This ensures ccache or sccache is used to cache the compilation of the
headerclean checks.
Task-number: QTBUG-82615
Change-Id: Ie944eb1d643e7271551c9f8337609741e419e9d8
Reviewed-by: Cristian Adam <cristian.adam@qt.io>
It appears there's a difference between the chosen macOS compiler path
in the CI versus the compiler path on my local machine.
In the CI the chosen compiler path ends up 'being /usr/bin/clang++'
whereas for me locally it's
/Applications/Xcode.app/Contents/Developer/Toolchains/XcodeDefault.xctoolchain/usr/bin/c++
For some reason the headersclean commands succeed in the CI, but
locally they fail for me saying that standard library includes can not
be found, unless an explicit sysroot flag is specified.
I assume that in the CI the '/usr/bin/clang++' compiler shim chooses some
implcit sysroot, whereas the longer Xcode compiler expects an explicit
sysroot.
It's probably also affected by the fact that in the CI we pass an
explicit CMAKE_OSX_SYSROOT to a non-standard Xcode location e.g.
/Applications/Xcode11.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX10.15.sdk
Note the '11' in the Xcode app name.
Locally I don't pass a custom CMAKE_OSX_SYSROOT, and my Xcode is
installed in a regular location e.g /Applications/Xcode.app.
The sysroot flag and path is added to regular CXX compilation rules
inside CMake's core (in cmLocalGenerator.cxx).
Reuse the same variables that CMake uses and add them to our headersclean
command rules.
Task-number: QTBUG-82615
Change-Id: Ic03ea27e39471f5fa168eb5970bf3d3f1d1be251
Reviewed-by: Cristian Adam <cristian.adam@qt.io>
Add 3 new convenience custom targets:
'qt_plugins', 'qpa_plugins' and 'qpa_default_plugins'.
Additionally, if we detect that an internal executable / test
links against Gui, add a dependency on the 'qpa_default_plugins'
custom target, so that if a developer configures Qt for the first time
and then calls ninja 'tst_foo_check', we ensure the test will launch
successfully because the default QPA plugin will also be built.
Change-Id: If6dd70844b5effdf8a293f65f8785855cc85b132
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
Move the existing “kitchen sink” test out of the way,
and rename it to “kitchensink”
Change-Id: I121260e640bb2810a94f8112fcea212e97055fb3
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
Return immediately if processing events caused the
predicate to become true.
This gives us a nice speedup on tests with call
qWaitForWindowExposed() or qWaitForWindowActive(),
for example tst_QGraphicsView::itemsInRect_cosmeticAdjust:
cocoa 1164ms -> 825ms ~30% decrease
offscreen 296ms -> 15ms ~95% decrease
Change-Id: Ifbab0ca662c082e7dfb609d0cb4bc1161709067b
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
At least for qdoc, hidden friends are class members.
Change-Id: I6eaa21565937cd49c0905ee47b8b82b0c1765bc1
Reviewed-by: Topi Reiniö <topi.reinio@qt.io>
Reviewed-by: Paul Wicking <paul.wicking@qt.io>
It was way too small, and should match the margins
for other content such as paragraphs, so that an
inline table in between two paragraphs has a balanced
spacing.
Change-Id: I2adb434ac7773e9796199f40d5318368ca380aca
Reviewed-by: Paul Wicking <paul.wicking@qt.io>
QList::parameter_type is defined and used to give better
performance e.g. for arithmetic types. Let's use it consistently
in QList API instead of const T &
Change-Id: I2e12bd83f55679b55a14fbb23ab6172a9cf7bbcc
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
We're ensuring its existence right in the next line.
Change-Id: If7a55511c9ef839a9f9ae70a7928b58ad768ae65
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
We need to make sure to create a config.opt file, even if it's empty.
This is the same fix we did for the .bat file in
dad9550305.
Change-Id: I79c507cfe0f9517d3033a773d3a5cf69eb13cfe6
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
For example, we must not prepend -l to -pthread, -framework and -lfoo.
Fixes: QTBUG-87760
Change-Id: Ie1bc7a76183c2c4980c519b5f23dde6c47ec85a9
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
...instead of qt-cmake-private.
Also, add ".bat" on Windows.
Change-Id: I2aa94ac76f2dadfb8e94d68b19bb379b3d45a93d
Reviewed-by: Oliver Wolff <oliver.wolff@qt.io>
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
The variable to check for a Windows host system was mis-typed.
Change-Id: I25b14b80d25bfec0c1a00e99833520b6fb6a4b02
Reviewed-by: Oliver Wolff <oliver.wolff@qt.io>
on Ubuntu 20.04
There is some issue with the glib event dispatcher.
Task-number: QTBUG-87728
Pick-to: 5.15
Change-Id: I4d64206898dd2c8356d5fc51a68c2e5759b38aac
Reviewed-by: Tony Sarajärvi <tony.sarajarvi@qt.io>
This makes the diff-ing of the config summary between qmake and CMake
builds correct.
