QMultiHash has access to two sizes: one of them is shared with QHash,
stored in QHashPrivate::Data::size, which counts keys; the other, which
is what our public size() function returns, is stored in
QMultiHash::m_size and counts plain (key,value) entries. We forgot to
update it in the non-const operator[] that created a node.
I've reviewed the rest of the code and can't find any more places where
the item count may be changed and m_size isn't updated.
[ChangeLog][QtCore][QMultiHash] Fixed a bug that caused an element that
was created by operator[] to not be counted, resulting in a hash map
with an incorrect element count and which could cause an assertion
failure depending on how the hash was later mutated.
Fixes: QTBUG-112534
Pick-to: 6.2 6.4 6.5
Change-Id: Idd5e1bb52be047d7b4fffffd17527ec274e1d99e
Reviewed-by: Lars Knoll <lars@knoll.priv.no>
In some cases, you might want to set the "credentials" flag for HTTP
requests. But you can't use it everywhere, because it may break the
use of public APIs. As a result, you're running into
Reason: Credential is not supported if the CORS header
'Access-Control-Allow-Origin' is '*'
on the client side. See
https://developer.mozilla.org/en-US/docs/Web/HTTP/CORS/Errors/CORSNotSupportingCredentials
for details.
This commit restores the former default "false", in order to make
WASM connections that worked prior 6.5 work again.
Fixes: QTBUG-112478
Pick-to: 6.5
Change-Id: I46f6a374c07038608c3484ac831a1dc5999fb120
Reviewed-by: Mikołaj Boc <Mikolaj.Boc@qt.io>
Reviewed-by: Lorn Potter <lorn.potter@gmail.com>
As I was investigating the NO_UNITY_BUILD_SOURCES issue, I realized that
we don't need to pass these quoted, especially now that we have moved
to `cmake_parse_arguments(PARSE_ARGV`, and we can check their existence
just by checking the `arg_*`, and that should be sufficient.
I also left a warning that we are aware of this.
Pick-to: 6.5
Change-Id: I4d939e80dc4671ea3ae9dc61516279f69ba2c5a5
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
The overhead of making new custom classes appears to be less than
constructing a generic std::function.
[ChangeLog][QtCore][QRunnable] QRunnable::create can now take
non-copyable functions as argument.
Task-number: QTBUG-112302
Change-Id: Ied870f13ca6c7eaa14ed6eff9c4e676c7b73881c
Reviewed-by: Marc Mutz <marc.mutz@qt.io>
Since 68de00e0d4, the check doesn't work
with flatpak versions lesser than 1.13.1. Checking the file in the root
directory works since flatpak 0.6.10.
Pick-to: 6.5 6.2
Change-Id: Icc83ea5de4a962b52a737c9842248df3b60b1331
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Examples in Qt's codebase must be tagged with specific categories
such that Qt Creator can group them thematically. This can be done
by way of using the `\meta category` construct in QDoc.
At the same time, we want the generated documentation to group the
examples by the same logic as in Qt Creator. Hence, QDoc was modified
to implicitly add a group for each category that is used in a
`meta category` invocation.
By design, QDoc exposes ways to list groups to users, but no way to list
the \meta command invocations. Letting QDoc implicitly add a group for
categories passed to the \meta command as a side-effect, therefore breaks
with the principle of least surprise and the single responsibility
principle.
An alternative solution makes use of QDoc's existing support for code
generation through macros. This patch introduces the macro
`\examplecategory` as a global macro throughout Qt to achieve the same
effect as the aforementioned change to QDoc. The macro takes an argument
enclosed in curly braces. This argument is the example category name. It's
used as meta information in the manifest files consumed by Qt Creator, and
added as the category group name for the QDoc side at the same time.
The introduction of this macro allows reverting the change to QDoc
itself, while maintaining feature parity for both Qt Creator and the
generated content.
