Some classes have circular dependencies:
* RuntimeIteration depends on RuntimeTask
* RuntimeContainer depends on RuntimeIteration
* RuntimeTask depends on RuntimeContainer
* TaskTreePrivate depends on RuntimeTask
* RuntimeContainer depends on TaskTreePrivate
Although there are a few std::unique_ptr deployed, the code at the
moment does not fully solve the problem: the classes have an inline
destructor, and the dependency has only been forward declared, resulting
in a build error.
Make some destructors out-of-line, and define them (as defaulted) late
enough so that all the graph of classes has been seen, in order to fix
the build.
Change-Id: If150b92154061e915ebac920806717d2288c0638
Reviewed-by: Jarek Kobus <jaroslaw.kobus@qt.io>
(cherry picked from commit 71e53405a309684e4adcb8669a5f3d04319ad767)
Reviewed-by: Qt Cherry-pick Bot <cherrypick_bot@qt-project.org>
(cherry picked from commit 3fda22bd8a45bfa5877994962af67e2b6b753043)
As a drive-by replace QList::append and operator<< with emplace_back,
to avoid repeating value_type.
Task-number: QTBUG-115841
Change-Id: I6fb790f959b41b0feb49e3a0f0b6b3631e24a232
Reviewed-by: Marc Mutz <marc.mutz@qt.io>
(cherry picked from commit 71d114588d9312e89195a32357ec402fc22b062e)
Reviewed-by: Qt Cherry-pick Bot <cherrypick_bot@qt-project.org>
(cherry picked from commit 2b2271bea3f94a72adf73aa04395a6b20c404f0e)
The class declares a protected destructor, which interferes with the
RO5. The only subclass in qtbase is move-only; QTextCodec (in
Core5Compat) is not even movable.
Make QStringConverterBase move-only as well.
Change-Id: I20d16df79f7ad8e7e7a8af339954cfa2d1482e96
Reviewed-by: Marc Mutz <marc.mutz@qt.io>
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
(cherry picked from commit eb533c81b8aa55f89605bb1d091afe4df4db763c)
Reviewed-by: Qt Cherry-pick Bot <cherrypick_bot@qt-project.org>
(cherry picked from commit 72d0b8f709c0c73fac6062b5a287ef5630f9b770)
Since the removed API uses QDebug. Requested in code review.
This is needed for a subsequent commit (removing qdebug.h from public
headers when possible), but I split out to its own commit to ease
backporting.
Change-Id: I76980634b9ebbae7cf010d99f5095aa94a825824
Reviewed-by: Marc Mutz <marc.mutz@qt.io>
(cherry picked from commit c0fcb2db242a40cdcc2bbfd00332390b64cab439)
Reviewed-by: Qt Cherry-pick Bot <cherrypick_bot@qt-project.org>
(cherry picked from commit 92afa1eaf8783355a2057849e0614342cc4306d4)
COLRv0 fonts are both scalable and in color, which was not
previously handled in the Freetype backend because it only
supported color fonts with embedded bitmaps. This patch
simply correctly detects the scalable color fonts and
render them through the same mechanism as other color
fonts.
[ChangeLog][Freetype] Added support for COLRv0 format color
fonts.
Task-number: QTBUG-131116
Change-Id: I290e7525c24b4cf2ec872a366e5193f54712a3a4
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
(cherry picked from commit 39df9e1858a4115bc19b6a4dee5d687c5ee00d21)
Reviewed-by: Qt Cherry-pick Bot <cherrypick_bot@qt-project.org>
(cherry picked from commit 37c389c779188e05bb057d2f7164cba2827d33b6)
QShortcut has a constructor that takes a _parent_ object and a
function to invoke, and it will internally connect its activated()
signal to that function. This connection will not provide a
context object (the parent is not used to this purpose).
Given it's not clear what the context should be if any (the parent?
`this`?), I'm not at ease at just supplying "something" as the context.
Instead, hide the constructor if we're under QT_NO_CONTEXTLESS_CONNECT.
There's another constructor where the caller can supply a context.
