Some CMake files currently assume that QtGui is always enabled and we
get a configure-time failure without these changes.
Task-number: QTBUG-86053
Change-Id: I28e32c180c32221f32519017bac6b518a19d5983
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
If a consumer wants to cross-compile a Qt app, they need to provide
the host Qt path location.
When building Qt in the CI we embed the Qt host path into the
generated CMake toolchain file for convenient building of other Qt
modules.
But once Qt is built, packaged and installed on a user's machine, most
likely the Qt host path will not be the same. In such a case, if the
user wants to use the convenience toolchain, they should explicitly
provide the Qt host path via the QT_HOST_PATH and
QT_HOST_PATH_CMAKE_DIR variables.
Show an error message if the values are not provided or don't exist
on the file system.
It's possible that in the future the Qt installer will patch the
toolchain file, or provide additional info, to point to the host
Qt installation so that the user doesn't have to do it manually.
But until that's done, a friendly error message is a good way to
inform the user what they should do.
Task-number: QTBUG-83999
Change-Id: I26291e3c47bb77375f8a5ce7b848c0382a660ca9
Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Public consumers of the qt toolchain file will most likely not have
their compilers in the same location where they were on the Qt build
machine.
Only set the compiler paths if none was set already, and the paths
actually exist.
This seems to become a trend in the generated toolchain file, and is
only a stop-gap solution.
A proper solution (two different toolchain files) may follow.
Task-number: QTBUG-83999
Change-Id: I7a603af447333a45c65b98e299ee109932d16517
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
Do follow the example of winmain.pro and replace /Zi with /Z7.
Manipulating CMAKE_CXX_FLAGS_DEBUG is not particularly elegant,
but CMake seems not to offer a more modern way yet:
https://gitlab.kitware.com/cmake/cmake/-/issues/19084
Fixes: QTBUG-86423
Change-Id: I280d06ad52b79d4be04052ab4e0d2113ef08414c
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
A previous change accidentally broke what we we set for
QT_HOST_PATH and QT_HOST_PATH_CMAKE_DIR.
The QT_HOST_PATH variable should use an absolute path as it was
done before.
The QT_HOST_PATH_CMAKE_DIR variable incorrectly used the value of
"${QT_HOST_PATH}" instead of "${QT_HOST_PATH_CMAKE_DIR}".
Fix both of these, and change the names of intermediate variables
to be consistent.
Amends a6a3b82ffb
Task-number: QTBUG-85240
Change-Id: I328a7edee12a13ff793684e8a0a4c2e03204eca4
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
The qfilesystemmodel.pro was recently moved to the gui tests subfolder,
but it uses widgets inside. Skip it in a no-widgets build.
Amends fb9ec8ad44
Task-number: QTBUG-86187
Change-Id: I955556ffddad483d4c25602ae126b8c2433091b2
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
This requires refactoring of QMetaSequence, as they share a lot of
common functionality. QMetaAssociation provides a low level interface to
an associative container.
Task-number: QTBUG-81716
Change-Id: I273e00abd82f1549ba8803c323d82aa3a2d12ded
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
Updated QIcc::fromIccProfile() and friends to not rely on
QByteArray pointer alignment. Used qFromUnaligned() instead
Task-number: QTBUG-84267
Pick-to: 5.15
Change-Id: I69ef7e011707bec27cd84693e7f0e92d79a577d1
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
[ChangeLog][QtCore][QString] QString is a prepend optimized
container similar to QList.
Task-number: QTBUG-84320
Change-Id: Id035aac859a6f404e389f2eeaf20f9eee5bff20c
Reviewed-by: Paul Wicking <paul.wicking@qt.io>
Amends f9940b15f7, which
amended 6e1d70ae12.
The bevel of the button is not defined by the contentsRect, but by the
borderRect that the stylesheet style calculates.
Change-Id: I9a0d5bf29a06ce2270014f0d144e33cc3a1a7473
Pick-to: 5.15
Fixes: QTBUG-86587
Reviewed-by: Richard Moe Gustavsen <richard.gustavsen@qt.io>
Otherwise we hit an #error statement in MSVC standard library.
Change-Id: Ib029edf0be8513a80f2640fd9ca75541615a0448
Reviewed-by: Giuseppe D'Angelo <giuseppe.dangelo@kdab.com>
test.bjon file doesn't exist, however, it's still included in json.qrc.
That causes the test to fail on Android because it's using that resource
file.
