Added specifically to support the deprecated CVOpenGLTextureCache on
macOS, because Qt Multimedia still needs a way to use that when the
applications requests using OpenGL instead of Metal.
Follow what we did for GL_TEXTURE_EXTERNAL_OES, and add a flag that
simply makes all our glBindTexture calls use the
GL_TEXTURE_RECTANGLE[_ARB] target.
Pick-to: 6.2
Change-Id: If818b13a9f520cdb8bdc16de84a3ca0e18ad6c33
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
QPA plugins might have to close popups for events that are not delivered
to QWindow or QWidget instances. For instance, the Cocoa plugin has to
explicilty close popups when the user clicks into the window frame.
Expose this functionality through a virtual in QGuiApplicationPrivate,
and move the QApplication implementation from a static helper into the
override.
Task-number: QTBUG-96450
Change-Id: I52be5710c8d7515b9ae2e4bbadb069df4b3ed546
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
This function implements the isTopLevel() logic, which
we now don’t have to duplicate at each call site.
Change-Id: Ic8b857aa7cd3c3c23d5e950d9f50b66c81ba3ccf
Reviewed-by: Morten Johan Sørvig <morten.sorvig@qt.io>
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
The current implementation got stuck always asking for
the parent of the same child
This patch will make sure we actually walk up the
parent chain.
Pick-to: 6.2 6.2.0 6.1 5.15
Change-Id: I9f67f6305e0143526f53952a563d496e760ac2e7
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
...in the native interface.
Using OpenXR is impossible on some platforms (for example, Android)
without knowing all three. The EGLContext alone is not enough, and
EGL offers no way to query the EGLConfig a context was created with.
https://www.khronos.org/registry/OpenXR/specs/1.0/html/xrspec.html#XR_KHR_opengl_es_enable
Therefore, expose all three so libs/apps can use the new way to query
these native resource without resorting to the old-style
nativeResourceFor* queries.
Change-Id: I7efb0a26b858150da55e711752af99426e744322
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
a332f3fabc disabled resolving all fonts
on the system for every font lookup, which was a significant startup
time improvement. But it also caused a regression: When a font has
an alias which shares the name of a proper font, then this would
not be resolved correctly.
This is fairly typical on Windows/GDI due to backwards-compatibility.
Instead of being collected under a shared typographical family, fonts
are disambiguated by adding the style name to the family name. The
proper typographical name is still available, but this is not
enumerated by the system.
So "Segoe UI" for instance, will be available as "Segoe UI",
"Segoe UI Light", "Segoe UI Bold" etc.
When we populate family aliases, we register that "Segoe UI Light"
is actually "Segoe UI" with Light weight, and prior to
a332f3fabc this would be done implicitly.
But after the optimization, we would only populate family aliases once
we stumbled over a font request for a non-existent font. For "Segoe UI",
we would simply return the regular weight font as the best imperfect
match.
The fix is to populate font family aliases not only when the family is
non-existent, but when the match is imperfect, e.g. if we are asking
for a Light weight font and only finding a regular one. User code can
still avoid this somewhat expensive operation by using the full
family names on Windows.
This also requires a fix to a test. When removeApplicationFont() is
called, we invalidate the font database, so it will be reset to a state
that does not contain the family aliases. Therefore we cannot guarantee
that it is identical to what it was before the test started, since this
depends on what has happened previously in the application.
[ChangeLog][QtGui][Text] Fixed an issue where some font styles and weights
would not be selectable. This was especially noticeable on Windows.
Pick-to: 5.15 6.1 6.2
Fixes: QTBUG-94835
Change-Id: I892855edd1c8e3d3734aace396f6000d897d2ec4
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
QWINDOWSIZE_MAX is needed to implement platform plugins and Wayland
shell integrations. Moving the #define from qwindow_p.h to
qplatformwindow.h makes it available from a supported header file.
This should not break anything, since qwindow_p.h includes
qplatformwindow.h
Change-Id: Ie003b1e9d6494695057d59aba5a17ad7fd51c654
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
The make the semantics overt and avoid the joys of yet another
approximation to pi/180 hard-coded in our source-code.
Change-Id: I9dcbaada2e7de119e385fc7056bc601ecd59311a
Reviewed-by: Shawn Rutledge <shawn.rutledge@qt.io>
On platforms such as iOS, the default window state is Qt::WindowMaximized,
so that calling show() will fill the entire screen. But sub-windows should
not get this treatment, as their geometry is typically managed manually.
Pick-to: 6.2 6.2.0
Change-Id: Ib17c2011a16baf6023fe368b94fa7f38e12bd777
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
QDesktopServices::openUrl(const QUrl &url) is already implemented
on Android. But even if it is possible to set an URL handler, the
mechanism to invoke it is missing. With this commit the URL handler
will work on Android like it is already working on iOS.
Task-number: QTBUG-84382
Pick-to: 6.2
Change-Id: Ic560bd380f1cc59586861aa1a6a3ea064276a39e
Reviewed-by: Assam Boudjelthia <assam.boudjelthia@qt.io>
This reverts commit 39f0633f0b.
The Deepin desktop environment is developed based on Qt, but some
applications use dialog boxes of GTK, and some applications use dialog
boxes of Qt. This doesn't look good. The initial commit was trying
to unify this by always using GTK dialogs.
But this turned out to not look good, so instead the plan is that
all application dialogs can be based on Qt. Reverting the original
commit as a result, as it is no longer applicable.
Pick-to: 6.2
Change-Id: I2824b7b2f292f3896c9f94ccce6979f77d66578b
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
Qt wayland platform plugin has improved quite a lot and it is now pretty
much usable on Gnome. It also improves user experience a lot on HiDPI
displays.
Pick-to: 6.2
Change-Id: I943e0bb969f384bdc3c603e290ce9c8358b70f63
Reviewed-by: Aleix Pol Gonzalez <aleixpol@kde.org>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Deepin prefers the GTK3 theme implementation with native dialogs etc.
Pick-to: 6.2 6.2.0
Change-Id: I54cea9d17c0d7f5e466c700adce5aa68f0cf5564
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
The isActive function does not determine if a window should be active,
but whether it is currently active. The way the documentation was phrased
may have lead people to believe the former.
Pick-to: 6.2 6.2.0 5.15
Change-Id: I05a4cb3d8784a2fefa24bdd42ea96cfdae22b9d1
Reviewed-by: Giuseppe D'Angelo <giuseppe.dangelo@kdab.com>
The QVariant returned by resource() can contain either a QPixmap
or a QImage. The code here is now more similar to the one in
qtextimagehandler.cpp.
Also, the quality is 0 when not set, in which case we want a nice PNG
rather than a very very low quality JPG with just a few large blocks of
same-color pixels.
Pick-to: 5.15 6.2
Change-Id: I49db542e2234c8068f85a636a81a7d8cdb7b5876
Reviewed-by: André Hartmann <aha_1980@gmx.de>
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
...just like create() would do when requesting a full rebuild.
Not relevant for the OpenGL backend, while D3D and Metal may get away
without doing this, but the Vulkan backend gives visible rendering
errors in Qt Quick once updateResources() is taken into use and a
scene manages to do the "right" amount and types of changes.
The most common source is a changing uniform buffer offset. Consider
how the Qt Quick scenegraph merges uniform data into a single buffer
in unmerged batches (i.e. when a set of geometry nodes cannot be
batched together in a single draw all). While each node gets its own
draw call, each associated srb refers to the same uniform buffer at
binding point 0, just with a different offset.
Without the layout-compatible-update optimization (that is
QRhiShaderResourceBindings::updateResources()) this is not something
that needs extra care: once an srb is built or rebuilt (by calling
create()), the offset, just like the resource itself is fixed and
cannot change. And once create() is called, it conveniently
invalidates all related data structures, leading to the subsequent
setShaderResources() rewrite descriptors (incl. the resource, the
offset, etc.) with Vulkan or do whatever is relevant with other
backends.
updateResources() only does the minimum amount of changes because we
know that the binding list layout has not changed. It turns out this
was a bit too minimal, because certain state tracking data structures
still need resetting, just as if we called create().
The problem is becoming apparent with non-layout data such as the
uniform buffer offset, storage buffer offset, or the storage image mip
level. It however works as expected when changing the resource itself.
E.g. if a binding point now references a QRhiBuffer different than
before, then there is no visible problems, regardless of the buffer
offset being different or the same. Hence being difficult to discover,
until the aforementioned Qt Quick renderer case triggers it.
Why is this?
Changing the resource (buffer, texture, sampler) itself works due to
the guarantees given by the QRhi resource model. Consider the
following:
ubuf is a uniform buffer
ubuf->create();
srb->setBindings({ references ubuf });
srb->create();
// use the srb in some draw calls
// ...
// later, when preparing the next frame we decide we need new data with
// a different size even:
ubuf->setSize(new_size)
ubuf->create();
// fill in new data to ubuf
// use the srb in some draw calls
at this point "magic" happens: note how there is no rebuilding of the
srb itself (no create(), no nothing), yet it will correctly pick up
the now-different native buffer objects underneath ubuf.
This works because there is a certain degree of state tracking
happening that allows recognizing that a resource referenced from an
srb got rebuilt and now has different native objects (e.g. a VkBuffer)
underneath, which in turn needs (with Vulkan) rewriting the associated
descriptor with the new native resource.
Incidentally, this also makes updateResources() work just fine as long
as it replaces the QRhiBuffer/Texture/Sampler reference for the
binding point. However, with the example snippet above there is no way
to change the other associated data such as the buffer offset. (that
would need rebuilding the srb with create(), and that resets all
related state tracking structures)
So once we encounter an updateResources() where the same QRhiBuffer is
now used with an offset different from before, that is not recognized
by setShaderResources() and (with Vulkan) it will not rewrite the
descriptor with the new offset. (unless the changes for another
resource in the binding list trigger it; the granularity here is quite
coarse, i.e. we either rewrite (with Vulkan) all descriptors or none
at all; this makes the problem even less apparent because now
rendering errors occur only when Qt Quick ends with an update where
only the uniform buffer offset, but nothing else changed)
Pick-to: 6.2
Change-Id: I82ee43aa358947288135ff72ec213e091342e9cb
Reviewed-by: Andy Nichols <andy.nichols@qt.io>
Be idiomatic and return the output iterator one past the last element.
Otherwise passing in a plain pointer (as exercised by the autotest now)
fails to function because we write over the same 4 elements again and
again for each binding.
Pick-to: 6.2
Change-Id: If74463fa5140ffa2b1d5be97b71868848ad46614
Reviewed-by: Andy Nichols <andy.nichols@qt.io>
...by the Qt Quick renderer, for example.
A typical Qt Quick material binding set serializes to 8 uints. This
would not demand a container like QVector. However, being implicitly
shared is essential here due to the intended usage (query the
serialized blob, put it into a cache key, hash it, compare it, all
without any copying and new allocs; we can afford an extra alloc
upon each srb construction, but don't want more afterwards in the
rendering engines)
Also make it clear in the pipeline docs that the optimization Qt Quick
is (soon going to be) doing is legal. (the srb ref in the pipeline can
be dead and dangling as long as every call to setShaderResources()
specifies a layout-compatible alternative)
Pick-to: 6.2
Change-Id: I97efbea1fa3516b10c9832adbab0a21b7bc0845d
Reviewed-by: Andy Nichols <andy.nichols@qt.io>
Use the same variable name in function
Amends 6869d2463a2e0d71bd04dbc82f5d6ef4933dc510
Pick-to: 6.1 6.2
Change-Id: If710a53993e84d048f9052f4fcf92eb57635f585
Reviewed-by: Giuseppe D'Angelo <giuseppe.dangelo@kdab.com>
On OpenSUSE we seem to get an Enter event after or while the QWindow is
closed, and that appears to reset the currentMouseWindow after we
cleared it in QWindowPrivate::destroy. Apply a workaround similar to the
focus_window to make sure that currentMouseWindow and
currentMousePressWindow don't point to destroyed objects.
Pick-to: 6.2
Change-Id: I61afdaa7d8d1d437058624012319460be2f2567f
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
The thread checking in in_pixmap_thread_test() is only safe, and only
makes sense, if we have a QGuiApplication. If we only have
QCoreApplication, only null QPixmaps may be created anyway, so
threading makes no difference.
Fixes: QTBUG-95358
Pick-to: 6.2
Change-Id: I93c983a3d6c271b0a19dfd0384862fc151459029
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
Reviewed-by: Sze Howe Koh <szehowe.koh@gmail.com>
Until now, after updating the bindings one had to always rebuild the
srb, which can be heavy esp. on Vulkan (release old objects, create
new layout object, descriptor sets). When updating the binding list in
a way that it is fully isLayoutCompatible() == true with the previous
list, this is an overkill. Internally, most notably in
setShaderResources(), we already should have everything in place in
all backends to recognize if the entries in the binding list refer to
QRhiBuffer/Texture/Sampler objects that are different than before, and
so apart from adding an alternative to create() in the API there is
not much else needed here.
Pick-to: 6.2
Change-Id: I2efdd4fd0b24c7ebba694a975ed83509744b044b
Reviewed-by: Andy Nichols <andy.nichols@qt.io>
customFilters defined in .qdocconf are not supported anymore by Qt
Assistant since Qt 5.13. Therefore remove them from all .qdocconf files,
also to avoid cargo-culting them to new help modules.