Change-Id: I21befcc86eeb07e65eadee1222757ecc5bab1d2c
Reviewed-by: Richard Moe Gustavsen <richard.gustavsen@qt.io>
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
(cherry picked from commit ade3bb1c5224a9155a2a2f965fa70740aa9ec8a7)
Reviewed-by: Qt Cherry-pick Bot <cherrypick_bot@qt-project.org>
(cherry picked from commit c82464ea754e0345859abb54f31b636541669f0b)
Cleanup CE_ItemViewItem branch in drawControl():
- use local rect instead rect, vopt->rect or option->rect for
consistency
- move out viewItemPosition comparisons into local booleans
- fix coding style
- merge two QPainter::drawLine() calls into one
Task-number: QTBUG-131585
Change-Id: I7cd2f96a2a733a55d5e7fb5956c59bd043317cb2
Reviewed-by: Wladimir Leuschner <wladimir.leuschner@qt.io>
(cherry picked from commit 44edab7c84be101025dc6f9a74f87944ea3a22e9)
Reviewed-by: Qt Cherry-pick Bot <cherrypick_bot@qt-project.org>
(cherry picked from commit 37effa5dd3b7345ceff38f55ce5b563d2748962c)
When calling insertRows(), the model of the child items was not
correctly updated. insertRow() is behaving correctly.
Fixes: QTBUG-131372
Change-Id: I8b6aef7ab97887c6eb46eb21b992d76e4d59b3a0
Reviewed-by: David Faure <david.faure@kdab.com>
(cherry picked from commit 5bab8448feb2661f871e71d71a95afaffcb1656a)
Reviewed-by: Qt Cherry-pick Bot <cherrypick_bot@qt-project.org>
(cherry picked from commit 3536eb32bf7e6b177fe600256dc9efd0ff653e36)
The progressbar content calculation assumes that
subElementRect(SE_ProgressBarContents) returns a rect at 0/0 which is
not the case for progressbars in rtl mode. Fix it by moving the topLeft
edge to 0/0 and translating the painter accordingly
Task-number: QTBUG-132459
Change-Id: I614589c9094b0da3c02867b45570d4dfdca61a18
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
Reviewed-by: Oliver Wolff <oliver.wolff@qt.io>
(cherry picked from commit 314a5b972646a4f2ef9c1f49ddfd335f546336ab)
Reviewed-by: Qt Cherry-pick Bot <cherrypick_bot@qt-project.org>
(cherry picked from commit bf535c4c66b8d39e7c9a067155bdf2f6ea26a924)
Mouse tracking is needed for windows 11 style since it is using
QStyle::State_MouseOver to draw the cells.
Fixes: QTBUG-129242
Change-Id: Ib79ac5a8c14ac8c70365e87120f80b4e4ee4c5b9
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
Reviewed-by: Oliver Wolff <oliver.wolff@qt.io>
(cherry picked from commit 85333b3d8cb89bf58c6c6c0db7c3db5c95b4cbfc)
Reviewed-by: Qt Cherry-pick Bot <cherrypick_bot@qt-project.org>
(cherry picked from commit 5ff8c2eb00aedc3fec76fe63e4914a6b4e5f32bd)
There is no need to search for high-res svg icons - a svg scalable by
design. Also the QSvgIconEngine does not support it.
Change-Id: Ib76c887900ce77268e9c0253cb697b8de9e7f10e
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
(cherry picked from commit c64774eef72411029f7c2fb611328f25b83ff8ca)
Reviewed-by: Qt Cherry-pick Bot <cherrypick_bot@qt-project.org>
(cherry picked from commit 2bb83f6b16d08d2c30bc6fbb226ad8033f312e88)
Even with SHGSI_LINKOVERLAY flag set, loaded Icon with
SHDefExtractIcon doesn't have any overlay, so we draw
it manually.
Fixes: QTBUG-131843
Change-Id: Iae4c2da12104a361b9a8cf05c78adcdb698f1f82
Reviewed-by: Christian Ehrlicher <ch.ehrlicher@gmx.de>
(cherry picked from commit fd7bc16e9fbdc63bd22ba90d0c20b36ccffd2bae)
Reviewed-by: Qt Cherry-pick Bot <cherrypick_bot@qt-project.org>
(cherry picked from commit 59861df8ce4c7c69cdf1c7a8738eeac81ff2d8c2)
In windows Backtab key event always comes as "Shift+Backtab".
So in order for Backtab key to be recognized as a shortcut sequence,
we also consider Backtab without shift modifier a possibility.
Fixes: QTBUG-94890
Change-Id: I20a7b404b57d8df5bea23765257a178f2e098ed0
Reviewed-by: Oliver Wolff <oliver.wolff@qt.io>
(cherry picked from commit 4d9f99c4ea70cdae58c80e1bfdbabefe3abb455f)
Reviewed-by: Qt Cherry-pick Bot <cherrypick_bot@qt-project.org>
(cherry picked from commit 6d8afaac84f995cfb008bd2d3eb59fe4fe10dd69)
Rework fullscreen and maximized/expanded states handling
by simplifying and re-organizing the code, removing some
unnecessary code. Also, use newer APIs and handling the
cutout regions.
