFPCLASSSS (yes, 4 "S")
instructions to confirm whether a finite input became zero or non-
finite, but a) that's only available with AVX512, so of little practical
use today and b) it has a 3-cycle latency. Like the comparisons above,
we'd need 4 of them.
Change-Id: I01ec3c774d9943adb903fffd17b8b9cb2ce805ce
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
No change in behavior. But instead of trying to guess if we're going to
cross a page boundary, we check for the cacheline: if we don't cross a
cacheline boundary, then we can't cross a page boundary either. Testing
bit 4 can be implemented with a shorter instruction than a test for bit
11.
f43de: 40 f6 c7 20 test $0x20,%dil
Change-Id: I664b9f014ffc48cbb49bfffd17b04817a0fb8c6b
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
It's the same aeshash() as before, except we're passing a template
parameter to indicate whether to read half and then zero-extend the
data. That is, it will perform a conversion from Latin1 on the fly.
When running in zero-extending mode, the length parameters are actually
doubled (counting the number of UTF-16 code units) and we then divide
again by 2 when advancing.
The implementation should have the following performance
characteristics:
* QLatin1StringView now will be roughly half as fast as Qt 6.7
* QLatin1StringView now will be roughly as fast as QStringView
For the aeshash128() in default builds of QtCore (will use SSE4.1), the
long loop (32 characters or more) is:
QStringView QLatin1StringView
movdqu -0x20(%rax),%xmm4 | pmovzxbw -0x10(%rdx),%xmm2
movdqu -0x10(%rax),%xmm5 | pmovzxbw -0x8(%rdx),%xmm3
add $0x20,%rax | add $0x10,%rdx
pxor %xmm4,%xmm0 | pxor %xmm2,%xmm0
pxor %xmm5,%xmm1 | pxor %xmm3,%xmm1
aesenc %xmm0,%xmm0 aesenc %xmm0,%xmm0
aesenc %xmm1,%xmm1 aesenc %xmm1,%xmm1
aesenc %xmm0,%xmm0 aesenc %xmm0,%xmm0
aesenc %xmm1,%xmm1 aesenc %xmm1,%xmm1
The number of instructions is identical, but there are actually 2 more
uops per iteration. LLVM-MCA simulation shows this should execute in the
same number of cycles on older CPUs that do not have support for VAES
(see <https://analysis.godbolt.org/z/x95Mrfrf7>).
For the VAES version in aeshash256() and the AVX10 version in
aeshash256_256():
QStringView QLatin1StringView
vpxor -0x40(%rax),%ymm1,%ym | vpmovzxbw -0x20(%rax),%ymm3
vpxor -0x20(%rax),%ymm0,%ym | vpmovzxbw -0x10(%rax),%ymm2
add $0x40,%rax | add $0x20,%rax
| vpxor %ymm3,%ymm0,%ymm0
| vpxor %ymm2,%ymm1,%ymm1
vaesenc %ymm1,%ymm1,%ymm1 <
vaesenc %ymm0,%ymm0,%ymm0 vaesenc %ymm0,%ymm0,%ymm0
vaesenc %ymm1,%ymm1,%ymm1 vaesenc %ymm1,%ymm1,%ymm1
vaesenc %ymm0,%ymm0,%ymm0 vaesenc %ymm0,%ymm0,%ymm0
> vaesenc %ymm1,%ymm1,%ymm1
In this case, the increase in number of instructions matches the
increase in number of uops. The LLVM-MCA simulation says that the
QLatin1StringView version is faster at 11 cycles/iteration vs 14 cyc/it
(see <https://analysis.godbolt.org/z/1Gv1coz13>), but that can't be
right.