Task-number: QTBUG-95987
Change-Id: I664391460637d2e859348da0338e1a4a3ee9f570
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Reviewed-by: Paul Wicking <paul.wicking@qt.io>
It is not true that isLayoutCompatible() can be called before create().
That used to be the case before the optimizations have been added.
The docs are still internal, but let's fix it up.
Pick-to: 6.2
Change-Id: Iee61848f058a06774550af6f38a3253956e4cfd3
Reviewed-by: Andy Nichols <andy.nichols@qt.io>
Relevant for Vulkan, given that descriptorCount is part of
VkDescriptorSetLayoutBinding, meaning two srbs with arrays of
SampledTextures should only be reported as compatible if the
array size matches.
Also reduces the prealloc size for the VLAs. For Qt Quick even a
lower number would be sufficient, but we still keep the number
something fairly high in order to play nice with Quick3D.
Pick-to: 6.2
Change-Id: Id92b7c09b051ebe54b1fa2bf4ba78950fe60ba27
Reviewed-by: Andy Nichols <andy.nichols@qt.io>
We want to close the window, end full screen mode on macOS, and free
platform resources. This is all done by QWindow::close. QWindow::close
closes the platform window, triggering a closeEvent to QWidgetWindow,
which then calls QWidgetPrivate::close_helper.
This way, closing a window via QWidget::close, QWindow::close, or
interactively by the user are all equivalent.
The QCloseEvent generated by the widget needs to be spontaneous for
window-system generated events (i.e. the user clicked the close button),
and non-spontaneous if the window closes because of a call to
QWindow::close. To keep track of whether the event originated in an
explicit call to QWindow::close, add a boolean to the QWindowPrivate.
Add a test case that verifies that the window resources is destroyed,
and that events are delivered as they should.
Done-with: Morten Johan Sørvig <morten.sorvig@qt.io>
Fixes: QTBUG-46701
Pick-to: 6.2
Change-Id: Iacb6a2c8d5e880b16b0c8f0c9257ed94bed36f5b
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
No need for timer event to reduce cache size when it is already empty.
May also avoid the "Timers cannot be stopped from another thread"
warning at exit, if the global cache object is then deleted by another
thread.
Fixes: QTBUG-96101
Pick-to: 6.2 5.15
Change-Id: Id1aeecfbb43a25a887ebd5cc7242749a74290bb0
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
We can't really compare two NaN's. Should use qIsNaN() for that.
Pick-to: 6.2
Change-Id: Ia514cabe65cfcdeafb39cab91ecdb66f8fae725c
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Replace the “size() / devicePixelRatio()” pattern with
a call to deviceIndependentSize().
Change-Id: I9d9359e80b9e6643e7395028cd43e3261d449ae7
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
The original code checks the existence of an ellipsis character first.If
this does not exist (glyph == 0,which is an invalid glyph index), then
it falls back to looking up '.'. But in the Tibetan environment,the
glyphIndex('.') also returns 0, so that it simply doesn't add any form
of "...", and cuts the text instead.
If both the attempts at getting something from the main font fails,
we can do a third pass on the "multi" font engine.
Fixes: QTBUG-95942
Pick-to: 6.1 6.2
Done-with: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@qt.io>
Change-Id: I251de3fe92e19be0462c58c2059ecf7d354bfbb0
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@qt.io>
Change it to take a QByteArrayView instead of a plain char *; all its
callers do know the size and propagating it enables the implementation
to call strntou?ll() rather than strtou?ll(), thereby escaping the
need for '\0'-termination.
Fixes: QTBUG-74286
Change-Id: Ie9394786e9fcf25c1d1be2421805f47c018d13bb
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Amends 53496e67f0b78645c6080e9218c7a36bc5a9d76d:
- move flags() out of the QT_DEPRECATED_SINCE block in the header
- add QT_DEPRECATED_SINCE block around source() implementation
Pick-to: 6.1 6.2
Change-Id: Id52fa6b04a13efbede3e6ac440060f90b283e773
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
Take the screen from the window, if the window was specified.
When it comes to Qt Quick, this is in fact required in order to
make the behavior on par with Qt 5, see
https://code.qt.io/cgit/qt/qtdeclarative.git/tree/src/quick/scenegraph/qsgthreadedrenderloop.cpp?h=5.15#n1336
try to mirror that in Qt 6 as well. There are still subtle differences
between Qt 5 and Qt 6 in particular when qt_gl_global_share_context is
set (do we set the screen from that or from the window?), for now leave
that question unsettled.
What exactly setting the screen on the context does is platform
specific, and can become relevant with multi screen, multi GPU
systems.
Pick-to: 6.2
Change-Id: Icc90b8fea87bf1e34ecf1dec0702f4d3c411db00
Reviewed-by: Christian Strømme <christian.stromme@qt.io>
Clean a bit, also enhance, e.g. mention textureLod for TexelFetch which
is how Quick3D uses it, and most notably expand the "in practice" notes.
As no application or library can be expected to write fully conditional
code based on all these flags, knowing when a certain feature is
(un)supported in a real run time environment is essential in order to
make informed decisions on when and in what manner (i.e. with or without
conditions), rely on a given feature.
Pick-to: 6.2
Change-Id: I15ea34e11bd345b36248f92de9b1fdb1fdc3e8d9
Reviewed-by: Christian Strømme <christian.stromme@qt.io>
We already honor AA_ShareOpenGLContexts and pass in the QGuiApp's
context as the shareContext for the QRhi's QOpenGLContext. Extend this
to also allow specifying a QOpenGLContext in the init params struct.
A good example of this is the backingstore compositor that serves
QQuickWidget and co. If one wanted to implement that with (an
OpenGL-based) QRhi, instead of direct OpenGL calls, then the ability
to create a QRhi that uses a context that shares resources with a
given other context becomes essential.
Pick-to: 6.2
Change-Id: I6bc5ff8e803d467f8795197ac1f12fdc0f73bbd1
Reviewed-by: Christian Strømme <christian.stromme@qt.io>
The mapping of the dash offset into the dash pattern was susceptible
to int overflow.
Pick-to: 6.2 6.1 5.15
Change-Id: I254ed40b323ef516d10abc8e6008b14cfc23d2a0
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
Since the dashing is computed even outside the clipping and device
area, painting very long dashed lines could consume unexpected amounts
of time and resources. Fix by placing a limit on the dashing, and fall
back to solid line drawing if hit.
Fixes: QTBUG-95594
Pick-to: 6.2 6.1 5.15
Change-Id: Ida05ecd8fe6df402c9e669206fd5cec4a9f5386a
Reviewed-by: Robert Löhning <robert.loehning@qt.io>
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
When we are unable to get the line thickness from a font, we
fall back to calculating this based on pixel size and weight.
But the font weight scale has changed in Qt 6, causing the
calculated font weight to be 10x as thick as it should.
This happened e.g. for Titillium Web, but only when using 100%
UI scaling, because the GDI engine was unable to get the correct
line thickness from the font, whereas the DirectWrite engine
succeeded.
[ChangeLog][Windows] Fixes an issue where underlines and other
decorations would be too thick for some fonts.
Pick-to: 6.1 6.2
Fixes: QTBUG-94529
Change-Id: Id2c88f0a145776151be0399ee646f28b3bdf9072
Reviewed-by: Thomas Hartmann <thomas.hartmann@qt.io>
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
When using NoFontMerging, no fallbacks should be resolved. If the
font does not support a specific character in the text, we should
display a box instead of merging it with another font.
But in practice, Qt would still apply the fallback mechanism for
one specific case: If the font itself does not support the script
of the text, we would get no match and do a search for a fallback
instead. Since NoFontMerging is set, we would then force this
as preresolved for *all* scripts in the QFont's private data
(logically, the match should only have a single response for
NoFontMerging).
The end result was that if you set the font family before updating
the text, you would get broken rendering. This can happen e.g. in
Qt Quick, where you could update the font family of a text label
while it contains characters which are not supported by the new
font. Qt would then pick a fallback instead. When you subsequently
update the text, the fallback would already be preresolved for
whatever script this is. If it does not support the updated text,
we would then see boxes, even if the requested font actually would
have supported it.
The fix is simply to do an additional pass if NoFontMerging is set
and we were not able to match with the specified script. Since
the same family might be available in different foundries, with
different writing system support, we still want to do a pass first
to see if we can match the exact script of the text.
Note that QRawFont::fromFont() exploited the bug by using
NoFontMerging for getting the fallback font for a specific
writing system. To keep this working without having to rewrite
fromFont() and risk introducing regressions, we add an argument
to make the findFont() function behave as before. It isn't
super-pretty, but since it is private API it is hopefully fine.
[ChangeLog][QtGui][Text] Fixed an issue with NoFontMerging and
changing font families dynamically, where boxes would be seen in
place of the correct text.
Pick-to: 5.15 6.1 6.2
Done-with: Andy Shaw
Fixes: QTBUG-81770
Change-Id: Ide9a36d7528a1040172c5864fa99e7a82eac4e83
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Konstantin Ritt <ritt.ks@gmail.com>
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
Size 0 is a "valid" answer for QIODevice implementations so we need to
make sure that we only enter the "try to workaround broken files" if we
know there is a size, otherwise the first read of length 4 that libpng
does breaks everything.
Change-Id: I1e396abd206ff90edae4372726f1d82d5d41ccf3
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
Like we do for OpenGL. Conveniently enough the QRhi feature flags are
readily available.
This should prevent errors such as:
MTLValidateFeatureSupport:3901: failed assertion `Base Vertex Instance
Drawing is not supported on this device'
on the iOS Simulator. It is not clear since which version or SDK this
became a fatal problem, but the base vertex/instance support is indeed
an optional feature according to the Metal Feature set tables, so not
calling the drawIndexedPrimitives variant taking baseVertex and
baseInstance when the reported iOS GPU family is too low is the right
thing to do regardless.
Pick-to: 6.2 6.1
Fixes: QTBUG-95795
Change-Id: I47c54a77a66a0410b86b8d4e5a1863dc730490f4
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
Reviewed-by: Andy Nichols <andy.nichols@qt.io>
This avoids rebuilding the same pattern. Caught by clazy.
Change-Id: Ibd0f2063617df1a9e975f58e34df556d1983afff
Reviewed-by: Richard Moe Gustavsen <richard.gustavsen@qt.io>
The documention currently incorrectly lists QNativeInterface::QX11Application
as available since 6.0.
Pick-to: 6.2
Change-Id: I13256a1504b2bd93296434438835791f12353089
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
On high-dpi displays or when you explicitly select an unhinted font,
we pick a DirectWrite font engine. This hit an uncovered code path
on Windows, because we relied on being able to get the HFONT from
the font engine.
To fix this, we introduce an alternative code path which gets the
HFONT based on the DirectWrite font when this font engine is active.
[ChangeLog][Windows] Fixed an issue where the characters in printed
text would look too small.
Pick-to: 6.1 6.2
Fixes: QTBUG-95720
Change-Id: Ifd609e92512e1f25f0ee2aace35cb5ccedf09030
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
When a key press comes in we may end up in QAppleKeyMapper::possibleKeys()
as part of checking whether the key press should trigger a QShortcut.
The function builds on QAppleKeyMapper::keyMapForKey(), which provides
a map from the given virtual key to all the possible Qt::Keys that can
be produced by applying different modifier key combinations.
The map is built using the Carbon function UCKeyTranslate, that takes
the current keyboard layout, virtual key, and modifiers, and produces
the resulting characters. The function also maintains a running dead
key state via one of the arguments. When mapping a dead key, the state
variable will be updated to the current dead key state, which then
affects the next call to the function (for the next key press).
The problem is that we're not calling UCKeyTranslate for each key press.
We are calling it in a loop, for a single key press, to build up a map
of all the possible characters produced by varying the modifier keys.
And in doing so, we are passing on the dead key state from one call
to the next, even if these are for different modifiers. The result is
that the first call, for the dead key, results in mapping to \0, as
UCKeyTranslate produces no output, it only modifies the dead key state.
And then the next call, for the next modifier key combination, results
in mapping to a character that incorrectly incorporates the dead key
state (resetting it in the process).
What we really want is to directly map the initial modifier combination
to the dead key terminator character, if one is defined. This is the
character produced if the dead key state is cancelled, for example by
pressing a key that's not defined in the dead key state.
To achieve this we pass kUCKeyTranslateNoDeadKeysMask as the translate
options to UCKeyTranslate, and always reset the dead key state before
every call. Another common way to achieve the same result would be to
call UCKeyTranslate a second time when detecting that the first call
produced a dead key state, for example with a synthetic space key, to
trigger the terminator output. But this can potentially fail if the
space key actually has a defined output in the dead key state.