For expanded mode, use transparent instead of translucent
so that the user can decide what color to use if needed,
and in any case using the translucent flags is deprecated.
You might still notice some artifacts as in QTBUG-88676,
a fix for that is outside the scope of this patch.
When going off of fullscreen mode one some cases you might
notice a white/black black at the bottom and that's because
QtRootLayout.onSizeChanged() is reporting wrong available
size which is also an existing issue and outside of this
scope.
Fixes: QTBUG-96105
Fixes: QTBUG-101968
Fixes: QTBUG-127394
Fixes: QTBUG-121820
Task-number: QTBUG-109878
Task-number: QTBUG-119594
Change-Id: I586775a1d0414ec0adbc968d50b9c1a1ce466422
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
(cherry picked from commit 0abcb9bef8a7cb85df006adfed51bc9258868ed2)
Reviewed-by: Petri Virkkunen <petri.virkkunen@qt.io>
Move calls for handling system UI visibility to QAndroidPlatformWindow
where they belongs.
Change-Id: I3802cf9d205ee6678f71b787c5ea4804d3aaeb29
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
(cherry picked from commit a2385e5c9dc9b07cb43264bf0309a80992f80ad8)
Reviewed-by: Petri Virkkunen <petri.virkkunen@qt.io>
No need to also check for Qt::WindowMaximized, if we're using
Qt::ExpandedClientAreaHint, we would expect the full size to
be used.
Fixes: QTBUG-110878
Change-Id: I52db0224aa96409a8f2d984e01192001f9b42440
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
(cherry picked from commit 10805e7d8af0bcbcba12bc896a01d645e1648741)
Reviewed-by: Petri Virkkunen <petri.virkkunen@qt.io>
Java methods of QtAbstractItemModel may get called from Qt GUI thread or
Java main thread. Situation is the same with internal proxy methods,
where they may directly or indirectly called from these two thread.
At the same time, a chain of calls may happen from these threads, so
some functions may get called indirectly from a single caller thread.
In this change, we are adding a map of QObject* to QRecursiveMutux
where the QObject is a QAbstractItemModel, either the proxy instance
or a reference to an underlaying QAbstractItemModel when the model
is implemented in C++.
Each call to any function locks the QRecursiveMutux instance associated
with the QAbstractItemModel instance. A connection from
QObject::destroyed to a lamda, helps to remove the map entry when
the model is destroyed.
Fixes: QTBUG-127467
Change-Id: I18dcdc207e201b3883526648b4e672e5868ed8ab
Reviewed-by: Assam Boudjelthia <assam.boudjelthia@qt.io>
(cherry picked from commit bf76387507dde71063ebe484bce1e7d1b8e09bc2)
Reviewed-by: Qt Cherry-pick Bot <cherrypick_bot@qt-project.org>
(cherry picked from commit 2b14b70b435fa1120dee3f08a1377aa26016d2b5)
The call to q_assert() isn't constexpr, so can't happen in a constexpr
function. Most compilers shrug this off as they can see the code
actually is unreachable, but apparently debug builds on QNX can't work
that out. Since nothing actually needs the function to be constexpr,
remove that qualifier. In the process, move the offset-case, which is
also unreachable, to the end of the switch, closer to the unreachable
end of the function (since I failed to spot that there were two
unreachable parts at first).
Fixes: QTBUG-132698
Change-Id: Ibdde7abd91d284e1c1af53e2b5e142fc234400a8
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Marc Mutz <marc.mutz@qt.io>
(cherry picked from commit 2578d72a7cb9b784a00efb04df6a31496d92dcdc)
Reviewed-by: Qt Cherry-pick Bot <cherrypick_bot@qt-project.org>
(cherry picked from commit cf9e8434eb1d6cb527f8f216c4c37c14cf713690)
Tool dependencies that are added using the
qt_record_extra_package_dependency function now are considered as
the third party dependencies. This allows using two different
approaches when looking for the dependencies of Qt tools.
All dependencies that are Qt tools now go to the path invented in
035fbd068b5a3fbc18b7868ecac9a6a6a2f6602c and use the
_qt_internal_find_tool_dependencies function which is designed to
look for Qt packages and uses CONFIG signature of find_package.
