A number of consequences of the new rhi-based backingstore
composition were not handled. Most importantly, the fact
that RasterGLSurface is not a thing anymore in practice
causes challenges because we can no longer decide just
based on the surfaceType what a QWindow with OpenGLSurface
would be. (a plain GL window or a GL window with a backing
store?) Also, the backingstore needs to be able to initialize
its backing QRhi by itself, because with eglfs going through
OpenGL is the only way.
Amends 68a4c5da9a
Fixes: QTBUG-102750
Change-Id: Ia1ca59d01e3012264a76b50e591612fdcc2a0bd6
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Andy Nichols <andy.nichols@qt.io>
A comment from the patch 6ffc8d8eb6
is a little ambiguous. This patch is just to give information
about the modification.
Task-number: QTBUG-72103
Pick-to: 6.3 6.2 5.15
Change-Id: I6bfc3ae926c118de0d969a4b44f438c24f8d4f72
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
If a user or a delegate asks a view to act on an editor that does
not belong to the view, report a warning instead of silently
ignoring the problem. This erroneous condition can happen for instance
if a user installs the same delegate on multiple views (something
that the documentation says _not_ to do) and the delegate indeed
emits signals related to the editors of one view -- the other views
don't know about that editor.
Change-Id: I2d10582ebb7aefca4acea306b8a57bcc3162050a
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
From the Intel Software Optimization Manual[1], Chapter 11 ("Multi-core
and Hyper-Threading Technology"), offers:
User/Source Coding Rule 14. (M impact, H generality) Insert the PAUSE
instruction in fast spin loops and keep the number of loop repetitions
to a minimum to improve overall system performance.
See section 11.4.2 for an explanation of why this is a good idea.
[1] https://www.intel.com/content/www/us/en/developer/articles/technical/intel-sdm.html#inpage-nav-5
Pick-to: 6.3
Change-Id: I7fb65b80b7844c8d8f26fffd16e94088dad1ceee
Reviewed-by: Sona Kurazyan <sona.kurazyan@qt.io>
qsimd.cpp:400:31: error: no member named 'size' in namespace 'std'
Pick-to: 6.2 6.3
Change-Id: I35544e2937c048e4a466fffd16e97e7a20d9da69
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
It's not necessary, and it breaks the qttools build (where we have a
global variable named 'id'), and thus will most certainly build a lot
of existing user code.
Amends e47c22480f.
Change-Id: I97a91c2cb23fdae65143cf14c81570cf88d529d5
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
Otherwise they will just use default compiler
Pick-to: 6.3
Change-Id: Id5813b99fbbb6b0d8b0ee658e06312b637a097c1
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
It needs fewer moving parts this way. In the process, to match the
QString strValue, rename a QString key to strKey, freeing the name key
for a view.
Change-Id: Ic58145fd665db265d136545ced59e293a2eed1eb
Reviewed-by: Marc Mutz <marc.mutz@qt.io>
In some circumstances, an empty QDBusInterface was created. However,
this can lead to a crash on some systems (observed in Ubuntu 20.04).
The crash happens, when no network is available.
Pick-to: 6.3
Change-Id: I37316db547f33f082b8aaa73494db1bdf5aded1d
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Broken meaning not being functional with shader model 5.
This can happen with certain virtual machines where there is an
accelerated DXGI adapter which then only supports a feature level
lower than 11_0, and so SM 5.0 shaders do not work. Similarly, GPUs
from 2009 and earlier may have a similar driver setup.
We do not particularly care about such devices, however we should make
sure it can be recognized early on (i.e. in create()) that something is
not right and fail the entire QRhi initialization. Otherwise no error
will occur until attempting to create the first shader via a
QRhiGraphicsPipeline and that's way too late.
This way clients such as Qt Quick can make an effort to retry with some
other settings, most notably the PreferSoftwareRenderer flag, in order
to pick the sw rasterizer based adapter (WARP).