Fixes: QTBUG-95471
Pick-to: 6.2 6.1
Change-Id: Icdae7639fd9a641a86c9d6615679bd93d380ff5c
Reviewed-by: Morten Johan Sørvig <morten.sorvig@qt.io>
When mapping virtual keys and modifiers to their corresponding characters
via a keyboard layout we may hit combinations that do not produce any
characters. This can happen if there's no <key> element defined for the
virtual key, or if the output attribute of the element is empty (despite
the spec saying there should always be one UTF-16 code point in the output).
https://developer.apple.com/library/archive/technotes/tn2056/_index.html
When that happens QAppleKeyMapper::keyMapForKey() will map the combination
to a null-QChar, resulting in the "key" being 0. We do not want to propagate
this back to the QShortcutMap machinery, as QShortcutMap does not validate
the keys coming out of QKeyMapper::possibleKeys(). In particular, it doesn't
check the isEmpty() or count() of the QKeySequences it creates. And even if
it did, QKeySequence itself seems to treat Qt::Key(0) + Qt::SomeModifier
as a non-empty sequence, so passing on 0-keys would still give weird bugs.
The user-visible result of passing back 0-keys is that QShortcutMap will
treat it as a partial match for any incoming key combination (as long as
some modifier is pressed that triggers the QShortcutMap machinery), which
resulting in eating the key press. This compounded the issue in QTBUG-95471.
Regression after fab3dfff7d.
Task-number: QTBUG-95471
Pick-to: 6.1 6.2
Change-Id: I2e51ec86f4df2a708e1757be827ab74859be3c8b
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
Reviewed-by: Morten Johan Sørvig <morten.sorvig@qt.io>
The "locked" flag was only added for a particular use in the Freetype
font engine in 070d9c00c488a5ee6811f04170cf488ead79bf80,but that usage
was refactored away long ago in afb326f071.
Pick-to: 6.2
Change-Id: I7347c91f49ba59041f6ff35bd5a6b537d6c704d4
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
The local/global positions for gestures should be handled the same as
for other events as they mean effectively the same thing.
Change-Id: Ic5ad995607ecd3daf385a7c7be3b67cbae312e7b
Reviewed-by: Morten Johan Sørvig <morten.sorvig@qt.io>
CreateDXGIFactory2() is always available from Windows 8.1
No need to load it dynamically.
Pick-to: 6.2
Task-number: QTBUG-84432
Change-Id: I84d82f30327df416e1fdbac256b63512900c07b7
Reviewed-by: Oliver Wolff <oliver.wolff@qt.io>
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Laszlo Agocs <laszlo.agocs@qt.io>
Previous fix hit too widely so some valid horizontal and vertical
lines were affected; the root problem being that such lines have an
empty control point rect (width or height is 0). Fix by caculating in
the pen width.
Pick-to: 6.2 6.1 5.15
Change-Id: I7a436e873f6d485028f6759d0e2c6456f07eebdc
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
The {mouse,touch}DoubleXDistance variables were initialized to -1,
which forces them into the DATA segments.
But this value is completely unused, since these variables are also
set in initThemeHints(), called from QGuiApplicationPrivate::init(),
before their only user, QGuiApplicationPrivate::processMouseEvent(),
can possibly execute.
By changing the default value to 0, then, we allow the compiler to
place these variables into the BSS segment.
Change-Id: I86144559ee2410b844f70a4f2499c8d3140d14cf
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Instead of appending to a QByteArray and then streaming that one, just
stream the components of a line directly. QTextStream's op<<(const
char*) is not subject to QT_NO_CAST_FROM_ASCII, etc., so can be used
unconditionally.
Change-Id: Idd97a75a1b5b939de7176d40880a2f328d01927d
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
The major use-case of the now private QX11Info from Qt X11 Extras was
getting hold of the Xlib display and XCB connection, for example in KDE:
https://lxr.kde.org/search?%21v=kf5-qt5&_filestring=&_string=QX11Info
A new native interface for QGuiApplication has now been added that
exposes these two properties, e.g.:
if (auto *x11App = app.nativeInterface<QX11Application>())
qDebug() << x11App->display() << x11App->connection();
To avoid type clashes one of the enum values of QXcbNativeInterface's
ResourceType had to be renamed.
Pick-to: 6.2
Task-number: QTBUG-93633
Change-Id: I2e366a2bb88bd3965ac6172ad000ae32209f43e7
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
The following code assumed, and all comments indicated, that the cpp
variable was limited to a max of 4. Yet, as coded, cpp could be five,
in which case the XPM would be corrupt, as the header suggested five
characters-per-pixel while the data was formatted in only four.
Add a warning and error out when we encounter this situation.
[ChangeLog][QtGui][QImage] Instead of writing a corrupt file, rejects
to write XPM files with more than 64^4 colors (more than four
characters per pixel) now.
Pick-to: 6.2 6.1 5.15 5.12
Change-Id: I458873cf7d179ab2e2dacd4c17dc837d640591a9
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
It's private API, but exported, so de-inline the dtor to pin the
vtable in QtCore instead of potentially duplicating it in every plugin
that uses the class.
Change-Id: Ia948985bb94677c8453a8e7be27a14085303aaf5
Reviewed-by: Giuseppe D'Angelo <giuseppe.dangelo@kdab.com>
Resize QPlatformWindow on DPI change, so that QWindow
size can stay approximately constant.
For example, a 100x100 QWindow at 100% scaling will
have a 100x100 QPlatformWindow. If the scaling is changed
to 200% then the QPlatformWindow is resized to 200x200,
while the size of the QWindow stays at at 100x100.
In practice the QWindow size will also change slightly,
due to inaccuracies in how we adjust for the size of the
non-client window area. This will be addressed in a later commit.
We can get DPI change independently of screen change,
so no resizing should happen in screen change events.
Disable the resize code in QGuiApplication for Q_OS_WIN,
and remove the WithinDpiChanged flag.
The new flow for handling DPI change is:
1) Send screen change (if any), so that the correct
screen will be used when calculating scale factors
during the following resize.
2) Resize the native window, which will trigger geometry
change events, possibly also for the QWindow.
3) Resize child windows; WM_DPICHANGED is sent to
top-level windows only.
Fixes: QTBUG-89294
Pick-to: 6.2
Change-Id: I0e2d44bae72d20ebdafc3d410db7be9964ad851b
Reviewed-by: Oliver Wolff <oliver.wolff@qt.io>
... instead of *(array + index).
Also fix a pointless cast from QRgb to int which the next line
implicitly undoes (because colorMap has QRgb as key, not int) and get
rid of the local variables that facilitated said fallacy in the first
place.
Change-Id: I71a92822ee2404646f6fb5533e40252f38e6b21f
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
... and fix spacing around operators and after flow-control keywords.
Change-Id: Iefaa03074536d13a655c91fb42aef6aa96c2665b
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
The code used the if (!contains()) { insert() } anti-pattern,
necessitated by Qt's deviation from the STL of allowing insert() to
overwrite an existing entry, causing two lookups of the same key.
Since QMap these days is a wrapper around std::map, fix by using the
real thing (a std::map) instead, which sports the non-broken insert()
semantics already, avoiding the need to play tricks like detecting a
size increase in order to find whether an insertion took place. It
also simplifies the loop later on, and we can transparently use pmr,
when available.
Change-Id: Iedd8d5691514a7705a55c27376446304b20af071
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
... and use one variable definition per line.
Change-Id: Ie8e6d6fb9e570cf715ab0b39c78d677f1e5a19a7
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
The xpm_color_name() function returned a pointer to a function-static
buffer. This is infamously non-reentrant, and an actual problem,
because we explicitly allow QImage operations (incl. saving to an
.xpm) from non-GUI-threads.
Fix by using the CSS pattern (Caller-Supplied Storage; also used in
the QAnyStringView(char32_t) and QAnyStringView(QStringBuilder) ctors)
to force the caller to allocate storage in its own stack frame. As a
consequence, we re-gain re-entrancy, but the returned pointer is now
only valid until the end of the full-expression, which necessitated
simplifying one caller (sorry!).
To see why said simplification is valid, observe that xpm_color_name()
writes a (now-explicit) NUL into returnable[cpp] and the old code read
max(cpp, 4) characters from xpm_color_name()'s result.
NB: cpp can be 5, even though the code comments say otherwise! :(
[ChangeLog][QtGui][QImage] Fixed a race condition when concurrently
writing .xpm files.
Pick-to: 6.2 6.1 5.15 5.12
Change-Id: I36d7173d53839a52f5cdf58324474c1b32c71f33
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
In order to filter out file names based on the user's settings,
QFileSystemModel used to have a loop that tested if a given file
name matched one of the filters. The problem is that each filter
(a wildcard) was converted to a QRegularExpression _inside_ the loop.
This causes a quadratic behavior (number of files * number of filters).
Instead, build the regexps once when the filters are set (or the case
sensitivity is changed, as that affects the filtering), and simply
_use_ them in the loop.
Simplify and correct some related code as a drive by.
Done-with: Jean-Michaël Celerier
Fixes: QTBUG-95383
Pick-to: 6.2 6.1 5.15
Change-Id: I6bc336364c145bb05793a8f867545d7715d35832
Reviewed-by: Samuel Gaist <samuel.gaist@idiap.ch>
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
One is a bad application or library in this case, but nonetheless
we should handle this more gracefully then just crashing due to
the QRhi already having been destroyed. Mainly because in Qt 5 one
could get away with the same: releasing OpenGL objects underneath,
for example, a QSGPlainTexture with no (or wrong) GL context did
not generate any user visible fatal errors. So we should not crash
in Qt 6 either with these code bases.
In debug builds or when QT_RHI_LEAK_CHECK is set, one will get the
unreleased resources warning printed in Qt 6, which is a step
forward compared to Qt 5. So there is still some indication that
something is badly designed, even if the application survives.
Task-number: QTBUG-95394
Pick-to: 6.2
Change-Id: I944f4f425ff126e7363a82aff926b280ccf1dfc3
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Andy Nichols <andy.nichols@qt.io>
It would have been nice to ensure that a device is registered already in
mouseEnteredImpl(); but in that context, NSEvent.deviceID is always 0,
and we can't find out anything else about the device.
QWindowSystemInterface::handleEnterEvent() doesn't currently take a
QPointingDevice either.
In handleMouseEvent() and scrollWheel(), deviceID seems unique for each
trackpad or Magic Mouse, but 0 for any plain USB mouse. There, the first
mouse that the user interacts with becomes primaryPointingDevice():
its deviceID is assigned to systemID (except if deviceID == 0, we use
1 instead, to avoid the auto-incrementing device ID assignment in the
QInputDevicePrivate ctor.) When scrolling occurs, we update the
capabilities to have PixelScroll if theEvent.hasPreciseScrollingDeltas.
So over time, QInputDevice::devices() should build up to a complete
list, with capabilities() also distinguishing plain mice from those that
have the PixelScroll capability. And in the common case that the user
has only one Apple pointing device, it becomes primaryPointingDevice().
Pick-to: 6.2
Task-number: QTBUG-46412
Task-number: QTBUG-63363
Task-number: QTBUG-72167
Change-Id: Id9771b4dfd765e49023bd57d42a2aa4d0635a3b2
Reviewed-by: Shawn Rutledge <shawn.rutledge@qt.io>
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
Apart from a more fitting, minimal, API, QDuplicateTracker also
transparently uses C++17 pmr::monotonic_buffer_resource to avoid, or
at least reduce, memory allocations.
Change-Id: I155f5518190c5f8f6d21fbec3fcecd6bcc7ff852
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
If Vulkan headers are present on the system when qtbase is configured,
QtGui and QtOpenGL should be compiled with Vulkan support.
If a user project uses a Qt built with Vulkan support, but their system
is missing Vulkan headers, the project configuration needs to succeed.
The project will get compilation errors if it uses Vulkan headers, but
that's intended.
This use case was broken when fixing Vulkan to be found when building
Qt for Android.
Fix the regression with a combination of things
1) Mark the WrapVulkanHeaders package as optional (already the case)
2) Use the include directories directly when compiling Gui and OpenGL
3) Propagate WrapVulkanHeaders::WrapVulkanHeaders link requirement to
consumers only if the target exists. It won't exist if Vulkan
include dirs are not found
This also requires some changes in pri and prl file generation.
For prl file generation, we don't want to link to the
WrapVulkanHeaders target, so we filter out all dependencies that
use TARGET_NAME_IF_EXISTS for anything that calls
__qt_internal_walk_libs which includes qt_collect_libs.
For pri files, we make sure to generate a uses=vulkan/nolink clause
by inspecting a new _qt_is_nolink_target property on the target.
We also don't add include dirs to the pri file if the new
_qt_skip_include_dir_for_pri property is set.
This is intended for Vulkan, because there is separate qmake logic to
try and find the include dirs when configuring a user project.
As a drive-by, fix nolink handling for WrapOpenSSLHeaders.
Amends bb25536a3d
Amends 7b9904849f
Pick-to: 6.2
Fixes: QTBUG-95391
Change-Id: I21e2f4be5c386f9e40033e4691f4786a91ba0e2d
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Alexey Edelev <alexey.edelev@qt.io>
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
When xdg-open is not found the KDE-specific kfmclient is considered.
That however is part of Konqueror and may not be present as well.
Plasma offers another option, kde-open5, which should be considered
before falling back to webbrowsers.
This is particularly for non-http URLs like tel: where opening in a
webbrowser is not wanted.
Pick-to: 6.2 6.1 5.15
Change-Id: I2b606562e21568fbe43f4593de67a1d467918cc4
Reviewed-by: David Faure <david.faure@kdab.com>
Before this change, next() was the only way to advance the iterator,
whether the caller was ultimately interested in just the filePath()
(good) or not (bad luck, had to call .fileInfo()).