To look for the non-Qt modules we should use more generic find_package
signature and _qt_internal_find_third_party_dependencies fits perfecly
for these purpose.
Note that it's known issue that the
_qt_internal_find_third_party_dependencies command also may find
"target" tools when crosscompiling, but this commmit doesn't reinvent
it. We already have various places which make the scoped "host"
dependencies lookups to ensure that "host" tools look for the "host"
dependencies in build systems like yocto.
Ammends 035fbd068b5a3fbc18b7868ecac9a6a6a2f6602c
Task-number: QTBUG-132340
Fixes: QTBUG-132622
Fixes: QTBUG-132616
Pick-to: 6.5
Change-Id: I9effba3d405ceec720a8a2a332250619cd56f598
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
(cherry picked from commit f881e06dd44c377772dd3b99793ab7552c7cd64e)
Reviewed-by: Qt Cherry-pick Bot <cherrypick_bot@qt-project.org>
(cherry picked from commit df5c9259af2e915e2ae8caa6122456e9f067614b)
The qthreadstorage.h header contains two implementations controlled by
the QT_CONFIG(thread) definition.
Commit 8553ffd8d147ecf6a713d12a360027b477dd59c8 changed the positions
of the namespace macros in the header, but didn't consider the fact
that we now include std and Qt headers inside QT_NAMESPACE in the
!QT_CONFIG(thread) case.
Fix it by moving all the includes to the top, and wrapping them in
the needed QT_CONFIG() checks. As a drive-by, drop the unused
type_traits include.
Found in Qt 6.9 API review, but picking down to 6.5, as that's what
the original commit did.
Note that 6.5 cherry-pick will require to replace <memory> with
QtCore/qscopedpointer.h because
6da1f72311b844b2232da3067ad6e1e24614e67c was never picked to that
branch.
Pick-to: 6.5
Change-Id: Ic9c5fe82df38a3bd64829c016de8dcbd9f060315
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
(cherry picked from commit 4a6ef7fc6881c7b5ab0b9f6d51a2738ce86cca93)
Reviewed-by: Qt Cherry-pick Bot <cherrypick_bot@qt-project.org>
(cherry picked from commit be31cbfce87294a6e59e472bd7efa96dfb7a40a4)
New languages added with v46
- Kara-Kalpak
- Swampy Cree
Several new Chinese-language locales have been added, including one
using Latin script, which invalidated some prior QLocale tests, which
have been adjusted to fit.
Some obsolete time-zone identifiers are now treated as deprecated
aliases. These have lost their AnyTerritory association, implying
changes to QTimeZone tests.
Many redundant likely sub-tag rules for unspecified language have been
dropped, in favor of simpler rules.
[ChangeLog][Third-Party Code] Updated CLDR data, used by QLocale, to
v46.
Task-number: QTBUG-130877
Change-Id: I92cf210422c7759dd829a7ca2f845d20e263d25b
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
(cherry picked from commit e316276b76b9c3768ca4e19a04d03308ef21fe12)
Reviewed-by: Qt Cherry-pick Bot <cherrypick_bot@qt-project.org>
(cherry picked from commit 9413c19cc1f394bc39a9f46d7d12a71fb42c8d1a)
QUniqueHandle is intended for handle types or opaque pointers that are
no-throw constructible, copyable, and assignable, and has nothrow
comparison and relation operators. This patch ensures that this holds
true, and makes most of the QUniqueHandle noexcept specifiers valid.
This will make it easier to implement the comparison and relation
operators in terms of the Qt comparison framework.
Task-number: QTBUG-132507
Change-Id: Id9fc3fac72a2337541e49c3b4613b8fd082796af
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
(cherry picked from commit 49f1877bbbb8f96d81144861558bdcc81e758a58)
Reviewed-by: Qt Cherry-pick Bot <cherrypick_bot@qt-project.org>
(cherry picked from commit bded4bd5fcb319eb1e95f61a3b6bb80f555279a8)
This removes boilerplate and the extra complication of std::addressof.
Task-number: QTBUG-132507
Change-Id: I07091ec0ac526975cf55361a9811fad77eb152c2
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
(cherry picked from commit 30ae70a110ee2460744e4fb78bae29b75effbf72)
Reviewed-by: Qt Cherry-pick Bot <cherrypick_bot@qt-project.org>
(cherry picked from commit 11abd8704650cf09d8259adc58fa7f7a2ccaa2ed)
So that the pthread_key is destroyed much later than now, at QtCore
unload time instead of a bit earlier in exit(). That's important so it
runs later than the Q_DESTRUCTOR_FUNCTION for QLibraryStore
(qlibrary.cpp), which is what unloads plugins, whose unload-time
destructors may be attempting to quit() & wait() for their worker
threads.