Change-Id: I1e93076069073a1a621459a7d90aec5c0c03f768
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
The macro BACKTRACE_HELPER_NAME is defined inside a block with
#ifdef __GLIBC__ which is not defined on Android, thus causing the test
to fail on Android.
Change-Id: I55e9b3e3bae2da182481239ad88107c36e3bd438
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
This prevented QDoc from parsing the properties correctly.
Change-Id: Ib3e02eb29a5c35034c30a3dd9fdf797d4c7e1550
Reviewed-by: Paul Wicking <paul.wicking@qt.io>
For this build, cxx11_future is not available and thus
QThread::create() as introduced by
ed114b728d does not work.
Revert back to implementing a QThread.
Pick-to: 5.15
Fixes: QTBUG-86575
Task-number: QTBUG-85676
Change-Id: I86a91f6bcdfc88804b35bf617362d92f37e51dea
Reviewed-by: Kai Koehne <kai.koehne@qt.io>
This was overlooked when doing the conversion to use UTF-8 as the
standard 8 bit encoding for text.
Fixes: QTBUG-54942
Change-Id: Ib7b1b75b4d694648ab7143f6930b6bb1dcad19c9
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
It has been deprecated and will live in qt5compat from now on.
Fixes: QTBUG-86480
Change-Id: I3744c7cee058d51d0fce633a174ab1a0f9235d2c
Reviewed-by: Karsten Heimrich <karsten.heimrich@qt.io>
The seconds and milliseconds default to 0; so omit them when they're
zero, unless there's nearby code that contrasts nicely with them.
Change-Id: Iea049015e976f5a5209fe87a2cbfdae9de49a559
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
It replaces the deprecated int margins property, which was a Q_PROPERTY.
This enables Qt for Python users to use property syntax for it.
Task-number: PYSIDE-1019
Change-Id: I958d2ef7f8e2afcc2be490e29d4b11c070569fbd
Reviewed-by: Christian Tismer <tismer@stackless.com>
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
They have existed throughout Qt5, and no longer used by any
QIconEngine I am aware of.
Change-Id: Iab0a978be808a60fb82379467e294e2457056bae
Reviewed-by: Morten Johan Sørvig <morten.sorvig@qt.io>
'userInterfaceStyle' is only available on iOS 12.0 or newer [-Werror,-Wunguarded-availability-new]
if (previousTraitCollection.userInterfaceStyle != self.traitCollection.userInterfaceStyle) {
Pick-to: 5.15
Change-Id: Id6340f70019d55ab2a0707b1aebd6d333e9544cb
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
QFile is no longer included implicitly via other headers,
so include it explicitly in the FreeBSD framebuffer code.
This is needed to make it compile (at all, although there
are also API changes to chase and general bitrot).
Change-Id: I3801a39503f545a24a3b2c58122774b5697358e9
Reviewed-by: David Edmundson <davidedmundson@kde.org>
Reviewed-by: Laszlo Agocs <laszlo.agocs@qt.io>
This can happen when there is a uniform block in the Vulkan
shader with an instance name, both in the vertex and fragment
shader. The members are mapped to members of a struct uniform
in GLSL since we do not use uniform buffers there.
Now if there is, for example, "ubuf.opacity" both in the vertex and
fragment shader reflection info, then it's still just one single
uniform in the GL program, using one location. Registering it twice
in our 'uniforms' list is harmless, but wasteful, since it means the
uniform value will be set twice upon each setShaderResources().
Change-Id: Ib646eaec333522560d631b3b81800ef610f09319
Reviewed-by: Andy Nichols <andy.nichols@qt.io>
They are redundant, not in use, and got added without implementation in
a81859a3c8.
Change-Id: Ifed1fbf97a8158c2801df09dac47bf1fc90795d4
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
Certain servers, like smtp.live.com, will send a request for a
certificate even though they don't require one. In Schannel this
manifests as a warning/info status (SEC_I_INCOMPLETE_CREDENTIALS).
In the cases where it's not needed we should suppress the warning and
try to connect anyway, which is done by calling
InitializeSecurityContext again when we get the status.
Pick-to: 5.15
Change-Id: I3c48140f2949d8557251a49a2b66946da9395736
Reviewed-by: Joshua GPBeta <studiocghibli@gmail.com>
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
This reverts commit b0c51f86f4.
The build failure caused by 58c1c6ee5c has
been fixed.
Change-Id: Ic7458d54c7a874588e8b1bfeca61df1842763656
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
This reverts commit 6e1a570dc8.
The build failure caused by 58c1c6ee5c has
been fixed.