Add a new function, nextFileInfo(), with returns fileInfo() instead.
Incidentally, the returned object has already been constructed as part
of advance()ing the iterator, so the new function is faster than
next() even if the result is ignored, because we're not calculating a
QString result the caller may not be interested in.
Use the new function around the code.
Fix a couple of cases of next(); fileInfo().filePath() (just use
next()'s return value) as a drive-by.
[ChangeLog][QtCore][QDirIterator] Added nextFileInfo(), which is like
next(), but returns fileInfo() instead of filePath().
Change-Id: I601220575961169b44139fc55b9eae6c3197afb4
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Some pathological cases were not caught by the previous fix.
Fixes: QTBUG-95239
Pick-to: 6.2 6.1 5.15
Change-Id: I0337ee3923ff93ccb36c4d7b810a9c0667354cc5
Reviewed-by: Robert Löhning <robert.loehning@qt.io>
Added missing #if-ery and deprecation macros to a QFont constructor
that was only documented as deprecated.
Fixes: QTBUG-94521
Fixes: QTBUG-95310
Pick-to: 6.2 6.1
Change-Id: I3d0418a3f7dca191a9068cc22627fe4deb7c53c5
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
All other equality comparisons except floating-point are value
comparisons, not interpretation comparisons.
[ChangeLog][QColor] Equality on HSL colors is now raw value
based like other color-formats, instead of being partially
interpretation based.
Pick-to: 6.2
Change-Id: I391464da172af6a2d69178b94203d36f7ce2c07d
Reviewed-by: Eirik Aavitsland <eirik.aavitsland@qt.io>
The fontFamilies and fontStyleName have an asymmetric getter and setter.
The setter takes QStringList, but the getter returns QVariant (?), for
no reason whatsoever (the underlying QFont APIs take QStringList,
and the docs say that they should return QStringList).
It's not entirely obvious that we can fix this without introducing
badly-named getters (getX) or somesuch, so for the time being, mark
this to be fixed in Qt 7.
Change-Id: I49706af00906a2c5d950ada4be3cbc76dbee432c
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Reviewed-by: David Faure <david.faure@kdab.com>
Fixes static analyzer warning 12b19393e18b2394a398806f633c6eee, and
amends a1a6e3d21b.
In the process, replace the "int& *= double" with correct integer
arithmetic that'll produce the intended result without going via
double.
Done-with: Edward Welbourne <edward.welbourne@qt.io>
Pick-to: 6.2
Task-number: QTBUG-8096
Change-Id: Ib2aa8ae46a1bfd4d121e61cf99141c0311502215
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Spotted during API review for 6.2
Pick-to: 6.2
Change-Id: Ib31020848ec6ffabc126c6984cc82547266d4a1d
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Declaration::lengthValue only supported 'px' sizes, but one can transform
any 'pt' value into 'px' by multiplying with 1.33.
Notes: this ignores display DPI, and instead follows the W3C definition
of 'pt' and 'px' as absolute lengths [1].
[1] https://www.w3.org/TR/css3-values/#absolute-lengths
1pt = 1/72th of 1 inch
1px = 1/96th of 1 inch
so the conversion is px = pt * (72/96).
Add unit test that verifies this using QPushButton's icon-sizes property,
also with changed font in preparation of adding support for 'em' and 'ex'
units in a follow up commit.
Task-number: QTBUG-8096
Pick-to: 6.2
Done-with: Cristian Maureira-Fredes <Cristian.Maureira-Fredes@qt.io>
Change-Id: I58782e7ad0e2ff9d89ed695f8a23b1e584cfed64
Reviewed-by: Andy Shaw <andy.shaw@qt.io>
The use of a freestanding function is not needed now that the name
doesn't alias the nativeInterface accessor function, and was just
adding complexity to the machinery.
People not familiar with the code will have an easier time following
the flow through the helper member function, and we no longer need
to declare our own export macros.
Pick-to: 6.2
Change-Id: I17530b7e89939cfc19ab8ffaa076b7129ae02dcf
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
The switch condition expression has already dealt width all the
conditions, so a `default` label is not necessary (it introduces a
warning). Delete it.
Fixes: QTBUG-95223
Change-Id: Iac3d9bb5d0b6fd8372d4dc16d920ce0a7b0511db
Reviewed-by: Giuseppe D'Angelo <giuseppe.dangelo@kdab.com>
Replace a manually-managed pointer + size with QVarLengthArray,
allowing use of pre-allocated space.
Change-Id: Ife3abea0b3b0f2577bb76c348513d52d46f79641
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Use unique_ptr to indicate ownership.
The QPainter dtor is now empty, = default it.
As a drive-by, fix a non-sensical if (p) p = nullptr;
Change-Id: I5c811eb7de6abf8a9f279cf3630d5f86db6b3864
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
We keep QVector2D storage Qt 6 BC (to avoid making QNativeGestureEvent
larger), but in Qt 7 we should return exactly the same value as given
(for what it's worth, in spite of this being overkill for panning a
reasonable distance).
Change-Id: Iecbd4c9b60ad9ae5e0466c7027b038ddb85b8c8b
Pick-to: 6.2
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
By switching out the static_assert for an enable_if we end up producing
a clearer error, at the call site:
/qt/qtbase/examples/gui/rasterwindow/main.cpp:69:9: error: no matching member
function for call to 'nativeInterface'
app.nativeInterface<QNativeInterface::QCocoaGLContext>();
~~~~^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
/qt/qtbase/src/gui/kernel/qguiapplication.h:176:5: note:
candidate template ignored: requirement
'NativeInterface<QNativeInterface::QCocoaGLContext>::isCompatibleWith<QGuiApplication>'
was not satisfied [with NativeInterface = QNativeInterface::QCocoaGLContext, TypeInfo =
QNativeInterface::Private::NativeInterface<QNativeInterface::QCocoaGLContext>, BaseType =
QGuiApplication]
QT_DECLARE_NATIVE_INTERFACE_ACCESSOR(QGuiApplication)
^
By using SFINAE for the TypeInfo we can also ensure that it works for
types that are not native interfaces, such as if the user tries to
call nativeInterface<QString>().
Since we can no longer use decltype(*this) to resolve the base type
we need to change QT_DECLARE_NATIVE_INTERFACE_ACCESSOR to take the
type as an argument, as we do for other QT_DECLARE_FOO macros.
Pick-to: 6.2
Change-Id: Ie3f7e01ab7c3eb3dcc2ef730834f268bb9e81e0c
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
Use unique_ptr to indicate ownership, even though it's conditional (on
QPaintEngine::autoDestruct()). That just requires a custom deleter.
Change-Id: Icf8e356c333f9617b2e5172b14f13197e63c9502
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
In QPanGesture this is called delta().
OTOH we have QWheelEvent::pixelDeltas().
Delta is a vector, and there's only one (with two components).
Native gestures hold incremental values: e.g. the pinch gesture event
provides an incremental amount of either zooming or rotation (so most
events have QNativeGestureEvent::value() very close to 0).
It's the same with the pan gesture's delta().
Add better docs for swipe and pan gestures.
Change-Id: Ia147c7c9a22e084c3700b1620dec46427d792bd1
Reviewed-by: Povilas Kanapickas <povilas@radix.lt>
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
It came up during 6.2 API review that we prefer all floating-point API
to be double-precision on 64-bit platforms, despite the awkwardness of
representing a displacement vector with something called a "point".
The docs for QPointF explicitly state "A QPointF object can also be used
as a vector: Addition and subtraction are defined..."
Amends 31f90e99b8
Change-Id: I01029661f2586640cbf846f49df164c176d17f7a
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
When saving to a QIODevice, QImage and QImageWriter will automatically
deduct the file format from the filename if it determines that the
device is a QFile. That did not work for a QSaveFile device. Fix by
using the common ancestor, QFileDevice, in the implementation.
Fixes: QTBUG-89022
Pick-to: 6.2
Change-Id: Ie01d80df4f29ca0d4ff30bf7e1b77605293c070e
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
This prevents us from first reserve()ing Prealloc elements, and then
possibly reserve()ing a larger number, which leaves the first bucket
list's memory unused.
Consequently, deprecate reserve().
Change-Id: Ifc0a5a021097f4589557e7b5e45d9d0892797ade
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
They're now in QWidget itself. Remove them from the API, but not the
ABI.
The QToolBar case is straight-forward. QMenu is a bit more complicated:
Since QT_CONFIG(shortcut) builds changed the signature of an existing
function instead of adding/removing an overload, we have to deal with
two cases: In a QT_CONFIG(shortcut) build, these overloads that take a
trailing QKeySequence parameter have been deprecated and therefore
cannot be removed. In a !QT_CONFIG(shortcut) build, the same functions
are 1:1 copies of QWidget functions and can be removed (from the API).
[ChangeLog][QtWidgets][QMenu/QToolBar] The addAction() functions have
been moved down into QWidget.
Change-Id: I49997b3440c137a1d4e3858d1d27d34a191e1eed
Reviewed-by: Richard Moe Gustavsen <richard.gustavsen@qt.io>
Use unique_ptr to indicate ownership. Use std::stack to enforce
stack-ness (QStack doesn't). Requires the addition of a clear()
function, for which we spin a SmallStack wrapper. It also helps to not
have to repeat the value_type...
Because we use unique_ptr, `state` and `states` are now exclusive
(before, `state` was always `states.back()`). Correspondingly, rename
`states` → `savedStates`.
As a drive-by, this fixes an off-by-one error in a qWarning() (checked
`> 1`, but reported one saved state too many).
Change-Id: I8faade59300401be802ddc52c64ce31b8f23dc52
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
... instead of raw pointers or QSharedPointer.
Raw pointers are, of course, a no-no in modern code. In particular,
when the result is then held in shared_ptr or QSharedPointer,
make_shared or QSharedPointer::create() should be used to reduce
number of memory allocations.
Since this is private API, we're free to use std::shared_ptr, which
does only half the atomic operations on copies, compared to
QSharedPointer, so is more efficient.
For either make_shared or QSharedPointer::create(), we need to work
around the private ctor, which we do by inheriting a member-function
local class from QColorTrcLut and make_shared'ing that. As a
member-function-local class, it has access to the otherwise private
parts of QColorTrcLut, including its default constructor. As a public
subclass, shared_ptr has no problem performing the derived-to-base
pointer adjustment in the return statement. This way, we can use
make_shared even though our target's class' ctor is private.
Change-Id: Icb11249b54cd5e544e692f6a0bf1f9dda1710454
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
...lest some code tries to set it to no possible effect.
Change-Id: I7aed05200c1abeda1d2f9d88ceb99a5ce6132dbf
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
Using the QTextFormat::FullWidthSelection property to select a line
would previously not take into account right-to-left text layouts.
With this patch, the whole line should now be drawn correctly for both
left-to-right, and right-to-left layouts.
Pick-to: 6.2 6.1 5.15
Fixes: QTBUG-91125
Change-Id: Ide7b340cd7b4060e7c00e55e0011a86ffdfb5eb4
Done-with: Volker Hilsheimer <volker.hilsheimer@qt.io>
Reviewed-by: Dmitry Shachnev <mitya57@gmail.com>
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
None of the member functions use sharedFromThis().
Change-Id: I4d1fc0c3c18830df231833ce0ef631e1bb1fc27e
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
Amends c0b3c06a7d.
The QCompleter used by QFileDialog uses starts-with matching of
the path parts, for all parts of the path. It uses the EditRole of the
underlying model for the data to match.
The path parts used for a path on Windows are "C:", "Users", "etc...",
and then finally whatever input has been typed so far. If we try to
match that against a starts-with rule against "Local Disk (C:)", then it
will fail and the completer will never have anything to show.
So, for the EditRole, return just the volume drive letter as we did before
c0b3c06a7d, not the volume name. Neither
can be edited anyway. This happens to then match the native file dialog
behavior, which doesn't complete "L" to "Local Disk (C:)" if the Computer
contents (ie. volumes) are shown.
Augment the QCompleter test case to cover that scenario (it already has
everything set up for this particular combination, even though we are
actually testing QFileSystemModel returning the correct data).
Fixes: QTBUG-94799
Task-number: QTBUG-78043
Pick-to: 5.15 6.2 6.1
Change-Id: I14844d91601e9b16dc07cff2ca48713acb7cdd09
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
In Qt 6, QHoverEvent is a QSinglePointEvent and carries more information
than it did in Qt 5.
Task-number: QTBUG-94971
Change-Id: I55b271e8741081ed9074f687b08f4111142a1bf0
Pick-to: 6.2
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
Introduce two new packages WrapVulkanHeaders and WrapVulkan similar to
the OpenSSL wrapper packages.
WrapVulkanHeaders uses FindVulkan and is marked as found if Vulkan
headers are found (that's the only part the Qt build requires).
The WrapVulkan package is currently not used, but is there for
symmetry.
The Vulkan feature is now disabled by default on QNX, because the
QNX toolchain file in the CI does not set
CMAKE_FIND_ROOT_PATH_MODE-like variables and CMake ends up finding
host Vulkan headers causing the build to break.
Pick-to: 6.2
Fixes: QTBUG-92157
Change-Id: I05309821f866456cd42e7f85bf8b76ba099df656
Reviewed-by: Alexey Edelev <alexey.edelev@qt.io>
Reviewed-by: Assam Boudjelthia <assam.boudjelthia@qt.io>
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
... by making the test class a friend of the CUT, as we do elsewhere
for the same reason.