I had to keep the destruction of the exiting thread's adopted thread
where it was: a function-local static's destructor. As noted in the
comment added by 1da7558bfd7626bcc40a214a90ae5027f32f6c7f, we need
QThreadData::cleanup() to run before the libraries start to unload
because we need to delete the event dispatcher.
Init priority 10 will place it after the QHash random seed and qsimd
initialization, on ELF systems and on Windows. On other systems (incl.
Apple ones), the order probably depends on the order of the files in the
CMakeLists.txt. I thought that would mean the destructor for thread/
would run before plugin/ but that apparently is not the case. But that's
fragile.
Amends 1da7558bfd7626bcc40a214a90ae5027f32f6c7f, which amended other
things...
Fixes: QTBUG-132697
Task-number: QTBUG-102984
Task-number: QTBUG-132381
Change-Id: Id7263d6ffe7a5949cd84e35d942ad0e02df1b455
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
(cherry picked from commit 2f69a05bd0cd7ce63890f709ff3ed7a4f78acd70)
Reviewed-by: Qt Cherry-pick Bot <cherrypick_bot@qt-project.org>
(cherry picked from commit 6ce164b3e0e13f3eb376baa4bdb467f9bb331fbb)
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
Amends 65093a84c2b94b1543fd4593bc45d491951d28d4, which changed how we
destroyed the main thread's QThreadData. This merges the call to
destroy_current_thread_data() for both types of Unix systems: those with
broken thread_local destructors and those with working ones. It turns
out that the function got called too early for us in those working
systems (see updated comment).
The clean up of the QThreadData is split into two different mechanisms:
* for any auxiliary thread, when it exits, PThread will call back to
destroy_current_thread_data()
* for the thread that called ::exit(), PThread won't, but ::exit() will
invoke set_thread_data()::TlsKey's destructor
This is different from the situation that existed prior to commit
65093a84c2b94b1543fd4593bc45d491951d28d4: first, there's no code in
qcoreapplication.cpp for this (all in qthread_unix.cpp). Second one may
call ::exit() from any thread, whether that is the thread that called
main(), the thread Qt thinks is theMainThread, or any other.
This commit moves the tst_QCoreApplication check for no extant objects
to a new test. I've chosen to add a new test instead of running a helper
binary via QProcess because we do have a couple of !QT_CONFIG(process)
platforms in the CI, and this is too important.
Credit to OSS-Fuzz for finding this, though it is not itself a fuzzying
problem (all tests of a given structure were crashing on exit).
Fixes: QTBUG-132381
Task-number: QTBUG-130895
Task-number: QTBUG-129927
Task-number: QTBUG-129846
Task-number: QTBUG-130341
Task-number: QTBUG-117996
Change-Id: Ie294dce7263b4189f89ffffd9155ec71d31b89d9
Reviewed-by: Robert Löhning <robert.loehning@qt.io>
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
(cherry picked from commit 1da7558bfd7626bcc40a214a90ae5027f32f6c7f)
Reviewed-by: Qt Cherry-pick Bot <cherrypick_bot@qt-project.org>
(cherry picked from commit 6dc19b55fc6d6dd3198b744f713101cf3b77c0c8)
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
This reverts commit c3a2b9f35a.
Unfortunately, some of the log messages happen WAY too late at process
exit time, when QtCore is unloading. In particular, the worst offender
is the one in QThreadPrivate::finish(), which is called after the last
QObject in the thread that called ::exit() was getting destroyed. This
implies we have Static (De)Initialization Order Fiascos happening (the
logging recurses into qlocale.cpp and the defaultLocalePrivate global
static has already been destroyed).
I need to revert this commit in order to fix QThreadData destruction.
The functionality is welcome back in QtCore once the fixes are in, so
long as whoever does it fixes the issues we're seeing in the CI (and I
can't reproduce on my development machine). I've created QTBUG-132429 to
track the reversal and see if the functionality can be brought back in.