Task-number: QTBUG-111891
Pick-to: 6.5
Change-Id: I311b98168253b45ac456ff3c1824db3d835191a9
Reviewed-by: Luca Di Sera <luca.disera@qt.io>
Add a testcase for toSQLTCHAR() to make sure to pass the correct number
of encoded characters to the odbc functions.
Pick-to: 6.5
Task-number: QTBUG-112375
Change-Id: Ib67fab678fc3d0b098aedfc6fa9ec2139f2e75c7
Reviewed-by: Marc Mutz <marc.mutz@qt.io>
A QSqlError is not equal when the native error code differs. The
database and driver text should not be considered during the
comparison because they might differ due to e.g. different locales.
[ChangeLog][QtSql][QSqlError] The comparison operators have been fixed to
take both error type and error code into account.
Change-Id: Ie7511f183f88dd454eb165c6ff237e51b79d1c08
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
To request them in QtCore only apps we need to use the CoreFoundation
event dispatcher on the main thread, which is currently blocked by
issues in dealing with deferred delete events.
Pick-to: 6.5
Change-Id: I9fe2575c2b3527b1b83fac208cfed6a0d7a8cb42
Reviewed-by: Juha Vuolle <juha.vuolle@qt.io>
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
We should be able to just use NO_PCH_SOURCES under MINGW condition
and get the same result.
P.S. The problem with `qwindowspointerhandler.cpp`, most likely caused
due to symbol collision with `qwindowsmousehandler.cpp`. We've
encountered it again when we were building with unity.
Pick-to: 6.5
Change-Id: I20a5091d5d1a329228b5ddb4694f5fd61d803554
Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
This is a bit of a cleanup, mainly around unity build, and a few other
minor things:
- I replaced the direct inclusion of sources files using
`set_source_files_properties`, and instead used
`NO_UNITY_BUILD_SOURCES` when possible. In most cases, they are being
excluded in their respective `qt_internal_extend_target` but sometimes
I had to make a new extension.
- In few cases, we had to manually exclude the NO_PCH files, so, I
instead passed them directly to the NO_PCH_SOURCES which also exclude
them from the unity build as well.
- Removed a few unnecessary ""
Pick-to: 6.5
Task-number: QTBUG-109394
Change-Id: I466576592c1d115a2da4d88672c1e4b9f995f2ed
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
We can now use NO_UNITY_BUILD_SOURCES instead, and we only need to
exclude one of the file to avoid the exclusion.
Task-number: QTBUG-109394
Pick-to: 6.5
Change-Id: I33642eece77e362f8a1de7400c14a30da7889239
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
It was just missing, and was ending up being Qt cmake_automoc_parse.
Pick-to: 6.5
Change-Id: I710e22bd86404d412dbf45aa960c924dd148fafc
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
The source of the problem was in `qt_set_target_info_properties` which
was not able to process the NO_UNITY_BUILD_SOURCES, and therefore
leaking it into the `TARGET_COPYRIGHT`, ie., the last argument. So, I
decided to pass Unity Build arguments before them, and closer to
SOURCES, which is nicer to read, and avoid similar situation. And
I reverted the work around in the amend commit, and passing the
arguments normally.
This happens because we pass an unfiltered ${ARGN} from
qt_internal_add_executable to qt_set_target_info_properties and that the
current change is merely a workaround that ensures they get
circumstantially filtered out, because the NO_UNITY_BUILD_SOURCES option
appears before any of the first TARGET_ props.
Amend cd12c1f332
Task-number: QTBUG-99238
Task-number: QTBUG-109394
Pick-to: 6.5
Change-Id: Idb37937cf53e708425402c90f55bda8816e27f29
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
Add CMake rules to skip syncing and installation of header files
that are recognized as non-module. Previously these rules were in
syncqt.cpp only and CMake ignored them when creating the installation
rules. Now we skip any post processing for the header files that:
- are public and located in the '3rdparty' directory unless the module
is the 3rdparty one
- are not a part of the module source tree unless they are generated
- have the _qt_non_module_header property set to TRUE
Pick-to: 6.5
Change-Id: I045cfc2b8074f0c086c975aae95f14845e3edfef
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
Teach QBluetoothPermission to understand new fine-grained Bluetooth
permissions, which were introduced for Android 12.
