Prevents buildup of autoreleased NSViews (drawables) in the outer pool
of the main runloop, which may not drain as often as we wish.
Change-Id: Ifcf7317c50ec243e0d957bf4a19aab8bf34d5dd6
Fixes: QTBUG-84762
Pick-to: 5.15 5.12
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
The trait is deprecated in C++17 and removed in C++20. Enforce
the same meaning by using a constexpr variable instead.
Change-Id: Ief13afc3f889af09094391e626037778d879c4f5
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
And silence those warnings in code that emits those signals.
Change-Id: Ic9013648060c9b84b59c44bb5a8c77e48f82d24f
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
Remove QDestopWidget public header, simplify the implementation that
maintains a Qt::Desktop type QWidget for each QScreen, and turn
QWidget's initial target screen into a QScreen pointer.
QApplication::desktop() now takes an optional QScreen pointer, and
returns a QWidget pointer, so that applications and widgets can get
access to the root widget for a specific screen without having to
resort to private APIs.
QDesktopWidgetPrivate implementations to look up a screen for an index,
widget, or point are now all inline functions that thinly wrap
QGuiApplication::screens/screenAt calls. We should consider adding those
as convenience APIs to QScreen instead.
Note that QWidget::screen is assumed to return a valid pointer; there is
code that handles the case that it returns nullptr (but also code that
trusts that it never is nullptr), so this needs to be defined, verified
with tests, and asserted. We can then simplify the code further.
Change-Id: Ifc89be65a0dce265b6729feaf54121c35137cb94
Reviewed-by: Shawn Rutledge <shawn.rutledge@qt.io>
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Many of these were generated by clazy using the new qevent-accessors check.
Change-Id: Ie17af17f50fdc9f47d7859d267c14568cc350fd0
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
...by deprecating everything that doesn't conform to the naming scheme,
and providing replacements. Continues what was started with QWheelEvent
in 7d29807296. However QMouseEvent::pos()
is left un-deprecated because it's so widely used.
Also quit returning QPointF by const-ref from accessors. It's plenty
small enough to return by value; we were never consistent about it anyway;
and it's good to avoid some problems with returning a reference to a
temporary in case the value is calculated in the accessor.
[ChangeLog][QtGui][QPointerEvent] Mouse, touch and tablet events now
have a standard set of QPointF accessors: position(), scenePosition()
and globalPosition(). Existing accessors that return integer QPoints,
and those with non-standard names, have been deprecated. You can use the
clazy qevent-accessors check to update your code accordingly.
Task-number: QTBUG-20885
Task-number: QTBUG-84775
Change-Id: I8e6f587da76d6d0bca6e965ce8ebc7e67b868011
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
Apart from a more fitting, minimal, API, QDuplicateTracker also
transparently uses C++17 pmr::monotonic_buffer_resource to avoid, or
at least reduce, memory allocations.
Change-Id: I025504c7d22fb7a8c943e3968e4613d45c08d0b3
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Safari on iOS now supports the Navigator.clipboard
API, but not the Navigator.permissions API.
Looks like we have not encountered this combination
Before. Add undefined check for permissions as well.
Fixes: QTBUG-84658
Pick-to: 5.15
Change-Id: I99ab08fd34bbb29a82661e24bf400c927f3604f6
Reviewed-by: Lorn Potter <lorn.potter@gmail.com>
It can be configured using the QFONTCACHE_MIN_COST
define when configuring Qt.
Change-Id: I41fb781099c4c0f03c378f10c8db4ea06ef4e9ff
Task-number: QTBUG-83127
Reviewed-by: Risto Avila <risto.avila@qt.io>
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@qt.io>
These methods are scheduled as a replacement for left/right/mid()
in Qt 6 with a consistent, narrow contract that does not allow
out of bounds indices, and therefore does permit faster
implementations.
Change-Id: Iabf22e8d4f3fef3c5e69a17f103e6cddebe420b1
Reviewed-by: Alex Blasche <alexander.blasche@qt.io>
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Apart from a more fitting, minimal, API, QDuplicateTracker also
transparently uses C++17 pmr::monotonic_buffer_resource to avoid, or
at least reduce, memory allocations.
The code is the first user of the collected data, so make that
available by adding QDuplicateTracker::appendTo(Container&) methods.
Change-Id: Ibd8810c0070db7e6b3ead6d6a569facdab88b646
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
The seemingly useless template parameters are needed to avoid a
(distinct) bug in MSVC 2019 < 16.6, and should be removed once we have a
newer version in the CI.
Task-number: QTBUG-82945
Fixes: QTBUG-83600
Change-Id: I5b22a2259aa16ae90eca7d4f3bd2e4fa1116a73b
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
This patch makes the 'accesiblebridge' plugin configurable.
Change-Id: I99f01fcd434be25bbbe5460bbc8cc1d76aa744b9
Task-number: QTBUG-83126
Pick-to: 5.15
Reviewed-by: Jan Arve Sæther <jan-arve.saether@qt.io>
Pairs are easy to use, but they have no semantics attached: Two
QPair<QString, QString> compare equal, e.g., even though one is used
as a FontAndStyle and the other as, say, a type/subtype. It also
carries no information for the reader of the code.
So, write a minimal struct with equality and qHash() instead.
Change-Id: I9514c9b7d6c2cc6a4831f9ca83ca2ee466d91553
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
This patch adds the missing documentation for the new QSslCertificate
PatternSyntax enum which replaces the one from QRegExp.
Fixes: QTBUG-84464
Pick-to: 5.15
Change-Id: Icf092f42ad4dff9207ca51dfd8b4fb8ed6443ff2
Reviewed-by: Paul Wicking <paul.wicking@qt.io>
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Remove setRoleNames() and all its now unneeded helper functions.