Task-number: QTBUG-120124
Task-number: QTBUG-132429
Change-Id: Iecf8f14529c7a2bb2185fffdfd328066098826b1
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
(cherry picked from commit 357351b7ab8ab2eee865d2449ffcca9c9f502fd3)
Reviewed-by: Qt Cherry-pick Bot <cherrypick_bot@qt-project.org>
(cherry picked from commit 60eaa42a2591b0f4fdd1f8352dea2f5002b4eb28)
This looks like another case of a base class with a protected, defaulted
destructor, which interferes with RO5. Explicitly redeclare the other
special member functions in order to suppress the warnings coming from
the deprecated copies.
Change-Id: Icf8c8817ef001d0503683d071c4f004db64caa2d
Reviewed-by: Marc Mutz <marc.mutz@qt.io>
(cherry picked from commit f6e99f347d243749af8d9beacd7c68852f0ff636)
Reviewed-by: Qt Cherry-pick Bot <cherrypick_bot@qt-project.org>
(cherry picked from commit 25296f86d6b1394035384c5aced0dbfaf0bdfd86)
Flat buttons were larger than normal buttons - therefore also add
marginsRemove(QMargins(2,2,2,2) for the flat case too.
Task-number: QTBUG-132433
Change-Id: Id010d45bf85c5d7d09ff8db8bda44b1be7d85297
Reviewed-by: Wladimir Leuschner <wladimir.leuschner@qt.io>
(cherry picked from commit a930fb758d71ef6bf353e8fd8078de2335c32085)
Reviewed-by: Qt Cherry-pick Bot <cherrypick_bot@qt-project.org>
(cherry picked from commit b7e0a7bfdebfcce46084164c971ad6ab9363184f)
I'm not entirely sure why function_ref_base declares its destructor. I
guess it's declared in order to make it protected, because the class is
non-polymorphic and inherited by function_ref, and one may want to avoid
slicing. On the other hand, function_ref_base is an implementation
detail, living in a `detail` namespace; so that sounds very unlikely.
In any case, the class is clearly meant to be copiable, so add back
(as protected) the missing special members.
Pick-to: 6.5
Change-Id: Ibf909e1746e65eecf8b8990839a6e4c9eb56ca13
Reviewed-by: Marc Mutz <marc.mutz@qt.io>
(cherry picked from commit 74b875641f40b15a4a59a0c55cd79c251b49c05e)
Reviewed-by: Qt Cherry-pick Bot <cherrypick_bot@qt-project.org>
(cherry picked from commit ac7ff59b6ab2694c55b64cd0108b7df4529d9828)
Like the patch in dev, but for 6.8 and below
We need to, following the RFC, send a Dynamic Table Size Update using
HPack once we have acknowledged the setting as long as the new size is
smaller, or if we decide to use the new, larger size.
It's further complicated by the fact that we need to send this update on
the next 'field block' (anything with headers in it), and we may have
multiple SETTING frames come in, and we need to then acknowledge the
_smallest_ one as well as the _final_ one. This is so the decoder on the
peer's side can know that we have set the smallest size, and trimmed our
tables thusly, before going to the larger size.
This could, for example, be used to clear the table.
Fixes: QTBUG-132277
Pick-to: 6.5
Change-Id: I95f006248dfcbc7952c7e184f92d8859270e9e70
Reviewed-by: Mate Barany <mate.barany@qt.io>
This somewhat duplicates some other tests, namely the QHttp2Connection
one, but tst_http2 is our only coverage of the HTTP/2 code prior to 6.8.
This change adds a test where the server sets the
SETTINGS_HEADER_TABLE_SIZE to 0, effectively disabling the dynamic
table. We, in an earlier patch, makes the HPack table error out if
a new header entry is about to be added if the table has not been
resized below a changed 'max'. So, we take advantage of that here and
set the new 'max' capacity without resizing the table, or updating
dynamic capacity causing it to fail if the client doesn't send the
expected Dynamic Table Size Update.
Task-number: QTBUG-132277
Pick-to: 6.5
Change-Id: I1ca8ca7828d5b83606e7adbcfc13c154fa1e3cab
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
(cherry picked from commit f7c3acd27ff48b24f99086be206acbe7e00e3208)
Reviewed-by: Qt Cherry-pick Bot <cherrypick_bot@qt-project.org>
(cherry picked from commit 9a1edf52d4b5af35d29b4b2563d1cebafe889385)
Reviewed-by: Mate Barany <mate.barany@qt.io>
Expose some new API to make it easier to test too.
Will be used in follow-up patches for a http2 bugfix.
The setMaxDynamicTableSize() method no longer sets capacity or evicts
entries. This can then be used to set the new maximum size which we will
accept later when we receive a Dynamic Table Size Update.
