CMP0099 changes the way of LINK_ONLY genex works. With CMP0099 set to
OLD LINK_ONLY genex only links the exact library binary/archive without
propagating other interface options from the target. This feature was
exploited by PlatformXInternal targets to avoid propagating of their
linker options. Nowadays when CMP0099 is forced to NEW by Qt scripts,
including user-facing, we cannot rely on LINK_ONLY genex.
Introduce _qt_is_internal_target property that is set for all Qt
executables and explicitly limits the propagation of the linker
options from PlatformXInternal targets.
Pick-to: 6.5 6.6
Fixes: QTBUG-113641
Change-Id: I3a0ecddb65886e435073feb24c1b47035130ba70
Reviewed-by: Alexandru Croitor (OOO) <alexandru.croitor@qt.io>
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Move the implementation qRequireVersion() to prevent having
to include qdebug.h which pulls in many other headers.
Amends b5d874e36f.
Fix missing include introduced by
3a553507a1.
Pick-to: 6.6
Task-number: QTBUG-114214
Task-number: QTBUG-97601
Change-Id: Iba68ffca95061666d9458ffa5700d07c7669da5b
Reviewed-by: Axel Spoerl <axel.spoerl@qt.io>
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reworked the code to list supported ciphers for Schannel
backend to provide the proper cipher names and cipher
properties.
Note: restricting the used ciphers for a TLS connection
is only supported for TLS 1.2 yet!
[ChangeLog][QtNetwork][QSslConfiguration] Add support for
listing supported ciphers with Schannel backend.
Change-Id: Idfc17335ec489315387a3e84b7a76a263be22378
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Currently items can get stuck in a hovered state when all fingers are
lifted from a touchscreen. This is because we don't react to the
WM_POINTERLEAVE event from Windows. With this patch we translate a
WM_POINTERLEAVE event to a LeaveEvent to remove the hover state from
QtQuick items.
Fixes: QTBUG-62912
Pick-to: 6.5 6.6
Change-Id: I8a6fb6b7ec77457854a75e20277565d1eb89bab6
Reviewed-by: Wladimir Leuschner <wladimir.leuschner@qt.io>
Reviewed-by: Jan Arve Sæther <jan-arve.saether@qt.io>
The signal can deviate from QWidget::isVisible().
This patch clarifies the signal's documentation.
Since its name suggests that it is in sync with QWidget::isVisible(),
A TODO comment is added to consider deprication in Qt7.
It appears to be more reliable to listen to hide/show events.
Pick-to: 6.6 6.5
Fixes: QTBUG-48161
Change-Id: I43aa16c2ecb4877abd8effb7da8e07576438d6d2
Reviewed-by: Richard Moe Gustavsen <richard.gustavsen@qt.io>
This was used as a work-around for Q5MutexLocker's UB downcasting
QBasicMutex to QMutex, and for some unspecified performance
improvements.
Now that Q6MutexLocker is a template, the two are almost
token-by-token identical, so we can remove this work-around and use
QMutexLocker directly.
Partially reverts aea500d5d7.
Pick-to: 6.6 6.5 6.2
Change-Id: I57ef5c53999869aa3454fbbaad884c1d18591b2a
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
On FreeBSD, the environ variable doesn't come from libc, but is instead
inserted during linking. See ccf74b5928
(5.6) for more information.
But instead of allowing undefined symbols in QtCore, let's use a weakref
to environ so this one symbol is allowed to be undefined. It won't be,
but the linker doesn't know.
FreeBSD appears to be the only BSD to require this. We used to apply the
same linker option to OpenBSD in Qt 5, but neither the OpenBSD or nor
the NetBSD ports trees[1][2] carry a patch for this, so I don't think
it's necessary.
[1] https://github.com/openbsd/ports/tree/master/x11/qt6/qtbase/patches
[2] https://github.com/NetBSD/pkgsrc/tree/trunk/x11/qt6-qtbase/patches
Pick-to: 6.6
Change-Id: I63b988479db546dabffcfffd17661c839014771a
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@gmx.de>
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
We so far only handled them if they came at the very start of the method
declaration.
This patch ensures that we also handle them after the meta-method tag
(but before the actual type).
Unifying parseFunction and parseMaybeFunction to avoid the need to
munally keep them in sync is left for another day.
Fixes: QTBUG-111330
Pick-to: 6.6 6.5 6.5.2 6.2
Change-Id: Ic94edb69f04b9150aea2c8e6d004a8b9e5cf12ec
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
The GTK platform theme (and possibly others) emit theme change
when the named icon theme changes, but that was not propagated
to QIconLoader.
We now call QIconLoader::updateSystemTheme(), but note that
if a user theme has been set we ignore the system theme update
and will end up with a stale value, even if the user theme is
later cleared. This will be fixed in a follow up commit.
Pick-to: 6.6 6.5
Change-Id: I40b537f3618f44d396db0c7ca67e515dfcdfba44
Reviewed-by: Axel Spoerl <axel.spoerl@qt.io>
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
Change-Id: I8e0c5d7b9049fb3c3248749db8dc8c616db293f0
Reviewed-by: Mikołaj Boc <Mikolaj.Boc@qt.io>
Reviewed-by: Piotr Wierciński <piotr.wiercinski@qt.io>
The qstdweb::bindForever function will be introduced soon for binding
callbacks with forever lifetime. For now, since submodules have to be
adapted first, introduce a stub function to port qtmultimedia which
uses Promise::make - it will have its interface changed with
relation to bindForever.