This allows to remove the duplicated enum and struct in favor of using
The Real Thing™, which means the test can no longer go out of sync
with the CUT anymore.
Change-Id: I87dc8bb4a5476ae4fc99e006c4690e96d2f530d2
Reviewed-by: Giuseppe D'Angelo <giuseppe.dangelo@kdab.com>
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
This is a partial revert of commit
f385b8827a.
This causes issues because we may overwrite genuine styles with
synthetic ones.
Lets say for instance that we register "Roboto Bold" and then later
we register "Roboto Thin". When we register "Roboto Thin" we also
register an alternative font which is called "Roboto" (because this
is the typographical family name of the font) with bold weight, because
we know Windows can synthesize this.
This would work fine, except that on Windows we also store the original
face name of the font as a user-pointer in the database. This contains
the legacy name of the font: "Roboto Thin". This will override the
font that is already stored. When we look up "Roboto" + bold weight in
the database later, we will find this synthetic font, replace
the requested family name with the legacy one "Roboto Thin" and use
this instead.
The right fix for now is to revert the cause of the regression. If we
want to re-fix the original bug, we might be able to reintroduce
f385b8827a and then make sure we always
prefer the "real" font when there are conflicts (this would mean marking
synthetic fonts in the database).
[ChangeLog][Windows] Fixed a regression where different font styles
and/or weights would not be available.
Fixes: QTBUG-94781
Task-number: QTBUG-91398
Pick-to: 5.15 6.1 6.2
Change-Id: I092022b14ebf1d56685eaa3b8efe55f015659adc
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
The FontConfig database had a static leak because we never
dereferenced the current config. To make address sanitizer output
less noisy, we clean this up on shutdown.
From FcConfigDestroy docs: "Note that calling this function with the
return from FcConfigGetCurrent will cause a new configuration to be
created for use as current configuration."
So this should be safe even if application execution continues after
Qt shutdown, but it could trigger creation of a new current config
in that case, if FontConfig calls are made.
Fixes: QTBUG-92477
Pick-to: 5.15 6.1 6.2
Change-Id: I596055a84edc1a1b06157e2adf6c8627c6802db1
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
Using QSystemTrayIcon::isSystemTrayAvailable() generates a new
connection from QDBusTrayIcon::isSystemTrayAvailable() with a new unique instance ID.
These were never closed, so calling
QSystemTrayIcon::isSystemTrayAvailable() frequently leads to crash when
it runs out of file descriptors.
Fixes: QTBUG-94839
Pick-to: 5.15 6.1 6.2
Change-Id: Ib71441a6b680d8633707cc02f9b6081c0f02472b
Reviewed-by: Dmitry Shachnev <mitya57@gmail.com>
Reviewed-by: Shawn Rutledge <shawn.rutledge@qt.io>
It could fit into QVariant's inline storage if we didn't waste so much
space.
Change-Id: Ie9e592a49c955b0a97a202e0bd875183396af993
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
Use autogeneratedd cpp exports in the Gui module.
Task-number: QTBUG-90492
Change-Id: Ifc5d3dd138ef9f86f6003ed9adb66a9eefdd64db
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
Without the workaround, and when using a Core GL profile, we hit a code
path in QSGOpenGLDistanceFieldGlyphCache::resizeTexture() that produced
corrupt glyphs on M1 hardware.
We fix this by enabling the workaround, so that the user doesn't have to
set QT_ENABLE_GLYPH_CACHE_WORKAROUND themselves.
None-Core contexts do not have this problem, but the logic in
QOpenGLContext does not account for recreated contexts with different
formats, so we can't limit the workaround to Core formats only.
With the unified memory architecture of the M1 hardware, the workaround
should have limited negative effects.
In Qt 6 this is not a problem, since Qt Declarative effectively always
uses the workaround code-path, but it's worth recording the fact that
we need the workaround.
Fixes: QTBUG-89379
Pick-to: 6.2 6.1 5.15 5.12
Change-Id: Icfd8b8b23c0dcda3fea8663d81d0e225134eec5e
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@qt.io>
Reviewed-by: Laszlo Agocs <laszlo.agocs@qt.io>
When QPainterPath was used for RTL text, either directly or because
the target text size exceeds 64 pixels, we would pass true for
"isRightToLeft" to DirectWrite, causing it to do adaptation internally
for this. But the RTL layout had already been handled by Qt, so we
would essentially reverse the layout twice and also move the text
to negative X coordinates.
Passing false instead fixes this, as it will then just use the
positions we pass in blindly.
Fixes: QTBUG-94175
Pick-to: 5.15 6.1 6.2
Change-Id: Ie9a47e56e97fc867ede10ab21ac5e3f003ddcacb
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
Remove the configure.json and configure.pri files that
were used for the qmake-based configure.
Remove the .prev_*.cmake files that were a by-product of
configurejson2cmake.py.
Pick-to: 6.2
Task-number: QTBUG-89536
Change-Id: Ie827562f7fd2513d59f69234d77b8b93124ea78e
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
Since memReq is initialized inside the loop, codechecker
complained that we might use uninitialized data if count == 0.
So we assert that this is not the case to get rid of the
warning.
Change-Id: I82152942d2bb61723e9e41489c87bfde9f8bec15
Reviewed-by: Laszlo Agocs <laszlo.agocs@qt.io>
msvc2019 emits:
warning C4244: 'argument': conversion from 'qsizetype' to 'int',
possible loss of data
Change-Id: I286c0d453f69654ba58f071aee28e557b9f9058d
Pick-to: 6.1 6.2
Reviewed-by: Jan Arve Sæther <jan-arve.saether@qt.io>
A misunderstanding seems to have caused several files to be compiled
with extra runtime arch options.
We need to disable multi-target functions though, because it doesn't
work with clang-cl headers.
Task-number: QTBUG-88081
Task-number: QTBUG-88434
Pick-to: 6.2
Change-Id: Ic5829540f9007428c80f9a7bd46af35540745ccd
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
It's only Qt::MiddleButton in Qt 6.
Pick-to: 6.1 6.2
Change-Id: Ia68bad910c617993e30e3ed1e117192469ec50eb
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Fixes build issue "no instance of overloaded function "qHash" matches
the argument list" on INTEGRITY
Pick-to: 6.2
Change-Id: Ia1273587840d55199846dc64d487d194f9a4d565
Reviewed-by: Laszlo Agocs <laszlo.agocs@qt.io>
Move QT_END_NAMESPACE macro out of the #ifdef definition
Pick-to: 6.2 6.1 5.15
Change-Id: I26b4e263b5ae0acebf035dbfe8c7e287cd740190
Reviewed-by: Kai Koehne <kai.koehne@qt.io>
Remove a conflicting definition of the same operator from
qrhi.cpp. Who knows how many ODR violations this hidden gem may have
caused...
[ChangeLog][QtCore][QDebug] Can now stream QVarLengthArray objects.
[ChangeLog][QtCore][QVarLengthArray] Can now be output via QDebug
streaming.
Change-Id: I12f7670e32029fa8c96b446e9a98ebf3b9732d0d
Reviewed-by: Giuseppe D'Angelo <giuseppe.dangelo@kdab.com>
It's deprecated. Port some unnamed enumerations (used only to
declare constants) to constexpr integers instead. Apply qToUnderlying
as needed.
Task-number: QTBUG-94059
Change-Id: Ifaa64ece966ce08df40dc71ffcfa7ac038110e0b
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
We need those events to trigger palette color group changes in QQuickItem
without having to connect every item to yet another QWindow signal.
Task-number: QTBUG-93752
Pick-to: 6.2
Change-Id: I8534808cdaab828e5876f8fda31567aeb1b4272a
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
The decoding of PNG_INFO_gAMA to QColorSpace was incorrect, the PNG
gamma is the inverse of the gamma value we use. We revert it
everywhere else, just not here.
Pick-to: 6.2 5.15
Change-Id: Ic0ae1963b2dde3004cac8a6430ddaf99e7096915
Reviewed-by: Eirik Aavitsland <eirik.aavitsland@qt.io>
Handle the case were a format was applied that does not align with
the fragments.
Pick-to: 6.2 6.1 5.15
Change-Id: I27593a50c587a5ec0290ff9d0a7a9767a3981cac
Reviewed-by: Konstantin Ritt <ritt.ks@gmail.com>
We currently only support full gamma corrected painting by using linear
color space images, the text logic however did its own gamma correction
on top of that.
Change-Id: I4ef41745f4bc1c3a5815426abbcee6968987e70d
Reviewed-by: Eirik Aavitsland <eirik.aavitsland@qt.io>
The lack of an unsigned pack 32bit->16bit without SSE4, meant
the premultiplied path would saturate color values over 32735.
Pick-to: 6.2
Change-Id: Ia76e0369fd11a6767e8db78bf9a10eea30d91d2c
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
QIntValidator::validate documents correctly that any input with "at most
as many digits as the top of the range" returns Intermediate. This is
needed to avoid input deadlocks where one can't go from 9 to 15 if the
range is 8 to 16.
Fixes: QTBUG-94269
Pick-to: 6.1 5.15
Change-Id: I6776e09fc231249fe78f9e6106492f8454b70a03
Reviewed-by: Andy Shaw <andy.shaw@qt.io>
Mainly to avoid leaking memory again. A possible reduction of memory
allocations would be a welcome side effect.
This reverts parts of 89f7389494.
Change-Id: I65a7529f532175967a4e408450aa55549b77d7e4
Reviewed-by: Laszlo Agocs <laszlo.agocs@qt.io>
This should have happened when changing the code in
f493d41722
Fixes oss-fuzz issue 31633.
Pick-to: 5.15 6.1
Change-Id: I7945c3f90651f8fa04df20fbaf0b0c7f68619407
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
Following the discussion in QTBUG-69452, the resulting change in
cc33dd0797, and the documentation of
QAction::shortcutVisibleInContextMenu, the intention is that the
attribute allows the overriding of the platform default.
However, QAction did ask both the attribute and the platform
integration, making the override impossible. Instead, ask only
the attribute, but default the value of the attribute to what the
platform integration provides.
[ChangeLog][QtGui][QAction] The shortcutVisibleInContextMenu property
defaults to the value of the Qt::AA_DontShowShortcutsInContextMenus
attribute, which in turn defaults to the platform integration. To
override the default, set the application attribute after
instantiating QApplication, or override the default for each
QAction instance.
Task-number: QTBUG-73990
Pick-to: 5.15 6.1
Change-Id: Iaba330913555d93d05efe1b3965a6aea39db5554
Reviewed-by: Richard Moe Gustavsen <richard.gustavsen@qt.io>
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
qmake had support for building with clang-cl as the win32-clang-msvc mkspec.
Task-number: QTBUG-89642
Task-number: QTBUG-88081
Change-Id: I0709c289f90fedb121620d1e67ef841602219816
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
This reverts commit 0336f08fec.
Reason for revert: This breaks tst_qquickpinchhandler;
We no longer seem to construct a QNativeGestureEvent of type
EndNativeGesture with it.
Change-Id: I942fa099d5c5a1c757accde0c67e407809ce2d72
Reviewed-by: Shawn Rutledge <shawn.rutledge@qt.io>
Add FindWrapSystemMd4c.cmake so that the old md4c target can be used
as well as the new one and set WrapSystemMd4c_FOUND.
Link to the imported target WrapSystemMd4c::WrapSystemMd4c if the
system library is used.
Add qt_find_package line to find the package in configure.cmake.
Fix the condition for enabling system-textmarkdownreader, it includes
testing for textmarkdownreader because even if the code would compile
correctly without it, it looks strange when the output says
"textmarkdownreader no" and under "using system libmd4c yes" even if
libmd4c is not used.
Use system include when system-markdownreader is enabled.
Add library mapping for libmd4c.
Change-Id: Id5d5b13d6691a8c1cdf627238887977c847c1e67
Reviewed-by: Shawn Rutledge <shawn.rutledge@qt.io>
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
This is needed in order to support hyperlinks in text content.
Task-number: QTBUG-67878
Change-Id: I1e990a5db8f0cf78e5cdcec7359e5aabffe42e3d
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
This way we can use it from qtdeclarative to parse styled text
Change-Id: Ic888a75a9700558e97b3e743d6d42fda121ddcba
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
The QSinglePointEvent is mostly responsible for setting
QEventPoint::state(); but in this case we need to decide based on
Qt::NativeGestureType, so it needs to be refined in the
QNativeGestureEvent constructor.
Fixes: QTBUG-94178
Change-Id: I9799fe5b8fea71f934311ae2f3bb8e033d132ec5
Reviewed-by: Povilas Kanapickas <povilas@radix.lt>
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
Handy when one has the data buffer to be read in a QBAV.
This also fixes an issue and compiler warning about passing a
qsizetype data length value as an int, and makes it possible to pass
a qsizetype-size length without going through QByteArray.
Makes the QByteArray overload redundant.