This flow makes sense with both encoder and decoder. The decoder would
usually be set by us when we send our SETTINGS frame, and then the
Dynamic Table Size Update would be sent by the peer.
For the encoder we would set the maximum size when we receive the
SETTINGS frame, and then we must defer the actual resizing until we
can send the Dynamic Table Size Update ourselves, usually along with the
next HEADERS frame.
Pick-to: 6.5
Task-number: QTBUG-132277
Change-Id: I959f5006eb09427d130735efe136da8c04453fa2
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
(cherry picked from commit be477819ac17850632bca6ce59ee2c8ef11191cd)
Reviewed-by: Qt Cherry-pick Bot <cherrypick_bot@qt-project.org>
(cherry picked from commit 2101e9146522efa3810b10312207f7700fdc4bc7)
Reviewed-by: Mate Barany <mate.barany@qt.io>
Android 15 (at least) seems to have changed how float arguments are
passed to native functions, breaking our (conceptually correct) code for
processing the va_arg list into a list of static argument types of the
function implementation.
To fix this, we have to move away from using a va_arg function, and
register a function with statically typed arguments instead.
Use a template function that we instantiate with variadic arguments
deduced from the actual function, using a factory-helper that generates
a JNINativeMethod struct with that template instantiation as the
function pointer. Move all of that into a struct where we can also
declare the signature string as compile-time constant without cluttering
the namespace with static objects.
We can now remove the helpers that took care of type promotion in va_arg
functions, and of the tuple-construction from a va_list.
As a drive-by, don't cast function pointers to void *; it's strictly
speaking undefined behavior in C and should have generated a compiler
warning, if not a hard error [1]. We must initialize the
JNINativeMethod::fnPtr member with the address of the function pointer
instead.
[1] https://port70.net/~nsz/c/c11/n1570.html#6.3.2.3p8
Also, declare the native method as the JNICALL calling convention. That
is only defined on Windows, so makes no difference in practice, but it's
the correct thing to do anyway.
Fixes: QTBUG-132410
Change-Id: I190b95fcbcd07cf99c6765fa426c3c351f91994a
Reviewed-by: Volker Krause <vkrause@kde.org>
(cherry picked from commit e91a17873ee4ae58d369b8eb70029cf895b31d03)
Reviewed-by: Qt Cherry-pick Bot <cherrypick_bot@qt-project.org>
(cherry picked from commit 103aa5afb2d95f883daea41715c97455550285b2)
Demonstrates the problem described in the bug report on Android
15/x86_64. QEXPECT_FAIL the relevant test cases until the next
commit fixes the problem.
Task-number: QTBUG-132410
Change-Id: I065fd29282ef42ed75a2ed8177ded183c92aa6e3
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
(cherry picked from commit 749367da8c3309c98b3285836c2bd8abcd7274b1)
Reviewed-by: Qt Cherry-pick Bot <cherrypick_bot@qt-project.org>
(cherry picked from commit e342a5261332c2ef2c3b0d4d14f22dbdbd4a99c3)
In the update from 6.7.2 to 6.7.3, commit `ea0f00d5d` changed how linker
flags like `ws2_32` are treated when generating pri/prl files.
Specifically, before the commit a flag like `ws2_32` was left
untouched. The above commit changed it such that such flags are
converted to `-lws2_32` (seemingly in order to better support FFmpeg,
according to the commit message). However, this change also affects
absolute paths if the file has no extension. That is, after the above
mentioned commit, an absolute path linker flag to, say, a dylib on macOS
without a suffix will result in prepending the `-l` flag. This will
result in errors during link time.
An example where this caused problems can be found in the nixpkgs PR
draft #367206 (https://github.com/NixOS/nixpkgs/pull/367206).
This adds a small check to ensure that `-l` is not prepended if the
linker flag is an absolute path without a suffix.
Change-Id: I2c7ce3aac6624e1a27c59af233e3da2c1ae7ba60
Reviewed-by: Alexey Edelev <alexey.edelev@qt.io>
(cherry picked from commit 714ae22b84553271a2809c393dd1a5a90089cc19)
It's kinda is important to verify that commit() worked; all the
follow-up tests assume it, so wrap the calls in QVERIFY().
Amends 7e5e7eeaa1.
Pick-to: 6.5 5.15
Change-Id: I01185d65a5ec06ed122e59c43d9cdcc3a7157259
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
(cherry picked from commit 97fe38ff0d1e60f92c0bfa6fb3571a0433e60352)
Reviewed-by: Qt Cherry-pick Bot <cherrypick_bot@qt-project.org>
(cherry picked from commit 3660ff9ff88a69ce9f1366e5870f6aa0323f1fda)
To ensure we don't overwrite it in this quite long function.