Change-Id: I720b69572c23afd78e6d0bea6cdf0740980c3b36
Reviewed-by: Cristian Adam <cristian.adam@qt.io>
In configure.json files we have inputs with dashes, e.g.
bundlex-xcb-xinput. In configure.cmake files, these are read in their
normalized form, e.g. INPUT_bundled_xcb_xinput.
Normalize the input names in QtProcessConfigureArgs.cmake like we
already do for feature names.
Change-Id: Iece414d40a0e9e2920580f2fda68e25cd32674c9
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
RFC2822 requires times to be in the format 'HH:mm' or 'HH:mm:ss'.
We did not have unit tests to check that malformed RFC2822
dates are rejected. This patch adds such unit tests for
truncated hours/minutes/seconds.
Change-Id: Id5b9390112e633e617722439ad59439e6aeba841
Reviewed-by: Andrei Golubev <andrei.golubev@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
In Qt Quick, when we deliver an item-specific QTouchEvent that contains
only the subset of eventpoints that are inside the Item's bounds,
traditionally the Item can accept the event to tell the delivery logic
that the event is handled and doesn't need to be delivered further.
But an Item cannot be expected to have total scene awareness; so now,
the delivery is "done" only when all eventpoints in the original event
are accepted. This behavior has been working well enough already due to
logic in QQuickWindow that iterates the points and accepts them if the
event is accepted; but it seems appropriate to move this enforcement
into QPointerEvent itself. Making setAccepted() virtual gives us a
useful degree of freedom.
Event-handling code should alternatively use QEventPoint:setAccepted()
or QPointerEvent::setExclusiveGrabber() to take resonsibility for only
a subset of the touchpoints.
Another way to put it is that we treat QPointerEvent::setAccepted() as a
convenience method: accepting the QEventPoints is what counts (at least
in Qt Quick).
Change-Id: Icec42dc980f407bb5116f5c0852c051a4521105a
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
Previous fixes made QPushButton correctly respect the style sheet boxing
model (as it's documented to do), ignoring clicks that were within the
margin area of the button (ie outside the bevel). However, a hover state
selector in the style sheet would still be used for the entire widget.
Turn on mouse tracking for widgets that have a hover state selector, and
handle mouseMoveEvent to set an explicit hovered state only when the mouse
hits the button. Use that state to initialize the style option.
Fixes: QTBUG-87706
Change-Id: I2f423b760c85cfab9faac4be44a5c7dcf2ba1c23
Reviewed-by: Richard Moe Gustavsen <richard.gustavsen@qt.io>
Leave the normalizedPos warnings, there is no equivalent function.
Change-Id: I50c72ab24b4855e36941aafdee30cdb0e94c1684
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
Reviewed-by: Shawn Rutledge <shawn.rutledge@qt.io>
On a mac that can build all of Qt with CMake, Ninja, and command line tools,
there should be no warning just because a build tool that
won't get used is not available.
Turn the warning into an informative message instead so that IDE
integrations (such as VSCode's CMake integration) don't flag warnings.
Change-Id: I250c0e5dd0633b36ff2b690a52ba7ce3ceb22218
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
g++-10.2 fails to compile qregion.cpp in RelWithDebInfo configuration,
with error message:
qregion.cpp:3617:12: error: ‘ET.EdgeTable::ymax’ may be used
uninitialized in this function [-Werror=maybe-uninitialized]
g++ is right here, for Count==1, 'ET' is not initialized but used.
This patch fixes this by Q_ASSUMEing Count > 1.
Pick-to: 5.15
Change-Id: I3b9f1f58de9f3811c60640a08334487e3f8f2b23
Reviewed-by: Andrei Golubev <andrei.golubev@qt.io>
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
glib event sources can have a name, but it is not required.
Internal to glib, it is common to give them a name, see for example
https://git.io/JTZ8g .
This patch gives a name to each glib event source created in qtbase.
Task-number: QTBUG-84291
Change-Id: I4f04526dcec082242312e3a66da2adf37a22e626
Reviewed-by: Giuseppe D'Angelo <giuseppe.dangelo@kdab.com>
We cannot rely on "QString a; a.insert(0, u'A');" to give
a.capacity() >= 3, this is clearly an implementation detail. Changed
the check to a meaningful one
Task-number: QTBUG-87416
Change-Id: I2e017c1292d360e32b85b903361027485c08ea74
Reviewed-by: Assam Boudjelthia <assam.boudjelthia@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Looking up a locale with unspecified language got the C locale, due to
taking a short-cut that would make sense if no locale were found for a
specified language. Stop assuming the language was specified.
Task-number: QTBUG-74287
Pick-to: 5.15 5.12
Change-Id: I8b3c232da584fb187ebb6c190729c377d0083808
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Moves them to class scope, which will avoid them showing up as
possibilities in error messages for missing operators.
Also consolidates how they are compared, so QRectF and QSizeF act
similar to QPointF.
Change-Id: I1e12cb7e5a5c65e85c32281878da03c6136c17de
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
Index of the last visible tab was not updated, if a new tab was inserted
after the current tab and before the last tab.