Also request the ACCESS_FINE_LOCATION permission together with
Bluetooth permissions, because it is required for Bluetooth to work
properly.
At this point QBluetoothPermission simply requests all Bluetooth
permissions, because we do not (yet) have a public API to
distinguish between them.
Task-number: QTBUG-109964
Change-Id: I98d304eb59c238656d6a92917f611c3b3ba9e760
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
Reviewed-by: Juha Vuolle <juha.vuolle@qt.io>
minimumSize.width() was used instead of minimumSize.height() for the
lower bound of height
Change-Id: I330ea79de8ba39a6c55f5ed9b3210f1ff7d10fb6
Reviewed-by: Aleksandr Reviakin <aleksandr.reviakin@qt.io>
Reviewed-by: Lorn Potter <lorn.potter@gmail.com>
MSVC warns due to assignment of std::distance's return value (__int64)
to int when warning C4244 is enabled.
Pick-to: 6.5
Change-Id: I78360ec1bbe861b96d6875af3a29b77f419a3843
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Marc Mutz <marc.mutz@qt.io>
QFileSelector::selectionHelper walks through the list of selectors
(on iOS, on my phone it's ("ios", "iOS", "en_NO", "unix", "darwin")),
creating paths and testing QDir::exists. This ends up in asset
file engine trying to load asset for something like:
"assets-library://assets/+iOS" etc. ALAssetsLibrary -assetForUrl:
returns nil for such url and we start iterating through the assets
library. On my phone (e.g.) this takes about ~6 seconds (to iterate
through pictures/videos I have), so the image picker is dismissed
~30 seconds after an image was actually selected in a picker view,
making an impression it's completely broken. Bail out early
on such url, we know we'll fail (with AssetsLibrary giving a warning
about invalid asset with UUID(null).
Pick-to: 6.5
Fixes: QTBUG-109120
Change-Id: Ia302f4151e3aade830e647a8a260479df2b29d4b
Reviewed-by: Richard Moe Gustavsen <richard.gustavsen@qt.io>
There is no supported OS with a "maybe" monotonic clock. All of the OSes
we currently support either always have a monotonic clock in the first
place or never have one. There's only one OS in the latter category and
that's INTEGRITY, lacking even the CLOCK_MONOTONIC constant.
Change-Id: Ieec322d73c1e40ad95c8fffd17463531dae2c5bc
Reviewed-by: Ahmad Samir <a.samirh78@gmail.com>
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
Reviewed-by: Marc Mutz <marc.mutz@qt.io>
One each of L1 and non-L1.
Will help porting that API to QAnyStringView.
Pick-to: 6.5
Change-Id: I51afc07c5b2384409c2627164e95265265fbb544
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
If the trailing frame just had PRIORITY we would early-return, though
this meant we didn't check if the frame had the END_STREAM flag set,
leading some requests to certain servers to hang.
Fixes: QTBUG-111417
Pick-to: 6.5 6.4 6.2
Change-Id: Iac174dc5aeca30d5d19fae35f303983de9841847
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
Reviewed-by: Konrad Kujawa <konrad.kujawa@qt.io>
Instead of adding more QString::fromMyFavoriteStringImpl(), just check
that
QStringView{myFavoriteStringImpl}.toString()
works.
It does.
Pick-to: 6.5
Task-number: QTBUG-111886
Change-Id: I337282611360b4a56a10c8acfd2d7d53ea196d5b
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
We use the macOS Security framework to check whether the application is
sandboxed or not, in which case we might have to limit or change some of
the functionality of Qt, such as which shared memory backend to use.