Change-Id: Iba8825cf0fd8003fb2eac5b1d30a61ec91b85ceb
Reviewed-by: Giuseppe D'Angelo <giuseppe.dangelo@kdab.com>
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
In similar way with vsp2, the interfaces are needed to support
qtwayland hardware layer for webOS.
https://doc.qt.io/qt-5/qml-qtwayland-compositor-waylandhardwarelayer.html
Change-Id: I14481373d696b501a774b9258da789554065a6ea
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
Reviewed-by: Elvis Lee <kwangwoong.lee@lge.com>
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@qt.io>
Triggers -Wunused-function warning (a.k.a. error) when building this
part of the code:
qpixellayout.cpp:292:25: error: unused function 'fetchPixel' [-Werror,-Wunused-function]
inline uint QT_FASTCALL fetchPixel<QPixelLayout::BPP64>(const uchar *src, int index)
^
qpixellayout.cpp:300:25: error: unused function 'fetchPixel' [-Werror,-Wunused-function]
inline uint QT_FASTCALL fetchPixel<QPixelLayout::BPP16FPx4>(const uchar *src, int index)
^
qpixellayout.cpp:308:25: error: unused function 'fetchPixel' [-Werror,-Wunused-function]
inline uint QT_FASTCALL fetchPixel<QPixelLayout::BPP32FPx4>(const uchar *src, int index)
Change-Id: I8ff40e3cf02f8a7ad7534f58ac23eeb8eb716d3a
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
When rotating M_PI_2 based on x-axis, quaternion to euler conversion
makes NaN for the x-rotation value. This patch fixes this corner case.
Fixes: QTBUG-93600
Pick-to: 6.1 6.0 5.15
Change-Id: Ice321a80ad90dba9cf3ee3a14ec7d3d047c21bd3
Reviewed-by: Laszlo Agocs <laszlo.agocs@qt.io>
Reviewed-by: Andy Nichols <andy.nichols@qt.io>
Followup to 31f90e99b8.
The number of fingers involved in a gesture will likely become a key
feature to make different behaviors distinct.
Change-Id: Ib41b2fa7ab4ea7d008a6479c018b3d475a62a8fd
Reviewed-by: Povilas Kanapickas <povilas@radix.lt>
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
Supported on OpenGL (and ES) 3.0+ and everywhere else.
Can also be a render target, targeting a single slice at a time.
Can be mipmapped, cannot be multisample.
Reading back a given slice from a 3D texture is left as a future
exercise, for now it is documented to be not supported.
Upload is going to be limited to one slice in one upload entry,
just like we specify one face or one miplevel for cubemap and
mipmapped textures.
This also involves some welcome hardening of how texture subresources
are described internally: as we no longer can count on a layer index
between 0..5 (as is the case with cubemaps), simply arrays with
MAX_LAYER==6 are no longer sufficient. Switch to sufficiently dynamic
data structures where applicable.
On Vulkan rendering to a slice needs Vulkan 1.1 (and 1.1 enabled on the
VkInstance).
Task-number: QTBUG-89703
Change-Id: Ide6c20124ec9201d94ffc339dd479cd1ece777b0
Reviewed-by: Andy Nichols <andy.nichols@qt.io>
This amends e2bdff3555.
The linearAdvance property has some history. First it was
a 16 bit value (allowing for 10.6 fixed point numbers). Then
it was turned into 22 bits to fit the 16 bits of Freetype
integer parts into it (16.6). Then back to 10.6. Then in
b7e4367387 it was turned
back into 16.6 again. But this was accidentally reverted
as part of a bad conflict resolution in
afb326f071.
Since there was no check for it, we would sometimes overflow
the linearAdvance, but only in the rare cases where the
width and height did not also overflow. Specifically this
is the case for whitespace, which always has a width of 0
regardless of the advance.
This change just moves the linearAdvance condition in
together with the other checks to avoid fragmentation, and
also adds this fun story to the commit log.
Pick-to: 6.1 5.15
Change-Id: Iaac09942f4c50d1aced4a160b6eabb11eb8e6373
Reviewed-by: Paul Olav Tvete <paul.tvete@qt.io>
Add NEON for RGB32 and RGBA64 writeback, and SSE2 for
RGBA64 writeback.
Change-Id: Id9ee803267a78f5bdff5beaa719e7a59c1dbb9fb
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
And avoid some needless shuffling.
Change-Id: I5101eb0582c17f5e3cb6db8e20d315ef38bd1863
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
Useful for some HDR representations and HDR rendering.
Change-Id: If6e8a661faa3d2afdf17b6ed4d8ff5c5b2aeb30e
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
Super/Hyper keys are detected during a direct mapping phase, but the
function returned before the translation to Meta could take place.
Task-number: QTBUG-62102
Pick-to: 5.15 6.0 6.1
Change-Id: I9f7ccfd757fa86dbc648951306deb1b43ccf4167
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
Reviewed-by: Andrey Butirsky <butirsky@gmail.com>
Upper-casing a Latin1 character might move it out of Latin1 range,
for example U+00B5 MICRO SIGN, which upper-case equivalent is
U+039C GREEK CAPITAL LETTER MU. If that's the case, then map the
original lower-case character.
Fixes: QTBUG-93858
Pick-to: 5.15 6.0 6.1
Change-Id: Ibd86fbb7c70d620c0884015ec1c40aa6351e6320
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
Reviewed-by: Andrey Butirsky <butirsky@gmail.com>
The list persists events, which reference objects. Those objects might
get deleted by the time we investigate the objects. We cannot change the
event to store a QPointer for BIC reasons, and for normal usage of the
events, that doesn't make sense either.
Instead, connect the objects destroyed signal to a lambda which clears
the events' object member.
In order to access the private member, we befriend the test class.
Change-Id: I036be7053dccde4bdf862173789564e89d729ee1
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Reviewed-by: Shawn Rutledge <shawn.rutledge@qt.io>
The existing conversions weren't handling gamma correctly and used
an ad-hoc definition of gray instead of based on true luminance.
[ChangeLog][QtGui] RGB conversions to grayscale formats are now
gamma-corrected and produce color-space luminance values
Change-Id: I88ab870c8f5e502ddb053e6a14a75102239a26f2
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
The small caps version of a font is kept as a QFontPrivate*
which is manually reference counted, but we neglected to
actually delete it when the reference count went to 0.
[ChangeLog][Fonts] Fixed a memory leak when initializing
a small caps font.
Fixes: QTBUG-93068
Pick-to: 6.1 5.15
Change-Id: Icc7fb7a59bf523da84d2e6fa026940a7d1230525
Reviewed-by: Konstantin Ritt <ritt.ks@gmail.com>
Reviewed-by: Robert Löhning <robert.loehning@qt.io>
The latter forces users to build with RTTI enabled, as the typeid
use is in our public headers. Surprisingly this is also the case
even without instantiating the relevant template.
Change-Id: Icd18a2b85b250e0b77960797e5c43b7eaf9bd891
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
In QFileSystemModel, in some cases the hostname in a UNC path is
converted to lower case and stored in the root node's visibleChildren.
When QFileSystemModel sets the UNC path as the root path, it tries to
get the row number for the host, but it didn't convert the hostname to
lower case before getting the row number, which resulted in the host
not found in the root node's visible children. As a result, it returns
-1, an invalid row number. Change the behavior to find the node for the
host using the host name case-insensitive and then get the row number.
Fixes: QTBUG-71701
Pick-to: 5.15 6.0 6.1
Change-Id: Ib95c7b6d2bc22fd82f2789b7004b6fc82dfcb13b
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Alex Blasche <alexander.blasche@qt.io>
Normally we only allow creating wrappers for texture objects. These
can then be used with a QRhiTextureRenderTarget to allow rendering into
an externally created texture.
With OpenGL (ES), there are additional, special cases, especially on
embedded. Consider EGLImages for example. An EGLImageKHR can be bound to
a renderbuffer object (glEGLImageTargetRenderbufferStorageOES), which
can then be associated with a framebuffer object to allow rendering into
the external buffer represented by the EGLImage. To implement the same
via QRhi one needs a way to create a wrapping QRhiRenderBuffer for the
native OpenGL renderbuffer object.
Here we add a createFrom() to QRhiRenderBuffer, while providing a dummy,
default implementation. The only real implementation is in the OpenGL
backend, which simply takes a renderbuffer id, without taking ownership.
Task-number: QTBUG-92116
Change-Id: I4e68e665fb35a7d7803b7780db901c8bed5740e2
Reviewed-by: Andy Nichols <andy.nichols@qt.io>
Change-Id: I53eaaea149324d2495e794ba8bd58544e648e48e
Reviewed-by: Janne Koskinen <janne.p.koskinen@qt.io>
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
This enables OpenWFD EGLFS integration plugin and makes EGLFS the
default platform for INTEGRITY builds.
Change-Id: I65332ca0ae244f40013df435828e2e359200b325
Reviewed-by: Janne Koskinen <janne.p.koskinen@qt.io>
Reviewed-by: Laszlo Agocs <laszlo.agocs@qt.io>
The BASE argument of qt_add_resources now denotes the root point of the
alias of the file. Before, BASE was merely prepended to every file that
got passed to qt_add_resources.
Old behavior:
qt_add_resources(app "images"
PREFIX "/"
BASE "../shared"
FILES "images/button.png")
Alias is "../shared/images/button.png", and pro2cmake generated
QT_RESOURCE_ALIAS assignments to fix this.
New behavior:
qt_add_resources(app "images"
PREFIX "/"
BASE "../shared"
FILES "../shared/images/button.png")
The alias is "images/button.png". No extra QT_RESOURCE_ALIAS assignment
is needed.
The new behavior is in effect for user projects and for Qt repositories
that define QT_USE_FIXED_QT_ADD_RESOURCE_BASE. Qt repositories will be
ported one by one to this new behavior. Then the old code path can be
removed.
Pick-to: 6.1
Task-number: QTBUG-86726
Change-Id: Ib895edd4df8e97b54badadd9a1c34408beff131f
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
A previous change to fix QTBUG-85431 has caused this issue by emulating
the behavior of the non-enhanced mouse event handler in a particular
case, where mouse move events that did not change position were ignored.
However, some of these events seem to be involved in the dock drag&drop
implementation. This issue is also reproduced in pre-5.15 releases,
predating the QTBUG-85431 fix, by disabling the enhanced mouse event
handler by setting the QT_QPA_DISABLE_ENHANCED_MOUSE env var. However,
the ignored events in the current issue seem to be non-client events
only, while the QTBUG-85431 issue was related to client mouse events.
So we can restrict the test added in the QTBUG-85431 fix and have both
issues fixed.
Fixes: QTBUG-92182
Pick-to: 5.15 6.0 6.1
Change-Id: I98c0c8597912c7f4fe58af375a5a560695a82746
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
A QPageSize::PageSizeId is a faithful representation of a QPageSize,
so the corresponding QPageSize ctor shouldn't be explicit.
[ChangeLog][QtGui][QPageSize] Conversion from a QPageSize::PageSizeId
is now implicit.
Pick-to: 5.15 6.0 6.1
Change-Id: I2d32da370c032949686757400cb7c28583d9d8ac
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Reviewed-by: Giuseppe D'Angelo <giuseppe.dangelo@kdab.com>
The initial approach for providing public access to native
interfaces via T::nativeInteface<I>() was based on the template
not being defined, and then having explicit instantiations of
the supported types in a source file, so that the accessors
were exported and available to the user.
This worked fine for "simple" types such as QOpenGLContext
and QOffscreenSurface, but presented a problem in the context
of classes with subclasses, such as Q{Core,Gui}Application.
To ensure that a native interface for QCoreApplication was
accessible both from QCoreApplication and its subclasses,
while at the same time preventing a native interface for
QGuiApplication to be accessible for QCoreApplication, the
nativeInterface() template function had to be declared in
each subclass. Which in turn meant specializing each native
interface once for each subclass it was available in.
This quickly became tedious to manage, and the requirements
for exposing a new native interface wasn't very clear with
all these template specializations and explicit instantiations
spread around.
To improve on this situation, while also squashing a few
other birds at the same time, we change the approach to
use type erasure. The definition of T::nativeInteface<I>()
is now inline, passing on the requested interface to a per
type (T, not I) helper function, with the interface type
flattened to a std::type_info.
The type_info requested by the user is then compared to the
available types in a single per-type (T) "switch statement",
which is a lot easier to follow for someone trying to trace
the logic of how a native interface is resolved.
We can safely rely on type_info being stable between the user
application and the Qt library as a result of exporting the
type info for each native interface, by explicitly ensuring
they have a key function. This is the same mechanism that
ensures we can safely dynamic_cast these interfaces, even
across library boundaries.
The use of a free standing templated helper function instead
of a member function in the type T, is to avoid shadowing issues,
and to not pollute the class namespace of T with the helper
function.
Since we are already changing the plumbing for how a user
resolves a native interface for a type T, we take the opportunity
to add a few extra safeguards to the machinery.
First, we add a static assert in the T::nativeInteface<I>()
definition, that ensures that only compatible interfaces,
as declared by the interface themselves, are allowed.
This ensures a compile time error when an incompatible
interface is requested, which improves on the link time
errors we had prior to this patch, and also offsets the
one downside of type erasure, namely that errors are only
caught at runtime.
Secondly, each interface meant for public consumption through
T::nativeInteface<I>() is declared with a revision, which
is checked when requesting the interface. This allows us
to bump the revision when we make breaking changes to the
interface that would have otherwise been binary incompatible.