Task-number: QTBUG-78648
Change-Id: Ia4a3a0bc1a09e2864d426d4d55978dc6f759127c
Reviewed-by: Andy Nichols <andy.nichols@qt.io>
swapchain->currentFrameRenderTarget()->renderPassDescriptor() is not
functional at the moment, it returns null. This is because no backend
ensures that the internal renderpass descriptor object is exposed via
that getter in a QRhiSwapChainRenderTarget. Whereas in a
QRhiTextureRenderTarget this would work by design because there the
setter must be called by the user.
Fix this up, providing better API symmetry, and also reducing the need
to pass along QRhiRenderPassDescriptor objects seprately alongside a
QRhiRenderTarget in some places, e.g. in Qt Quick.
Change-Id: I42c4e9aaee3202c1d23bd093d840af80c5f8cd0f
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Andy Nichols <andy.nichols@qt.io>
Required for porting away from QLatin1Char/QLatin1String in scope of
QTBUG-98434.
Change-Id: I58e00618fc8c82fad8d35a66c1d46e459bfc9de6
Reviewed-by: Marc Mutz <marc.mutz@qt.io>
macdeployqt doesn't know or care about universal build (yet), and
will just scan all the lines produced by otool -L as if they were for
a single architecture, ignoring any lines it doesn't recognize.
This fails when the second arch starts off with a reference to the
binary we're processing, which then gets interpreted as a dependency
of itself. If that binary doesn't have a path, we end up looking for
a match in /usr/lib, which we can't find in the best case, or actually
end up deploying from an incompatible Qt version in the worst case.
Ideally we'd teach macdeployqt about the multiple slices, bur for now
we work around it by skipping any references to dependent libraries
that match the binary itself.
Fixes: QTBUG-102607
Pick-to: 6.2 6.3 5.15
Change-Id: Ia039fafdce90896e61aab90b9c8ac52543cc2c8c
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
Reviewed-by: Morten Johan Sørvig <morten.sorvig@qt.io>
Assorted QMetaType and QVariant methods deprecated in Qt 6 were still
in use - the deprecation warnings were suppressed by setting
QT_DISABLE_DEPRECATED_BEFORE=0 - fix them and clear that suppression.
Change-Id: I1aa8f45dcb5a18449b060b346c80ad70536896ab
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Retain the macro, so as to automate checking
QTest::currentTestFailed() after return from the lambda.
Change-Id: If5873a533224ea3b809cef02255642a7d071d642
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Give the macro and its parameters clearer names, wrap the macro body
in do {...} while (0), use QCOMPARE() rather than QVERIFY()ing
comparisons, convert the result type to QMetaType::Type rather than
QVariant's deprecated equivalent, save one macro parameter by using
decltype() on another. Simplify one check by pulling out a common
sub-expression as a local variable.
In the process, give the settings keys meaningful names rather than
just "key" with a numeric suffix.
Ideally this test would be reworked to be data-driven, instead of
using macros, but the use of qvariant_cast<> to convert a read-back
value, along with the variable type of the input value compared to the
result of that cast, precludes this (as far as I can see).
Change-Id: I4e083cd0dea14d723910873829352e15740006bd
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Move the large block unrelated to the macro-driven parts to before the
macro is defined. This makes it clearer that it has nothing to do with
the macro-driven part of the test. This is, in any case, a more basic
test of behavior, so may as well go first. Rename the tested key in
the process.
Change-Id: I783fcd812e36320e9bb9765bab3ec09b56d6e6e6
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Move its static populateWithFormats to before the class is declared
and make the _data() methods that call it (including one that did so
via another) inline so that we can see which are this simple - and
thus which aren't.
Change-Id: I71863244ba0e4bd188424b6a3f8d86d5d9f635a9
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
The functions it defines depend on many many types in QtCore for which
it did not have a #include; both files that included it thus had to
pull in QtCore to compile. Put that #include where it belongs and
clean out many specific QtCore includes that it makes redundant.
Change-Id: Ie9d9ec325d4879d771cb14baecb06fecbdaf62c5
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Although it loses type information for numeric types, it does preserve
the type of most other values.