Calls to SecStaticCodeCheckValidityWithErrors should ideally not be done
on the main thread, as the function may not return immediately, and we
get runtime analysis warnings about this in Xcode:
This method should not be called on the main thread as it may lead to
UI unresponsiveness.
To improve the situation we spin up a short lived thread at library
load that resolves the sandboxing state, ready to be queried when
needed from Qt.
Pick-to: 6.5
Change-Id: I52cdc1bf6aef05c3b93b43f67b3fb46035996b3a
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
QHash/MultiHash & QMap/MultiMap::erase() do no longer take an iterator
in Qt6. Clean up the examples by only providing one example, the rest is
common c++ which should not be handled in the Qt documentation. Also
mention erase_if() and remove references to the (soon to be deprecated)
Java-style iterators.
Pick-to: 6.5 6.2
Fixes: QTBUG-105109
Change-Id: I47b11f3b5dcc173494e5c6f9ad0167c613b12209
Reviewed-by: Marc Mutz <marc.mutz@qt.io>
This will be the building block for emplacement support in QPromise.
Task-number: QTBUG-112270
Change-Id: Ie267e76fe078a8f7b3ef8c7e0d2abfcd22062ff2
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
The T&& result argument is not deduced (T is the template class'
template argument, so it's an rvalue reference, not a Universal
Reference, assuming that we don't support QFutureInterface<U> where U
is a reference type).
So std::forward<T> will always be a std::move(), so use that directly
instead of raising eyebrows in the reader of the code by using
forward<> to feed a function called _move_Result().
Pick-to: 6.5
Change-Id: I805df4686b5b74da57f8212b052b4056943a15fa
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
The T&& result argument is deduced, therefore result is a Universal
Reference and as such can bind to lvalues, not just rvalues. If passed
an lvalue, the function would happily move away from it, which is not
what the doctor prescribed.
Catch the attempt with a static assertion.
Pick-to: 6.5
Change-Id: Iff8c1abd5dcb1043ed94ba76570be5ba3d6f92da
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
If the move can fail, so can most likely the copy. Use of
move_if_noexcept() therefore trades a faster potentially-throwing
operation for a slower potentially-throwing operation.
It is true that a throwing move ctor may leave the source object in a
partially-formed state while the copy ctor may not (exceptions prove
the rule, hello auto_ptr), but we shouldn't pessimize lazy coders
(missing noexcept) to cater for borderline-buggy code (throwing move
ctor that doesn't provide the strong exception guarantee). Yes, you
can construct such situations (and something as simple as
std::array<std::pmr::string, 10> is already affected), but we don't
cater for these anywhere else in Qt, so why do it here?
Change-Id: I728af8e87ed86f24326530f704a48df1f728ac98
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Add the move ctor and move operator for QSqlIndex, also add an explicit
testcase for QSqlIndex
Task-number: QTBUG-109938
Change-Id: I46cc6a24c2e7d5b23d2ac3427cafd01b9ba257ed
Reviewed-by: Marc Mutz <marc.mutz@qt.io>
The documentation suggested to use a non-existing data() function.
Fix it to mention the actual QPermission::value() method.
Pick-to: 6.5
Change-Id: I870021ef5a04edc26553d8b0376da007687b2f05
Reviewed-by: Giuseppe D'Angelo <giuseppe.dangelo@kdab.com>
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
This will be helpful in a number of places, in particular in order to
support enums of different sizes in QML. We record the type as string in
the JSON output and as QMetaTypeInterface in the generated C++.
Task-number: QTBUG-112180
Change-Id: I943fac67f8b25b013d3860301416cdd293c0c69e
Reviewed-by: Shawn Rutledge <shawn.rutledge@qt.io>
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
Both the full-width and partial-width paths in
QWasmBackingStore::updateTexture now correctly compute source and
target rects.
Task-number: QTBUG-112414
Change-Id: I30b0952609960f521119d3d628d2a8036f8b1fe5
Reviewed-by: Morten Johan Sørvig <morten.sorvig@qt.io>
We were missing the NSEventTypeMouseMoved in resolving the event type for
handleFrameStrutMouseEvent, resulting in passing on QEvent::None as the
event type to QWSI.