Measured performance of CPU cycles, on an Intel Core i9-7940X (Skylake,
no VAES support), normalized on the QString performance (QByteArray is
used as a stand-in for the performance in Qt 6.7):
aeshash | siphash
QByteArray QL1SV QString QByteArray QString
dictionary 94.5% 79.7% 100.0% 150.5%* 159.8%
paths-small 90.2% 93.2% 100.0% 202.8% 290.3%
uuids 81.8% 100.7% 100.0% 215.2% 350.7%
longstrings 42.5% 100.8% 100.0% 185.7% 353.2%
numbers 95.5% 77.9% 100.0% 155.3%* 164.5%
On an Intel Core i7-1165G7 (Tiger Lake, capable of VAES and AVX512VL):
aeshash | siphash
QByteArray QL1SV QString QByteArray QString
dictionary 90.0% 91.1% 100.0% 103.3%* 157.1%
paths-small 99.4% 104.8% 100.0% 237.5% 358.0%
uuids 88.5% 117.6% 100.0% 274.5% 461.7%
longstrings 57.4% 111.2% 100.0% 503.0% 974.3%
numbers 90.6% 89.7% 100.0% 98.7%* 149.9%
On an Intel 4th Generation Xeon Scalable Platinum (Sapphire Rapids, same
Golden Cove core as Alder Lake):
aeshash | siphash
QByteArray QL1SV QString QByteArray QString
dictionary 89.9% 102.1% 100.0% 158.1%* 172.7%
paths-small 78.0% 89.4% 100.0% 159.4% 258.0%
uuids 109.1% 107.9% 100.0% 279.0% 496.3%
longstrings 52.1% 112.4% 100.0% 564.4% 1078.3%
numbers 85.8% 98.9% 100.0% 152.6%* 190.4%
* dictionary contains very short entries (6 characters)
* paths-small contains strings of varying length, but very few over 32
* uuids-list contains fixed-length strings (38 characters)
* longstrings is the same but 304 characters
* numbers also a lot contains very short strings (1 to 6 chars)
What this shows:
* For short strings, the performance difference is negligible between
all three
* For longer strings, QLatin1StringView now costs between 7 and 17% more
than QString on the tested machines instead of up to ~50% less, except on
the older machine (where I think the main QString hashing is suffering
from memory bandwidth limitations)
* The AES hash implementation is anywhere from 1.6 to 11x faster than
Siphash
* Murmurhash (marked with asterisk) is much faster than Siphash, but it
only managed to beat the AES hash in one test
Change-Id: I664b9f014ffc48cbb49bfffd17b045c1811ac0ed
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
So that it hashes to the same value as QString{,View}.
In order to test this, you must either run on a CPU other than ARM and
x86, or disable the AES hasher. I did that and can confirm siphash and
murmurhash do work with on-the-fly conversion from Latin-1.
Change-Id: I664b9f014ffc48cbb49bfffd17b03e5e62ec4e89
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
This complements the previous commit by adding the heterogeneous lookup
in operator[]. Unlike the members of the previous commit, this one may
insert into the hash, in which case it needs a way to cast from the
heterogeneous type K to the actual Key type.
Change-Id: I664b9f014ffc48cbb49bfffd17b037c1063dfb91
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
This implements support in QHash and QMultiHash for lookups of
heterogeneous key types that produce the same hash value. This is
implemented by duplicating each of the following functions into an
overload on Key and one a template that is enable_if-constrained to a
key type that meets the requirement:
* contains
* count
* equals_range
* find
* operator[] (const only)
* remove
* take
* value
* values (QMultiHash)
The non-const operator[] may insert into the hash, so it's not part of
this commit.
Change-Id: I664b9f014ffc48cbb49bfffd17b037852f0fd192
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
The typical examples are views and their corresponding owning
container. Since they are different types, the search is heterogeneous,
hence the name.
We have to do it differently from the Standard Library. There, because
the hasher is a template parameter to std::unordered_{map,set}, it can
be a structure with overloads for different types, all of which the
implementer guarantees produce the same hash for input that also
compares equal. For QHash/QSet, we don't have a template parameter.