This amends commit ee8a5f4917
Pick-to: 6.3 6.2 5.15
Change-Id: Id4d0e11eabff1874869b940c05055e0bebef5963
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Since we set android:extractNativeLibs to "true" in AndroidManifest.xml,
a warning is thrown by the gradle build, which needs the property
packagingOptions.jniLibs.useLegacyPackaging to be set to true.
Pick-to: 6.2 6.3 5.15
Change-Id: I725d2663d68ab88abfda630f8a61666510043967
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Ramon Sadornil <ramon.sadornil@qt.io>
Reviewed-by: Ivan Solovev <ivan.solovev@qt.io>
Some missing values has to be defined in AndroidManifest.xml, otherwise,
warnings might be thrown.
* android:exported="true": because the manifest sets an intent-filter,
and it then has to explicitly to avoid the warning [1].
* android:allowBackup="true": this has to be explicitly set, we set it
to the default value here [2].
* android:fullBackupOnly="false": SDK 23+ use this to deteremine to
user auto backup or not, we set it to the default value here [3].
[1] https://developer.android.com/guide/topics/manifest/activity-
element#exported
[2] https://developer.android.com/guide/topics/manifest/application-
element#allowbackup
[3] https://developer.android.com/guide/topics/manifest/application-
element#fullBackupOnly
Pick-to: 6.2 6.3 5.15
Fixes: QTBUG-101320
Change-Id: I0872dc00e48a867154ec9ded26620383fb747918
Reviewed-by: Nicholas Bennett <nicholas.bennett@qt.io>
Required for porting away from QLatin1Char/QLatin1String in scope of
QTBUG-98434.
As a drive-by, fix qsizetype -> int narrowing conversion warnings for
the touched lines.
Change-Id: I133b80334b66e0a5ab9546dd8e1ff0631e79601e
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
Replace the uses of QLatin1String with corresponding string literals
where applies, or with QLatin1StringView.
As a drive-by, fix formatting of a warning message in
qopenglengineshadermanager.cpp.
Task-number: QTBUG-98434
Change-Id: I12f4479637a8fe265138f240a1310d2d576454a7
Reviewed-by: Marc Mutz <marc.mutz@qt.io>
WM_SETTINGCHANGE may be triggered in many different reasons,
we don't have to verify whether the system changes the global
theme or not in all cases. Although not officially documented,
there's a widely known and used technique to detect whether the
user actually changes the personalize settings or not, that is
when wParam is 0 and lParam is "ImmersiveColorSet", this
combination indicates system's personalize settings has been
changed. We can get rid of most unneeded verify of system theme
by only execute the logic in this specific case.
Change-Id: Iaf934c29975b3b2090fd692776f80b1125d3ddb3
Reviewed-by: Oliver Wolff <oliver.wolff@qt.io>
CUPS 2.2 adds accessible IPP printers to the list of destinations
that can be used. The "printer-uri-supported" option will be
present for those IPP printers that have been recently used.
https://github.com/OpenPrinting/cups/blob/master/cups/dest.c#L1611
Pick-to: 5.15 6.2 6.3
Change-Id: I49b4d2f1ee6d96e9bcc04ed1cc2ed48d7a6e7563
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
GCC 12 doesn't like that we're calling preparse() on the empty
QByteArray() (whose .data() == &QByteArray::_empty), despite the
condition being impossible.