Pick-to: 6.5 6.2
Change-Id: I144b6459c5d4fda2839d25cd52ddb441dc7c3a47
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
Some dead code was left after transferring event support to QWasmWindow.
Remove it now.
Change-Id: I40e15bc62bcbb6fff071f53c45223c8a2d12e348
Reviewed-by: Lorn Potter <lorn.potter@gmail.com>
An implementation of C++20 std::to_address, a prerequesite for
QSpan<>.
The test cases are inspired by libstdc++'s test suite, just to avoid
missing some cases, but the to_address implementation is white-room.
Fixes: QTBUG-108430
Change-Id: I4c092fdd7a56c0b279068e341bbf91a725ca3b1f
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Render the Up/Down and Plus/Minus images for QSpinBox in high-dpi mode.
Found during investigation of QTBUG-112019.
Pick-to: 6.5
Task-number: QTBUG-112019
Change-Id: Ie531311b657180790b0e12a36663012a5f4da6be
Reviewed-by: Eirik Aavitsland <eirik.aavitsland@qt.io>
The handleMouseEvent function already takes a QEvent::Type, where clients
pass in the corresponding mouse press/release/move type. The same applies
to the handleFrameStrutMouseEvent.
To avoid the chance that clients call these functions with a conflicting
event type (handleFrameStrutMouseEvent with MouseButtonPress instead of
NonClientAreaMouseButtonPress e.g.), we remove handleFrameStrutMouseEvent
altogether and just let clients use the handleMouseEvent function directly
with the correct event type.
Change-Id: I4a0241c39aedac0d2d8d5163ba05cde72605959c
Reviewed-by: Lars Knoll <lars@knoll.priv.no>
We need to pass the list of generated files to make the correct
filtering when generating the module master header. Otherwise
generated headers will endup in the module master header.
Pick-to: 6.5
Fixes: QTBUG-112458
Change-Id: Icd20fe5999db379d0eafbff7dad20765e1a96350
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
By erroring out we can ensure that new warnings are not
accidentally introduced and merged.
As a drive-by: fix indentation of the --qt option. "--dot" needs three
tabs, so "--qt" definitely needs it too.
Fixes: QTBUG-108119
Change-Id: I56107c0744957293338080cf37350f1e1c743093
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Reviewed-by: Kai Köhne <kai.koehne@qt.io>
This reverts commit 101e57d37c.
We have a separate change to fix Qt Wayland Compositor.
Change-Id: Ia00f73c9864921523fed205a813496ab417810a8
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
Reviewed-by: Liang Qi <liang.qi@qt.io>
The stream-based XML serialization API resides in corelib/serialization.
Move the steambookmarks example there. The Qt XML documentation is
updated to no longer refer to this example code directly and refer to
the direct location in the example documentation instead.
Task-number: QTBUG-110647
Pick-to: 6.5
Change-Id: Id36fb04a6acb7b8d1eb008f61568fe0abc221e3d
Reviewed-by: Marc Mutz <marc.mutz@qt.io>
QListView generates a move, when an item is dropped directly behind
itself. This causes unexpected behavior, e.g. item widgets getting
discarded.
This patch prevents a move from being generated in that case.
It adds an autotest to tst_QWidget, to verify item widgets and item
data do not get discarded in case of a no-op drag&drop.
Fixes: QTBUG-100128
Pick-to: 6.5 6.2
Change-Id: I02a755320a7b71dad218293c9c94c88da6507423
Reviewed-by: Jan Arve Sæther <jan-arve.saether@qt.io>
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Make it easier to find the function that actually gives focus to
windows that need it.
Pick-to: 6.5
Change-Id: I0586b89b42a1be260e36e166c81b686e394d8b5e
Reviewed-by: Nicholas Bennett <nicholas.bennett@qt.io>