Amends 7e5e7eeaa1.
Pick-to: 6.5 5.15
Change-Id: Ie5f81b50d2d66ba9e064810f8cbc1a7af298cc2b
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
(cherry picked from commit 67a906b5564a1ccf03e547aa78fc6fd910d9cd15)
Reviewed-by: Qt Cherry-pick Bot <cherrypick_bot@qt-project.org>
(cherry picked from commit b2dad9724d5c92582851e56ba73383644ea8b32b)
The docs said only the first placeholder would be considered; in fact,
it considers only the last.
Fix the docs and add a test (also for QTemporaryDir).
Amends b4bb4449aea7592afdb9b9134bb90c40fe29735a.
While 6.6 and earlier are unaffected by the incorrect documentation,
pick this to all active branches in order to verify the behavior is
the same everywhere.
Pick-to: 6.5 6.2 5.15
Fixes: QTBUG-132597
Change-Id: Ia7a37f714f834191b07420d31ca9024702b537cc
Reviewed-by: Ivan Solovev <ivan.solovev@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
(cherry picked from commit e7abbc7bf3ab06359b2847b1e1a46c38d8bc6581)
Reviewed-by: Qt Cherry-pick Bot <cherrypick_bot@qt-project.org>
(cherry picked from commit 58221c19559541d7da24fa68f05310aeee230e26)
Don't try to update the left and right boundaries when no valid idxRect
was found.
Also fix the early exit check - we can exit when the idxRect is
completely below *or* completely above the current viewport.
This amends 2f9c72028d.
Fixes: QTBUG-132670
Task-number: QTBUG-124173
Change-Id: I51f9e12c66268318e597facfbe4df74367d1089a
Reviewed-by: David Faure <david.faure@kdab.com>
(cherry picked from commit 9d4b5fa017199627a42477a842a4e1ef349bb1ae)
Reviewed-by: Qt Cherry-pick Bot <cherrypick_bot@qt-project.org>
(cherry picked from commit c74f0ccf1552271cc7a01bd9baa3567091894623)
To avoid illegal access, the change ensures that the QAccessibleInterface
is non-null and valid before accessing its QAccessibleActionInterface.
The check for the validity was also added to existing code which
previously only checked for the QAccessibleInterface not being null.
Fixes: QTBUG-132059
Change-Id: I69fc4f9bb052ded8f188032d324666d0c00b9c3c
Reviewed-by: Michael Weghorn <m.weghorn@posteo.de>
Reviewed-by: Lars Schmertmann <lars.schmertmann@governikus.de>
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
(cherry picked from commit 279c891ddf0ad10dd86c8fc836ce385df57593c4)
Reviewed-by: Qt Cherry-pick Bot <cherrypick_bot@qt-project.org>
(cherry picked from commit a30613197731cd77dd4dac8e3d7d8205718c37f0)
When making an object array, we use a LocalFrame to get access to the
conversion routine, but manage the frame of local references manually in
chunks of 100 objects. If any of the conversion routines uses the
LocalFrame's implicit frame management, then we push and pop frames out
of order.
Disable the frame's implicit frame management, and only use our chunk
management.
Amends 37638c84efaf2810ad49da0b987f19287d8c4ad6.
Change-Id: I93be77ffc8750bb90fd7daed8c273169b03a29e4
Reviewed-by: Assam Boudjelthia <assam.boudjelthia@qt.io>
(cherry picked from commit fe6222b7561faa52f12a961d7c5417e364cf0236)
Reviewed-by: Qt Cherry-pick Bot <cherrypick_bot@qt-project.org>
(cherry picked from commit 6a6ce2474e6b6c24dadf8ed2b007c7b3000cf6da)
Giuseppe D'Angelo
Isak Fyksen
Eirik Aavitsland
Moss Heim
Ahmad Samir
Eskil Abrahamsen Blomfeldt
Christian Ehrlicher
Morteza Jamshidi
Assam Boudjelthia
Soheil Armin
Edward Welbourne
Alexey Edelev
Ivan Solovev
Mate Barany
Jøger Hansegård
Thiago Macieira
Mårten Nordheim
Volker Hilsheimer
Volker Krause
Peter Kling
Marc Mutz
Jan Moeller
Friedemann Kleint