Reported at https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105348
In function ‘void* qt_cbor_decoder_read(void*, void*, size_t, size_t)’,
inlined from ‘void* read_bytes_unchecked(const CborValue*, void*, size_t, size_t)’ at cborinternal_p.h:234:51,
inlined from ‘uint32_t read_uint32(const CborValue*, size_t)’ at cborinternal_p.h:271:25,
inlined from ‘uint64_t _cbor_value_decode_int64_internal(const CborValue*)’ at cborparser.c:333:23,
inlined from ‘uint64_t _cbor_value_extract_int64_helper(const CborValue*)’ at cbor.h:372:50,
inlined from ‘void QCborStreamReader::preparse()’ at qcborstreamreader.cpp:769:59,
inlined from ‘QCborStreamReader::QCborStreamReader(const QByteArray&)’ at qcborstreamreader.cpp:825:13,
inlined from ‘QCborStreamReader::QCborStreamReader()’ at qcborstreamreader.cpp:790:37,
inlined from ‘QtPrivate::QMetaTypeForType<QCborStreamReader>::getDefaultCtr()::<lambda(const QtPrivate::QMetaTypeInterface*, void*)>’ at qmetatype.h:2302:65,
inlined from ‘static constexpr void QtPrivate::QMetaTypeForType<QCborStreamReader>::getDefaultCtr()::<lambda(const QtPrivate::QMetaTypeInterface*, void*)>::_FUN(const QtPrivate::QMetaTypeInterface*, void*)’ at qmetatype.h:2302:20:
qcborstreamreader.cpp:706:18: error: array subscript ‘unsigned int[0]’ is partly outside array bounds of ‘const char [1]’ [-Werror=array-bounds]
qbytearray.h:92:23: note: object ‘QByteArray::_empty’ of size 1
Pick-to: 6.2 6.3
Change-Id: If05aeeb7176e4f13af9afffd16e8485d34e36100
Reviewed-by: Ulf Hermann <ulf.hermann@qt.io>
The algorithm in QTextEngine::bidiReorder() doesn't like an empty level
and visual order list. Found by GCC 12's slightly misleading warning:
qtextengine.cpp:3838:29: error: ‘levels’ may be used uninitialized [-Werror=maybe-uninitialized]
qtextengine.cpp:1174:6: note: by argument 2 of type ‘const quint8*’ {aka ‘const unsigned char*’} to ‘static void QTextEngine::bidiReorder(int, const quint8*, int*)’ declared here
qtextengine.cpp:3835:28: note: ‘levels’ declared here
3835 | QVarLengthArray<uchar> levels(nItems);
| ^~~~~~
Pick-to: 6.2 6.3
Change-Id: If05aeeb7176e4f13af9afffd16e8490f131f160e
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
GCC is right that symbolsGroup2 could remain uninitialized for some
conditions. So always initialize it.
qxcbkeyboard.cpp:284:48: error: ‘symbolsGroup2’ may be used uninitialized [-Werror=maybe-uninitialized]
Pick-to: 6.2 6.3
Change-Id: If05aeeb7176e4f13af9afffd16e84af7d7806b0d
Reviewed-by: Zhang Hao <543985125@qq.com>
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
We're already tracking the last view along with the last event, so
use that instead of always assuming the drag started from the window's
content view.
Fixes: QTBUG-102584
Pick-to: 6.2 6.3
Change-Id: Ibe8506b0f91ecdfe6e7fe91e17c129b4d7552880
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
Reviewed-by: <stefan.wastl@native-instruments.de>
In the sequence for the early 32-bit time_t leap years, I somehow
managed to get two of the entries swapped. This amends
commit 1f4b237dad
Pick-to: 6.3
Change-Id: I5cb9080f8c6a24e2163508e640747c6fbbe93dbd
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
In the same ways as for blacklisting. In the process, move reporting
of an unrecognised tag outside the global-data loop, as the tag might
be, or start with, a global data tag; and also report the global data
alongside the test-specific data. Indent the lines reporting tags, as
this makes the structure of the lists more evident. Converted the tag
argument to QLatin1String() to facilitate searching and matching.
Updated documentation, including expanding on the option of specifying
the data tags on the command-line. Drive-by: fixed a link in the
documentation, removed a snippet that was nowhere used.
[ChangeLog][QtTest] When specifying what tests to run on a QtTest
command-line, it is now possible to include a global data tag in the
same way as a function-specific data tag or in combination with it.
Thus if the test reports itself as tst_Class::function(global:local),
command-line option function:global:local will select that specific
test and function:global will run all data-rows for the function on
the given global data tag.