Change-Id: I67acaa2d1dd5fc88c205b054252be2e6e26898c7
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
Seems this information is obsolete, get rid of it.
Fixes: QTBUG-86607
Pick-to: 5.15
Change-Id: I0250e32b3c312c7da0363dd1b0d7f676bbfa0115
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
This reverts commit f0873a1c62.
The build failure caused by 58c1c6ee5c has
been fixed.
Change-Id: I5cec78b353089ca024e905795bc99cad423c2d4e
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
...and fix the build errors the original change caused.
This reverts commit 127fb8bb55.
Change-Id: I4006b32734a51c5d101dd73c957f81d2a0f84ba1
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
clang on Windows shows an error:
qtbase/src/gui/text/windows/qwindowsfontdatabase_ft.cpp:122:25: error: loop variable 'key' of type 'const QString' creates a copy from type 'const QString' [-Werror,-Wrange-loop-construct]
for (const auto key : keys) {
^
Change-Id: Icdfa5e2a6753450213dcb96df70247c5a71c2293
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
If we have a platform plugin we ask the platform to quit, and if
not we fall back to the base implementation of QCoreApplication
that sends Quit events directly.
This allows the platform to involve the rest of the system in the
process. The platform will then come back with a spontaneous quit
via QWSI::handleApplicationTermination(), which will then send
the corresponding Quit even from QGuiApplication like normal.
Task-number: QTBUG-45262
Task-number: QTBUG-33235
Task-number: QTBUG-72013
Task-number: QTBUG-59782
Change-Id: I0000aaf7192e4b905933c5da0e53901c6c88f26a
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
Undeclared identifier ssize_t etc. Just use int. Not like we can have
more than a handful of physical devices anyway.
Change-Id: Ie1fb7ab9794a7d39e84db864c2be6dbdd5d97f50
Reviewed-by: Andy Nichols <andy.nichols@qt.io>
In each case, simply give WithBraces as default for mode, since that's
what _q_uuidToHex() used internally.
Task-number: QTBUG-85700
Change-Id: I9f6fddb259703917129d4be742bbdd2eb1647f44
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
These were now always defined, hence redundant.
Leave the #define in place so that we can verify we actually do always
define it, in a #else of an existing #if check on it.
Change-Id: Iea4c3dbc8f9982268bcf81da5ef17fe2ebf5c462
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
This follows up on commit d273076b44
which left a comment asking for the clean-up this finishes.
Task-number: QTBUG-85700
Change-Id: I1c6896a42a09b873302ad7ec8273879f2a4a4ce6
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
Some tests, particularly the asynchronous ones that depend on the
QTRY_*() macros, have call-backs in which a test can fail, but the
macro used to test for failure only returns from the call-back, so the
test doesn't know to fail.
Make sure the QTRY_*() macro gives up if that happens, so that the
test function at least gets control back and can notice that it's
failed. Even if they don't check, they'll fail sooner, where they
might otherwise have been stuck in a loop that would never exit until
the watchdog timer shoots the test down (and Coin ends up with a
debugger back-trace and no output from later tests).
Change-Id: I622a53117de5e97d23dd22e04e5cd20361a54651
Reviewed-by: Mitch Curtis <mitch.curtis@qt.io>
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
QTime() is invalid. Only the seconds and milliseconds are optional,
for valid times. Use startOfDay() for zoned times, QTime(0, 0) for
fixed-offset ones.
Change-Id: I3b65d5c3733ac83dc3a6c214859c3f56a480bb94
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
With 6e8985e357 merged we can now use
generic erase logic provided by array operations. This commit aligns
QString with QList/QByteArray
Task-number: QTBUG-84320
Change-Id: I83e9349e2461afd98737df25613aa2d0fd817a71
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Nothing used defDateTime(), nothing but it used defTime(). The only
other use of defDate() were from one test, which might as well make it
a local variable.
QDateTime's default is invalid, making invalidDateTime() redundant,
and the invalidDate() and invalidTime() it called had no other users.
Change-Id: I0e07ee58478bfe8ba680eafb52e2f73a962edd33
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
The QFlags constructor taking a plain zero literal was deprecated in
commit af2daafde7.
Pick-to: 5.15
Fixes: QTBUG-86585
Change-Id: I2fc68c725ba649218bd9fffd1633d6251649d2bd
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
Apparently these APIs don't have any replacements that we
can use, so we have to continue to depend on them for now.
Silence the compiler warnings related to them, to avoid
unnecessary noise.
Change-Id: I1838e3c82bedd31529fdad8debc577dca17613e5
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