Since the user will never see this interface due to the
revision check, they will not end up calling methods that
have been removed or renamed.
One advantage of moving to a type-erased approach for the
plumbing is that we're not longer exposing the native
interface types as part of the T::nativeInteface symbols.
This means that if we ever want to rename a native interface,
the only exported symbol that the user code relies on is
the type info. Renaming is then possible by just exporting
the type info for the old interface, but leaving it empty.
Since no class in Qt implements the old native interface,
the user will just get a nullptr back, similarly to bumping
the revision of an interface.
Change-Id: Ie50d8fb536aafe2836370caacb22afbcfaf1712a
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
takeItem and takeChild do not signal the change correctly to the
external world, this change fixes the problem
Fixes: QTBUG-89145
Pick-to: 6.1 5.15
Change-Id: Ib4844ace53007068a2cd62eba64df99e6e45fdc0
Reviewed-by: David Faure <david.faure@kdab.com>
Could in theory be triggered with a 0 line spanning elipsis.
Pick-to: 6.1
Change-Id: I2166ee354d2f7488e1fcddfcb8c949c8ca2452fe
Reviewed-by: Eirik Aavitsland <eirik.aavitsland@qt.io>
Set clip->count to 0 unconditional not just when hasRegionClip or
hasRectClip is true.
Pick-to: 6.1
Change-Id: Ib3d1c4dc24373df3d4dbc393241226a8730bb9fc
Reviewed-by: Eirik Aavitsland <eirik.aavitsland@qt.io>
Previously, resizeMaximizedWindows() would use the device
independent screen size as the source of truth when
setting window sizes. However this size may have been
rounded, which means that e.g. a fullscreen window may
fail to cover the entire screen.
Instead, use the native screen size as the true screen
size. Set QPlatformWindow geometry, and let the
platform update QWindow geometry via geometry change
events.
Pick-to: 6.1
Fixes: QTBUG-87334
Change-Id: If6e4852dea46ab03c83e469808c0047bc933ee47
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
Since Windows can synthesize certain font traits for us, we
used to register these in the font database so that we could
match against them. But after change
469b139169, this in principle
no longer happens, because we opt out whenever there is a
style name (which there usually is, this could be e.g.
"Regular" for a normal font). The result of this was that
if we looked for a bold variant of a font, we would not find
it.
In cases where a multi-engine was used, the request for bold
would still survive in the multi engine's fontDef, so we would
still pick it up later and apply the synthesis. But when
NoFontMerging was set, then we would override the weight in
the fontDef with the one from the font database.
Since the comment documents that the additional registrations
are there to make sure all the variants that Windows can
synthesize are available for matching, it does not make sense
to skip them just because the font has a style name. So this
is a partial revert of 469b139169.
Note: This exposed an error in QFontDatabase::isSmoothlyScalable().
The style parameter here is not the "styleName" (as in sub-family),
but actually predates that API. Instead it is the "style" as
returned by QFontDatabase::styles(), which may be the style
name, but it can also be the generated description of the style
and weight. In the latter case, we would return false for fonts
that are actually smoothly scalable, which is incorrect. This
caused a failure in tst_QFontMetrics::metrics(). To remedy this,
we add an additional condition, and also match the style if it
matches the generated descripion of the style key.
[ChangeLog][Windows] Fixed an issue where bold/italic would not
be synthesized for fonts if QFont::NoFontMerging was set.
Pick-to: 5.15 6.1
Fixes: QTBUG-91398
Change-Id: Id2166a47ae2d386536cf6e5e27ff09165ae8a23a
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
The leak can be reproduced by running the test project from QTBUG-63557
with asan, although that report is about yet another leak.
[ChangeLog][gui][QRhiGles2] Fixed a memory leak in QRhiGles2
Task-number: QTBUG-63557
Pick-to: 6.1
Change-Id: Ic4d346abb36a5666feb3ceb881865b029f5a6945
Reviewed-by: Laszlo Agocs <laszlo.agocs@qt.io>
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
from qt/qtx11extras 0e67fb41cfc4b4bfbaa7dc75f8ddebdf5a08e836.
The plan is to expose these as native interfaces, so this is a first
step.
Task-number: QTBUG-83251
Change-Id: Iecba8db9a4f616a08a3750ddaae08cc30ec66f89
Reviewed-by: Liang Qi <liang.qi@qt.io>
The fast-path can't antialias edges, and shouldn't be used when
that is requested.
Pick-to: 6.1 5.15
Change-Id: I3a0ce120ab96a6f95d11c165d1f5b356dae009fe
Reviewed-by: Eirik Aavitsland <eirik.aavitsland@qt.io>
The machinery is not needed for all translation units, so keep it out
of qglobal.h.
Change-Id: Ib0459a3f7bc036f56b0810eb750d4641f567f1fe
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
We end up excluding more than 65536 from 65536 of rowHeight.
Perhaps better fixed earlier, but I can't figure this logic out right
now.
Pick-to: 6.1 5.15
Change-Id: I5721c469441f15ac112180f971c857cd67edbf96
Reviewed-by: Eirik Aavitsland <eirik.aavitsland@qt.io>
It has a problem with very small targets, and coordinates can't exceed
the same bounds we have on dimensions.
Pick-to: 6.1 5.15
Fixes: QTBUG-93475
Change-Id: If5b3af324f4e525cee3dc448ba41fdd8a91cc880
Reviewed-by: Eirik Aavitsland <eirik.aavitsland@qt.io>
Most nested block elements are merged together, so while we shouldn't
do real inheritance we need to do it when block elements are combined.
Pick-to: 6.1
Fixes: QTBUG-91236
Change-Id: I9e37b15f705db92c79a620d0d772f25d0ee72b8d
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@qt.io>
Also add a feature to the textedit example to set this value.
[ChangeLog][QtGui][CSS] The CSS text-decoration-color attribute is now
supported in rich text spans with underlines, overlines and strikethrough.
Fixes: QTBUG-82114
Task-number: QTBUG-39617
Change-Id: I0065cb5431833da55b0f503ce7ff2b83b74b718a
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
We were not using the capitalization details from specialData in the
regular path of QTextEngine::itemize() causing it to be ignored.
Pick-to: 6.1 5.15
Fixes: QTBUG-90840
Change-Id: I7bb71fad4009f6d0685905a946c36ac1d24d8d3c
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@qt.io>
QWindow receives keyboard originated context menu events, so it should
also receive events originating from a right-button mouse event.
Remove the incorrect statement about the special acceptance flag for
context menu events. There is no such thing, the event gets delievered
after the corresponding mouse press/release event.
Fixes: QTBUG-59988
Task-number: QTBUG-93486
Change-Id: I148310440604e74f600932cc1898fa152c483a61
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
The interface is already declared in qopenglcontext_platform.h,
so the inclusion in qcocoanativeinterface.mm was unintentional.
Change-Id: Ic6d4a86527d03787efd858c1a6568933be8a4327
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
When a user requested bold and there was no available font for
this (which is quite common with CJK fonts and in fact is the case
for the official Japanese font on the system), we should synthesize
the boldness.
This was done by checking if the requested font weight boldness
matched the one in the font's traits, and if not, we flag the
font boldness to be synthesized.
But when initializing the font, we would first override the requested
weight with the selected font's weight, *before* performing the check
above. So even if there was a mismatch, we would not catch this and
as a result, e.g. the system Japanese font would never be bold.
[ChangeLog][macOS] Fixed an issue where boldness would not be
correctly synthesized for families with no bold variant.
Fixes: QTBUG-85634
Pick-to: 5.15 6.1
Change-Id: I36da59d7689455e29cca283cb0724a0841095918
Reviewed-by: Konstantin Ritt <ritt.ks@gmail.com>
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
Two legacy bitmap fonts are misreported as TMPF_VECTOR on
Windows: Roman and Script. This causes them to be marked as
scalable, and the automatic fallback to NativeRendering in
Qt Quick does not kick in - causing the text elements to look
empty instead.
To work around this, we exploit the peculiarity that the type
of these two fonts is reported as "0" in the enumeration, which is
not a valid value. No other fonts on the system is reported as type
0, so we simply detect this error case and mark the fonts as
non-scalable, which is the safer choice.
[ChangeLog][Windows] Fixed text in "Roman" and "Script" bitmap
fonts not showing in Qt Quick applications.
Pick-to: 6.1 5.15
Fixes: QTBUG-85826
Change-Id: Id889f0dedb1d529e6dd64c6da9e17e303f4a9d04
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
With extreme painter scaling, linearAdvance may be too large to fit in
an unsigned short.
Fixes: QTBUG-91758
Pick-to: 6.1 5.15
Change-Id: I7bbe6e77ec9bcef4aa5259da1d3000ed1a8eb27a
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@qt.io>
A) When a QEdidParser gets built, the code populates a QMap by parsing
the system's pnp ids file. With > 2000 entries in the file, that's 2000+
allocations for the QMap, plus 4000+ allocations for key/value
(QStrings), plus a bunch of temporaries created for processing.
What's more, on XCB and EGLFS, a QEdidParser gets built *per screen*;
the map is not shared and gets rebuilt for each screen.
It's however completely unnecessary to keep this map in memory. The
lookup is required only once per monitor change event, which, apart from
application startup, is an extremely rare event.
When such an event happens, and we have to lookup a vendor id, we can
just process the pnp database "on the fly", therefore removing the
constant memory usage from the application run.
In order to also avoid an allocation spike, just don't build a map at
all: simply process the file, using no memory allocations, one line at a
time while searching for the vendor id. I was unable to find
information about the file format, so I am not sure if it's guaranteed
that it's kept sorted by id (this could allow for a faster fail path).
B) In case of a cache miss, a hardcoded vendor table is used to try and
identify the manufacturer.
Look up in the hardcoded table using binary search, rather than a linear
scan, and don't allocate memory for each element processed in the list
(!). The size of the list (2000+) is big enough to completely justify
binary search.
C) Drive-by, remove a TOCTOU bug when opening the system file.
D) Drive-by, fix the includes in the header to IWYU.
Change-Id: I57c7cbd09a145c6efe3023c227ed36b24bed96f9
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
When we are having nested lists then we need to ensure that the HTML is
outputted correctly so that the closing list and item tags are placed
in the right order.
[ChangeLog][QtGui][QTextDocument] The output of toHtml() now handles
nested lists correctly.
Fixes: QTBUG-88374
Pick-to: 6.1 5.15
Change-Id: I88afba0f897aeef78d4835a3124097fe6fd4d55e
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
It used to be ignored because we couldn't disable it, but that works
fine now. So re-enable it.
Fixes: QTBUG-91171
Change-Id: I4cf966211bb200b73326e90fc7e4c4d3d4090511
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@qt.io>
Print a warning if the name for a given location does not match. This is
relevant only to OpenGL and only to < 330 GLSL versions: there the
location qualifier is simply not present in the actual shader code and
all matching is name-based. This leads to hard to discover problems in
Qt Quick ShaderEffects for instance, where providing a custom fragment
shader but using a built-in vertex shader is common, and one may be
tempted to use a name other than qt_TexCoord0 for the input variable in
the fragment shader. Unfortunately this breaks, sometimes silently, when
not having location qualifiers. (and we won't, neither in GLSL 120 or
150 which are the standard for Qt Quick)
Make this situation recognizable by printing a warning to the debug
output.
Pick-to: 6.1
Task-number: QTBUG-92500
Task-number: QTBUG-93370
Change-Id: I0d0bcc135e23a228783f7633f872e39c4e43bb93
Reviewed-by: Christian Strømme <christian.stromme@qt.io>
For optimized/release builds, cl 19.28 produces some false positive
"potential divide by 0" warnings even for code where no valid code path
can lead to such a division.
Fixes: QTBUG-92940
Pick-to: 6.1
Change-Id: I461f9104d7b9dc2d74839bf30bebb2b9d4cf2db3
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
Reviewed-by: Eirik Aavitsland <eirik.aavitsland@qt.io>
Move parsing to initHighDpiScaling() and store the parsed
factors in a vector. The factors are later applied in
updateHighDpiScaling() where the factor list is compared
to the screen list.
Change-Id: I59fc67b84932dbe69868a3683686a91e51333104
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
In Qt Quick we now need to keep track of which QQDeliveryAgent is
responsible when a point is grabbed, either passively or exclusively.
When we re-deliver to that grabber, we need to do it via the same agent,
so that the same scene transform is used, and the grabber will see the
event in the correct coordinate system. It's easier to track this
mapping here instead of in a separate map in Qt Quick.
(This is an alternative to 40330b8f0a717098982d1f54f34a18a8262b1f55:
it was not possible to use QFlatMap, because we need to keep the passive
grabbers in the same order as they were added. We don't use a QList of
structs, because QPointerEvent::passiveGrabbers() needs to return a
QList of just the grabbers, and it's not as efficient to construct that
list in the accessor.)
Change-Id: I457114f816736749d2ea5ee48fa03524eb93d2d0
Pick-to: 6.1
Task-number: QTBUG-92944
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
Reviewed-by: Andy Nichols <andy.nichols@qt.io>
This reverts commit 40330b8f0a.
It was a bad idea to use QFlatMap here, because it is a sorted map, but
we need to keep the passive grabbers in the same order as the grabs happened.
So need to go back to an earlier version of the patch that uses two parallel QLists.