Fixes: QTBUG-102269
Change-Id: I7c86d6026933b05f7994bfc3663a2ea6a47f5f38
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
The compositor should not show the resize cursor for fullscreen and
maximized windows, since windows in those states aren't resizable.
Pick-to: 6.3
Change-Id: I58762b7729a02f9a3c78edafe5d9d5b364fb552f
Reviewed-by: David Skoland <david.skoland@qt.io>
Reviewed-by: Lorn Potter <lorn.potter@gmail.com>
The canvas cursor should always be restored to "default" if no cursor
is set, otherwise we get the I-beam cursor since the canvas is
contenteditable.
The code which handles setting and restoring the override cursors
handles this correctly, however we also need to set the default
htmlCursorName for cases where setOverrideWasmCursor() is not called.
Pick-to: 6.3
Change-Id: I23be7847d3a30ab78e3a8c9ec402285246bfefec
Reviewed-by: David Skoland <david.skoland@qt.io>
Reviewed-by: Lorn Potter <lorn.potter@gmail.com>
If Qt is configured with -platform linux-g++-32 , make sure
to add the -m32 compile options for all built targets.
On 64 bit host OSes that provide both 32 and 64 bit libraries
we need to exclude the 64 bit libraries from being picked up.
The locations of the libraries are distro-specific.
This change by default excludes the Ubuntu x86_64 libraries
paths.
Opt outs are provided, which when used, forces Qt builders to
specify their own ignore paths in a custom CMake toolchain file.
The compile option and default path exclusions are added to the
Qt-generated CMake toolchain file as well, so they are reused
when building other Qt repositories.
Note that there is no foolproof way to tell CMake to ignore all
x86_64 packages / libraries, even if CMake 3.23
CMAKE_IGNORE_PREFIX_PATH is used, because there might not be
a single sysroot to exclude.
Both x86 and x86_64 libraries can co-exist in the same sysroot,
e.g in /usr
One would have to list each package / library directory in
CMAKE_IGNORE_PATH manually.
Additionally, the PKG_CONFIG_LIBDIR environment variable is also set
to Ubuntu specific prefixes, to ensure that pkg_check_modules ->
pkg-config don't pickup x86_64 libraries.
Fixes: QTBUG-101963
Change-Id: Ib17c8d2cd0ba33b2cf748772245bcd558de9120c
Reviewed-by: Jörg Bornemann <joerg.bornemann@qt.io>
Include new CMake and qmake snippets and reorganize some of the sections
structure.
Task-number: QTBUG-100369
Pick-to: 6.3
Change-Id: I47cc9b24818c26740e49e6365bf881874942b79c
Reviewed-by: Topi Reiniö <topi.reinio@qt.io>
Add the new qdocinc cmake and qmake snippets and reorganize some of the
sections structure.
Task-number: QTBUG-100369
Pick-to: 6.3
Change-Id: Iafa31856500e458c14feeb92bee660f44a65ffa9
Reviewed-by: Topi Reiniö <topi.reinio@qt.io>
Move some of the detailed information into its own overview page and
reorganize the contents structure. Also, escape the _ character.
Task-number: QTBUG-100369
Pick-to: 6.3
Change-Id: I396877912bed66537e4a97ee3d109ff92e15d843
Reviewed-by: Topi Reiniö <topi.reinio@qt.io>
clang compiler recently got smarter and detects "pseudo used variable"
patterns where we declare a variable and only use it in self-increments,
self-decrements and other similar expressions
Errors:
qtbase/src/corelib/text/qlocale.cpp:3898:9: error: variable 'group_cnt' set but not used [-Werror,-Wunused-but-set-variable]
int group_cnt = 0; // counts number of group chars
^
qtbase/src/corelib/text/qunicodetools.cpp:1372:21: error: variable 'uc' set but not used [-Werror,-Wunused-but-set-variable]
const char16_t *uc = text + from;
^
and more of the kind
Remove the ones that have no usage, mark others with [[maybe_unused]]
Pick-to: 6.3 6.2
Change-Id: Ib2d0722110e3da8c39e29ec78c0ec290d064c970
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