One alternative solution would be to detect the existence of
qHashEquals(T1, T2) or qHashHeterogeneousEquals() or something.
Change-Id: I664b9f014ffc48cbb49bfffd17b0318c0775a2b5
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
They were literally identical (as in, they used exactly the same letters
in the exact same order), but changed on whether *this was const or not.
So pass *this as a parameter so we can have just one implementation.
QMultiHash doesn't need this change because its non-const equal_range()
calls the const version after making a copy.
Change-Id: I6818d78a57394e37857bfffd17ba06aee2e7db0b
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Complements commit 7fe5611365 by moving
the default value selection into a lambda that is passed to the Impl
function. This makes the two pairs of overloads in each of the classes
be a single line.
Change-Id: I6818d78a57394e37857bfffd17ba067beb3a42fa
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
std::string has them too, see constructor 10 in [1]. There, they are
StringViewLike, something we may want to do here too, which would also
lower the precedence of the constructor in the overload resolution.
This will be needed for the non-const heterogeneous QHash::operator[].
[ChangeLog][QtCore][QByteArray] Added a constructor to create a
QByteArray from QByteArrayView.
[ChangeLog][QtCore][QString] Added a constructor to create a QString
from QStringView.
[1] https://en.cppreference.com/w/cpp/string/basic_string/basic_string
Change-Id: I6818d78a57394e37857bfffd17b9ab03bd0253e6
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
The switch from <string> to <string.h> dropped the definition of
std::size. Include <iterator> since <string.h> is no longer used to
pull in std::size.
Amends: dc2ae08e02.
Pick-to: 6.7
Change-Id: Ib742538eb5d21c77fcae7ee9abb6d5329bbcfd63
Reviewed-by: Ahmad Samir <a.samirh78@gmail.com>
The relational operators were removed in
8652c79df0 with the argument that they
do not have a total order. Back than it was a valid argument, because Qt
did not support any of the C++20 ordering types.
Now Qt has its own implementation for all three ordering types, so we
could technically add relational operators and claim that QVariant
provides partial ordering.
However, that could potentially lead to many bugs and/or unexpected
results, if people start using QVariant as a key in std::map/QMap
containers.
We do not want that to happen, so we only use the helper macros to
implement (in)equality operators.
This commit also extends the unit tests, providing more extensive
testing of (in)equality operators and the pre-existing
QVariant::compare() method.
Fixes: QTBUG-113234
Change-Id: I783f3b5df552da782627f4ed0a5bb1b577753a23
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Mitch Curtis
Volker Hilsheimer
Giuseppe D'Angelo
Tor Arne Vestbø
Thiago Macieira
Laszlo Agocs
Tatiana Borisova
Fredrik Ålund
Friedemann Kleint
Albert Astals Cid
Mårten Nordheim
Bartlomiej Moskal
Axel Spoerl
Alexey Edelev
Marc Mutz
Miguel Costa
Joerg Bornemann
Allan Sandfeld Jensen
David Faure
Shawn Rutledge
Wladimir Leuschner
Morten Sørvig
Jøger Hansegård
Alexandru Croitor
Assam Boudjelthia
Timur Pocheptsov
Liang Qi
Venugopal Shivashankar
Even Oscar Andersen
Keith Kyzivat
Oliver Wolff
Ahmad Samir
Anton Kudryavtsev
Paul Wicking
Lucie Gérard
Isak Fyksen
Vladimir Belyavsky
Ulf Hermann
Mate Barany
Piotr Wierciński
Yifan Zhu
Eskil Abrahamsen Blomfeldt
Øystein Heskestad
Ivan Solovev
Edward Welbourne
Christian Ehrlicher
Paul Olav Tvete
Timothée Keller
Jarek Kobus
Kai Uwe Broulik
Wang Zichong
Tinja Paavoseppä
Kai Köhne
Mikko Hallamaa
Dennis Oberst