Change-Id: I0a83751aace35700655d4cc7c79278325994cbdd
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
Macros and the await helper function from qfunctions_winrt(_p).h are
needed in other Qt modules which use UWP APIs on desktop windows.
Task-number: QTBUG-84434
Change-Id: Ice09c11436ad151c17bdccd2c7defadd08c13925
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
Port away from random-access of the returned sequences. That's
neither necessary nor does it make the code more readable than
the alternative single-pass implementation used here.
As a drive-by, make applying NRVO more likely by re-using the
`result` variable for all returns after its declaration.
Change-Id: I2c3bbaefa6b6f08ebf0b90fb7be62e3c6243f19d
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
And not pure swap. As per policy, QVariant must leave the rhs empty.
Change-Id: I2d5e0f584c4d4fffd05a0a5bfae27ddcb72430e9
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>
Any integer modulo 1 can never be anything else but 0, so the statement
could never be true. The intention is to abort in case of an odd number
of bytes, as this would indicate malformed input that can't be decoded
into a QString.
Note that QTranslator will then silently continue to search for valid
translations, and not print any error message at runtime, or otherwise
inform the user or developer that an input message file contains
malformed content.
Change-Id: I957b337ee035f3aca013e0859f8ee70553d9a97b
Coverity-Id: 11014
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@gmx.de>
Details of MSVC version became irrelevant some time ago; the code now
just tests for MSVC. In any case, the reader doesn't need to know how
the function is implemented, as long as it does what it claims to do.
We do not, in fact, use a random number source when generating the
return value for comparisons involving only one nullptr; we quite
consistently treat the nullptr as less than the other string. Make
explicit that this is true even if the other is empty.
Change-Id: Ifd9b00fdf8e814fcf933a05821201670fecfd646
Reviewed-by: Andrei Golubev <andrei.golubev@qt.io>
Reviewed-by: Paul Wicking <paul.wicking@qt.io>
It's a bit annoying that you have to include <QStringTokenizer> to use
the tokenize() methods. The separation is a historic artefact, just fix it.
Change-Id: Ied4cc0c8dd2bb6735bf29fbb26fdbf47d07db264
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
Apart from a more fitting, minimal, API, QDuplicateTracker also
transparently uses C++17 pmr::monotonic_buffer_resource to avoid, or
at least reduce, memory allocations.
Change-Id: I0a0d1e31fd35d483e0036045847a3759b593c71c
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
In some places QByteArray does interpret bytes as characters and a
sequence of them as a text string; everywhere else, however, it now
refers to them as bytes.
Add a section to the class doc explaining how char* pointers are
interpreted; the handling of '\0'-termination is surely familiar to
most readers, but making it explicit provides contrast for the
explanation of there being no such special treatment of '\0' bytes
when a stard address and length are given.
Also mention in the class doc that interaction with QString is done
via UTF-8 encoding and can be avoided by defining a macro. Shorten, in
light of that, the description of that encoding where it appears in
affected methods.
In the process, clean up lots of phrasings.
Change-Id: Ic97dce4a20752e277eeab35a06737322b2074692
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
This amends commit 9dd8e655cd to fix the
various references to section {8-bit Character Conversions}, which is
now renamed {Character Case}.
Change-Id: I1f777ab359c616a1b12f05a12b1cd7397f3ccb18
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
The QByteArray code doesn't use QLocale or QString.
Change-Id: I60966ecad35fdaaef9930ba2746bf732fa9f3cd7
Reviewed-by: Andrei Golubev <andrei.golubev@qt.io>
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
A check to prevent crash-on-exit if QLocale::setDefault() was called
after the default store was torn down had the side-effect of causing
setDefault() to not record data if it was called before the first
access to the default store caused it to come into being.
Change the check to test that the default store has been destroyed
and, if not, create it if it doesn't exist yet. This refines commit
4d6572aac0 (as modified by commit
11c5c078c7).
Fixes: QTBUG-83016
Fixes: QTBUG-83415
Pick-to: 5.15
Change-Id: Icbce9bd9c75d0258d403e2f90957561b5a18bdf3
Reviewed-by: Andrei Golubev <andrei.golubev@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Replace the implementation of blockingMappedReduced():
after calling non-blocking version of mappedReduced() we are getting
the future object, so we may call in sequence result(), which
will block and return the result when the all tasks are done.
The same is done with blockigMapped(), which calls blockingMappedReduced()
with a custom reduce function.
Looks like with this pattern we can reuse the non-blocking version
for implementing blocking version of mapped / filtered methods.
Task-number: QTBUG-83918
Change-Id: I7f240cfbd04834d551ff79d717b72194a26996d7
Reviewed-by: Sona Kurazyan <sona.kurazyan@qt.io>
The pipeline value is set by window property _q_platform_qnxPipeline.
Task-number: QSR-263
Change-Id: I13536936b2335a97a6eeb5a94c4e0f552308d15e
Reviewed-by: Samuli Piippo <samuli.piippo@qt.io>
Tor Arne Vestbø
Volker Hilsheimer
Giuseppe D'Angelo
Eirik Aavitsland
Shawn Rutledge
Marc Mutz
hjk
Eskil Abrahamsen Blomfeldt
Morten Johan Sørvig
Martin Jansa
Lars Knoll
Fabian Kosmale
JeongBong Seo
Tasuku Suzuki
Christian Ehrlicher
Samuel Gaist
Oliver Wolff
Jean-Michaël Celerier
Allan Sandfeld Jensen
Edward Welbourne
Alex Trotsenko
Jarek Kobus
Teemu Holappa