Pick-to: 6.1
Change-Id: I9e6013c2565986fe1eb9fd754f8259766f83bee5
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
QHighDpiScaling has two init/update functions:
- initHighDpiScaling(): called once during QGuiApplication construction
- updateHighDpiScaling(): called whenever (relevant) screen configuration changes
Currently the calls to updateHighDpiScaling() are made from
multiple places including platform code. Simplify by calling
it from two locations:
- QWindowSystemInterface::handleScreenAdded()
- QGuiApplicationPrivate::processScreenLogicalDotsPerInchChange()
Replace comment about early calls to qt_defaultDpi with a
test which calls qt_defaultDpiX/Y with no screens attached.
(Looking at the qt_defaultDpiX() implementation, it is unlikely
that there will be a problem as long as updateHighDpiScaling()
is called before QGuiApplication::primaryScreen() starts returning
a non-null value.)
Change-Id: I447db42894617495843a5cb531a1322b000fed62
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
In Qt Quick we now need to keep track of which QQDeliveryAgent is
responsible when a point is grabbed, either passively or exclusively.
When we re-deliver to that grabber, we need to do it via the same agent,
so that the same scene transform is used, and the grabber will see the
event in the correct coordinate system. It's easier to track this
mapping here instead of in a separate map in Qt Quick.
Pick-to: 6.1
Task-number: QTBUG-92944
Change-Id: I69f769c694d0da24885cdf4087e5032022bff629
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
And adjust configure.json files that were stale
Amends d385158d52
Change-Id: I851838a12c3773a6e8119ebc7f1de941ae7fe224
Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
It's not clear now whether trackpad gestures on Windows will need to be
so different than on macOS; however, any reasonable int value can be
stored in a qreal, and in Qt Quick we like to use floating-point numbers
for all "real" values and measurements. So since we need to add more
storage, and quint64 m_intValue has never been used, we now replace it
with a QVector2D, which should have the same size. The intended use
is that PanNativeGesture will include a displacement, probably in
pixels, by which the viewport or some target item should be panned or
moved. The naming of deltas() is flexible enough to support any gesture
with some incremental 2D valuators, though, just as value() has
gesture-dependent semantics.
fingerCount() will be useful for Qt Quick pointer handlers to filter
out events that have the wrong number of fingers, e.g. to require that
either a 3-finger pan gesture or 3 individual touchpoints are required
to activate DragHandler { minimumPointCount: 3 } (assuming we implement
pan gesture support in DragHandler).
Fixes: QTBUG-92179
Task-number: QTBUG-92098
Change-Id: I5462aea9047beed6e99075294a62011edd8a59f5
Reviewed-by: Povilas Kanapickas <povilas@radix.lt>
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
We now generate metatypes by default and every instance of GENERATE_METATYPES now causes a warning.
Change-Id: I83a124de3f3bb1fc770cfae0eed955c03e95b310
Reviewed-by: Ulf Hermann <ulf.hermann@qt.io>
The fact that KernedAdvances does not work correctly on a majority
of fonts is confusing. So we warn users that this is the case and
point them to the correct function to use instead.
Task-number: QTBUG-92930
Change-Id: I70f4b09ea1050fceabbff25a9c91008d1754f772
Reviewed-by: Konstantin Ritt <ritt.ks@gmail.com>
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
When we can't get the icon for the first time, we would use the data
in the cache later, and we will never get the new icon when the system
theme is updated while the application is running.
Pick-to: 6.1
Change-Id: I839ad9983918561a1dc6bc842f85477bba53f64a
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
Flatpak doesn't allow to own random name with PID. Even after adding
such a permission into manifest, all flatpaked apps have PID 2, so only
one Qt application at a time can have tray icon.
Even though unique name is a part of the spec, no tray hosts really
check it and SNI implementations without unique name run just fine
inside and outside of Flatpak.
This fixes the inability of Qt applications to have tray icon in Flatpak
outside of KDE.
Pick-to: 6.0 6.1 5.15
Change-Id: Ieea6dc335b7a74537a51929f6e70ca68c84228fb
Reviewed-by: Dmitry Shachnev <mitya57@gmail.com>
Reviewed-by: Shawn Rutledge <shawn.rutledge@qt.io>
Add SOLARIS cmake platform definition.
Add settings for QT_DEFAULT_MKSPEC so that qplatformdefs.h can be found.
Solaris has its gssapi symbols in libgss.
Solaris supports @ORIGIN.
Solaris ld does not support --dynamic-list needed for reduce relocations.
Make solaris fail the reduce relocation test.
getauxval is specific to GNU libc and some other libc implementations on
Linux but sys/auxv.h is not. The bootstrap uses sys/aux.h as the only
indication for getauxval. This breaks builds on Solaris, so only make
sys/auxv.h an indicator for getauxval on linux or glibc based systems.
Solaris uses X11 so add it to the X11_SUPPORTED list.
Solaris network libraries for sockets etc are in socket and nsl.
ifreq does not have a member ifr_ifindex on Solaris, it uses
ifr_index. Add test to check if ifr_index is a member of ifreq.
The first struct in the in_addr union on solaris is defined as four
uint8_t, therefore four arguments are needed for its initializer list.
Change-Id: Ieed4c1bbac8559a7ae1db9c4e1e91f609f150270
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
No other platform needs it, so avoid building and shipping unused code.
Change-Id: I23b68b58eb0523459014df915ad8516c22adbcc1
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
In C++ we can give names to classes, so just use that, without
C-isms (typedef struct).
Change-Id: I27239d8d5c28864b3f4f7bd4013cc47c045b4b04
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
No need to spam the summary with this unsupported feature.
Pick-to: 6.1
Change-Id: Ie644a7077762d818fd62a22a8b6d3db025f92c81
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
When the descent of the item is 0, ascent is the
height of the item, base(base = si.ascent)> sl.base.
At this time, sl.descent is not considered. The
calculated y value may be <0.
Fixes: QTBUG-86823
Fixes: QTBUG-92468
Pick-to: 5.15 6.0 6.1
Change-Id: I9cf088dec9162595e52ff72aa90ec3153a30fb72
Reviewed-by: Shawn Rutledge <shawn.rutledge@qt.io>
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
Xcode allows building a project targeting either the device or
simulator sysroot in one single build dir, but for the sysroot
switching to work there should be no linker or compiler flags
referencing absolute paths of a specific sysroot.
During CMake configuration of a project targeting iOS, all found
system libraries will be within one single sysroot, either the device
one or the simulator one, whichever one was passed to
CMAKE_OSX_SYSROOT. CMake will then generate the Xcode project
and pass those absolute paths, which makes sysroot switching within
Xcode not work.
To avoid that, the CMake documentation recommends passing linker and
framework flags of the form '-lfoo' and '-framework bar' instead of
absolute paths. Xcode then takes care of setting the correct framework
search path.
Zlib is one of the libraries found in the iOS sysroot and thus passed
as absolute path.
To avoid that, create a new FindWrapZLIB find script. The target it
creates will pass the absolute path to the library on non Apple
platforms and an -lz linker flag on Apple platforms (macOS and iOS).
To avoid issues with target global promotion when system PNG package
is found, ensure that a found ZLIB::ZLIB target is promoted to global
manually in src/gui/configure.cmake.
Pick-to: 6.1
Change-Id: I8bd8649be4f680a331ad51925f27cb9d13ac5e5f
Reviewed-by: Cristian Adam <cristian.adam@qt.io>
The engine pointer is guaranteed to be initialized in line 1838, so we
can assert that the correct value is passed into the function to fix
static analyzer warning 1d9b8ce922ee0891fb0d477dc17fdb8d.
Change-Id: I773bbaa579afec0d7a79d4393ee66fd26ba9629b
Reviewed-by: Konstantin Ritt <ritt.ks@gmail.com>
Reviewed-by: Jonas Karlsson <jonas.karlsson@qt.io>
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@qt.io>
Various checks for "engine" in the previous code suggested that
engine might be nullptr by the time we want to populate the
out-parameters.
This must not be the case, and QFontDatabase::load asserts
already that a valid engine is loaded and returned.
Fix static analyzer warning 7f68daa282c72e8cc172c681eb02f559 by
asserting it here as well.
As a drive-by, change the tested out-parameter to the last one
in the list of optional parameters.
Pick-to: 6.1
Change-Id: I3d9ff0f5f7c4740014301c073480d14fef54e2fb
Reviewed-by: Jonas Karlsson <jonas.karlsson@qt.io>
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@qt.io>
Prior code was naively assuming the character after a high surrogate
would necessarily be a low surrogate, which is buggy.
Fixes oss-fuzz issue 29718.
Pick-to: 6.0 6.1 5.15
Change-Id: I10f023c4b5024a0d76fea0a3672001063591ec6d
Reviewed-by: Konstantin Ritt <ritt.ks@gmail.com>
Reviewed-by: Robert Löhning <robert.loehning@qt.io>
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
With this enumeration value, users can get the width of
the current font TAB character more conveniently
In addition, added some documentation about where
users may use this enumeration.
Fixes: QTBUG-92205
Done-with: Volker Hilsheimer <volker.hilsheimer@qt.io>
Done-with: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
Change-Id: I9794b7553e9299e351f9182de02866d07a1393fc
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
We want right - left, but the numbers we have are (1 - left) and right,
so we need right - (1 - left) = right + left - 1.
Pick-to: 6.1 6.0
Fixes: QTBUG-91957
Fixes: QTBUG-92485
Change-Id: I238cbbe1eebddff1ce56da38127899cdbd21db0e
Reviewed-by: Eirik Aavitsland <eirik.aavitsland@qt.io>
When stroking a dashed path, an unnecessary amount of processing would
be spent if there is a huge number of dashes visible, e.g. because of
scaling. Since the dashes are too small to be indivdually visible
anyway, just replace with a semi-transparent solid line for such
cases.
Pick-to: 6.1 6.0 5.15
Change-Id: I9e9f7861257ad5bce46a0cf113d1a9d7824911e6
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
The Unicode 13 language tags are available since Harfbuzz
2.6.7.
For compilation with earlier versions, also update the
hardcoded tags to match the ones in Harfbuzz (and in
iso-15924):
https://unicode.org/iso15924/iso15924-codes.html
Task-number: QTBUG-90217
Change-Id: I4e2cfdf61a5c4189b8aac51644c9d80c816aeebc
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
(cherry picked from commit 910ac477f013f2af690b66f1ac41d65ab7527843)
Reviewed-by: Qt Cherry-pick Bot <cherrypick_bot@qt-project.org>
QPixmap doesn't have a devicePixelSize() method, but
deviceIndependentSize is a relevant method here.
Change-Id: Ie71e6cc88534792a896ea17ec37b1b9737c2f986
Reviewed-by: Paul Wicking <paul.wicking@qt.io>
Reviewed-by: Nico Vertriest <nico.vertriest@qt.io>
Some parts of the GLX code is only enabled for Linux. This makes
builds on other X11 supported platforms break when GLX is found.
To fix this enable these parts of the code when Qt feature
xcb-glx-plugin is enabled. xcb-glx-plugin has to be made public
in order for QT_CONFIG(xcb_glx_plugin) to work correctly in
these parts of the code.
Change-Id: I6bf78b6a64787ed88c8e2fae40675244c9198c37
Reviewed-by: Laszlo Agocs <laszlo.agocs@qt.io>
Reviewed-by: Liang Qi <liang.qi@qt.io>
By setting these variables after the platform window has been destroyed,
it will enable any platform specific code to handle changes in terms of
focus and so on before the destruction of the window changes these
directly. For example, this will prevent a problem with iOS where it
keeps track of the current focus object as the signal indicating that
the change to focus will still be emitted. This is before the variable
is set to nullptr if there is no parent window.
Change-Id: Ie540c10760d06dc62e163ccf6f6edea200b43bbf
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
Laszlo Agocs
Volker Hilsheimer
Tor Arne Vestbø
Morten Sørvig
Richard Moe Gustavsen
Eskil Abrahamsen Blomfeldt
Paul Olav Tvete
Edward Welbourne
Lars Schmertmann
Tang Haixiang
Jan Grulich
David Faure
Zhang Hao
Sze Howe Koh
Shawn Rutledge
Kai Köhne
Eirik Aavitsland
Ivan Solovev
Fan RuiJie
Albert Astals Cid
Waqar Ahmed
Jonas Kvinge
Povilas Kanapickas
Yuhang Zhao
Tasuku Suzuki
Andy Shaw
Marc Mutz
Giuseppe D'Angelo
Oliver Wolff
Alexandru Croitor
Nicolas Fella
Robert Löhning
Nico Vertriest
Allan Sandfeld Jensen
WenTao Song
Oliver Eftevaag
Topi Reinio
Friedemann Kleint
Ulf Hermann
Paul Wicking
Andreas Buhr
Alexey Edelev
Joerg Bornemann
Tim Blechmann
Kimmo Ollila
Albert Astals Cid
Cristian Adam
Fabian Kosmale
Niclas Rosenvik
Jan Arve Sæther
Elvis Lee
Andrei Golubev
Inho Lee
Andrey Butirsky
Dongmei Wang
Andre de la Rocha
Luca Beldi
Maximilian Goldstein
Fan PengCheng
Ilya Fedin
Qt CI Bot