When the new tab is inserted before the last visible tab, the index of
the last tab increments by one.
When the new tab is inserted after the last visible tab, then the newly
inserted tab becomes the last visible.
Fixes: QTBUG-86898
Change-Id: I2f4b7f705261ec35a5aa7b883ecdddba25f007b7
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
Extract checkPositions helper function to re-use in the new test.
Task-number: QTBUG-86898
Change-Id: I5c8241b5701cd8c8c3e21607c385217d4b75e728
Reviewed-by: Jordi Pujol Foyo <jordi@vikingsoftware.com>
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
Reviewed-by: Christian Ehrlicher <ch.ehrlicher@gmx.de>
Expose event would not be sent when window was resized
Fixes: QTBUG-69155
Pick-to: 5.15
Change-Id: I81bf2d54f830a0dabf15398e1f25b55ff7ff4479
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
Reviewed-by: Rami Potinkara <rami.potinkara@qt.io>
Reviewed-by: Andy Shaw <andy.shaw@qt.io>
Reviewed-by: Heikki Halmet <heikki.halmet@qt.io>
When using OpenFile portal and passing a file descriptor, we open the
file with qt_safe_open() which by defaults make the file descriptor
writable. However we didn't specify in options that the FD is writable
which leads into rejection on xdg-desktop-portal side as there is a
mismatch between writable FD and read-only request.
Pick-to: 5.15
Change-Id: I7a430339a9615f0a054e777e0f3de56e219d1706
Reviewed-by: Aleix Pol Gonzalez <aleixpol@kde.org>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
This was the case on Android, and caused a crash
Fixes: QTBUG-83916
Pick-to: 5.15
Change-Id: Iabc17237498091d815241730a2b14ece5a45d0b9
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
Reviewed-by: Assam Boudjelthia <assam.boudjelthia@qt.io>
Failure to copy m_inputEvent and m_pointerEvent actually left them
uninitialized, and resulted in random behavior in Qt Quick when
Flickable clones a pointer event for later replay.
Remove the comment about copying events being a "bad idea" in Qt 4,
while we're at it. Copying became more common in Qt 5, and we
probably won't be able to stop doing it now.
Change-Id: I40b6ba5ad696e7aaafbeefbca86eca00cab40616
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Unlike the 32-bit version, we can't go to a bigger integer type to do
the multiplication with. So instead accept looping. Both libstdc++ and
libc++ implement std::uniform_int_distribution this way anyway, but in a
far more complex way.
There is no looping if the "highest" is a power of two. The worst-case
scenario is when "highest" is one past a power of two (like 65). In that
case, we'll loop until the number is in range. Since all bits have equal
probability of being zero or one, there's a 50-50 chance that the most
significant useful bit will be set[*], in which case we'll need to loop
and we again get the same probability. So on average, we only need two
iterations to get an acceptable result.
[*] There's also a possibility that the other bits are such that the
number is still in range. For 65, we'd need the other 5 bits to be zero
(64 is a valid result), but the probability of that is only 1/2^5 =
3.125%. The bigger "highest" is, the closer we get to zero, so
approximate by saying that never happens and instead calculate that the
most significant useful bit is the controlling one.
[ChangeLog][QtCore][QRandomGenerator] Added 64-bit versions of the
bounded() functions. They are useful in conjunction with Qt 6's 64-bit
container sizes, so code that used to call bounded(list.size()) in Qt 5
will continue to compile and work in Qt 6.
Fixes: QTBUG-86318
Change-Id: I3eb349b832c14610895efffd16356927fe78fd02
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
When building tst_qmetatype.cpp, clang generates the warning that
class 'AlignmentDummy' does not declare any constructor to initialize
its non-modifiable members
Turn the class into a struct, which doesn't generate such warnings.
Change-Id: I61013a10418238a11824b18ff1e927bbafa46ec2
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
And remove the limitation to 8 types from the documentation.
Change-Id: I92c67368e53d69fd851886c621f3f894f638bae9
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Selective application of clang-format to follow our coding style where
it is significantly off.
Change-Id: I0ff4ed146fe53922691d5473d0c236f31d478a04
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
...for QT_BUILD_TOOLS_WHEN_CROSSCOMPILING.
qt_internal_export_additional_targets_file now gets two lists of target
names when run from qt_export_tools:
- TARGETS containing actually existing targets, and
- TARGET_EXPORT_NAMES containing the target names as they appear in the
additional target info file.
Operations that require actual targets are run on the TARGETS, in the
additional target info file only TARGET_EXPORT_NAMES are written.
This distinction is required for the case where the host Qt lacks a
tool that is built in the target Qt.
Example: host Qt is built with DEVELOPER_BUILD=OFF, target Qt is built
with DEVELOPER_BUILD=ON. Then the host Qt lacks qmljs, but it is built
in the target Qt. TARGETS contains qmljs_native, and
TARGET_EXPORT_NAMES contains qmljs.
Fixes: QTBUG-87693
Change-Id: I615aed996bfcbe654274defcda8c1cb2cc4b7b4e
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