Task-number: QTBUG-112296
Change-Id: I578e1633574e11877a8f5ec14d0f00dfd7c766bf
Reviewed-by: Aleksandr Reviakin <aleksandr.reviakin@qt.io>
Reviewed-by: Piotr Wierciński <piotr.wiercinski@qt.io>
Reviewed-by: Lorn Potter <lorn.potter@gmail.com>
QFontEngine tries to generate glyphs for ASCII control characters. Those
characters have no glyphs and should therefore not be generated.
Task-number: QTBUG-113848
Change-Id: Ib15a50c642d8bc92bc007ab5522aebc5a7b2993d
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@qt.io>
Some of our tools don't have the `-h`, or `-v` flag, or it could be
that the `-v` flag also prints the entire `--help` as well, e.g.,
`androiddeployqt`. When running in Jenkins, this may lead to a message
box being shown and consequently stopping the build. By customizing the
flag per tool, and limiting the TRY_RUN to tools that support `-v` or
`--version`, we can avoid this.
Also removed TRY_RUN from `macdeployqt` which doesn't need it anyway.
Amend 41b32cd2c4
Pick-to: 6.5.2 6.5 6.6
Fixes: QTBUG-114530
Change-Id: I78e3344d2553c0050c285ae86f2310bd373c6c57
Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
The method is more than a plain getter, so keep it with the other
logic of the class.
Pick-to: 6.5 6.6
Change-Id: I34aa185a51f04e3db3c1918f9723e53f33e5e9e1
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
The example was moved to qtnetwork but its documentation
still referred to old paths under /examples/corelib.
Add documentation dependency to qtwidgets as RSS listing
example's docs link to a number of widget classes.
Pick-to: 6.6 6.5
Change-Id: Ief1c9dc2ca38ba3da1fe3039500292147ec4cc7d
Reviewed-by: Andreas Eliasson <andreas.eliasson@qt.io>
The test checks whether windows repaint correctly by sampling the
background after their backing stores have been flushed.
Change-Id: Ib544457074d7d477a4acdc5c331ef83e5ba471d2
Reviewed-by: Piotr Wierciński <piotr.wiercinski@qt.io>
Reviewed-by: Mikołaj Boc <Mikolaj.Boc@qt.io>
If the style changes while a popup is open (or about to close),
the popup will change style as well before it's hidden. And
this can result in the popup window briefly ending up smaller
than what it needs to be, in order to fit all the menu items.
In that case, it will show 'up' and 'down' widgets in the
menu that auto scrolls it when hovered. And all this can
happen for a split second while the menu is about to close
(as a result of the user clicking on a menu item).
A bug happens because of this if you click on the last menu
item in the list, and this causes the style to change. In
that case, the 'down' widget will end up directly underneath
the mouse for a split second, which will trigger an auto-scroll
timer to start. This timer will trigger a bit later, after
the popup has been hidden, and scroll the list view a bit down.
The result is that the next time you open the popup, it ends up
at the wrong place on the screen in a failed attempt to center
the current index on top of the combobox.
This patch will make sure that we always scroll the list view
to the top before we start calculating where the popup should
be placed on the screen. Otherwise the geometry ends up wrong
since the popup will anyway be resized (if possible) to fit
all the menu items before it's shown and should therefore not
take scrolling into account.
Pick-to: 6.6 6.5
Fixes: QTBUG-113765
Change-Id: I61b5b832904de471c2303fc67325feec322b1449
Reviewed-by: Axel Spoerl <axel.spoerl@qt.io>
Like commit 91f6460aff which added support
for Windows. This API is documented and has apparently been present in
FreeBSD for a long while.
The DragonflyBSD API is very similar, but I don't have one to confirm
that I've coded correctly. OpenBSD and NetBSD may have similar APIs, but
I haven't even researched them. We're open to contributions, though.
Change-Id: I63b988479db546dabffcfffd1766bc431fed614b
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
This allows us to use absolute times on Linux (today) and FreeBSD
(soon), plus simplifies both QMutex and QSemaphore.
Change-Id: I63b988479db546dabffcfffd17675a182aa528fa
Reviewed-by: Ahmad Samir <a.samirh78@gmail.com>
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Makes it cleaner when I add more OSes.
Change-Id: I63b988479db546dabffcfffd1766bee2e6370b94
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Reviewed-by: Ahmad Samir <a.samirh78@gmail.com>
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
Setting parents for WASM platform windows is now supported. This means
that windows now reside in a hierarchical window tree, with the screen
and individual windows being nodes (QWasmWindowTreeNode), each
maintaining their own child window stack.
The divs backing windows are properly reparented in response to Qt
window parent changes, so that the html structure reflects what is
happening in Qt.
Change-Id: I55c91d90caf58714342dcd747043967ebfdf96bb
Reviewed-by: Morten Johan Sørvig <morten.sorvig@qt.io>
QDoc allows code-expanding macros to be defined so that the user can
abstract away common code in a simple to use command.
Furthermore, macros allows for a different expansion based on the
currently generated format so that format-specific requirements can be
satisfied when required.
When format specific expansions are provided, QDoc considers the
resulting text as if it should be generated verbatim.
The "\raisedaster" macro is used to generate an superscript asterisk.
The macro is provided for both HTML and DocBook.
As the macro has format-specific expansions, the expanded code will be
read verbatim.
Nonetheless, while the HTML expansion expands to actual HTML, the
DocBook expansion expands to an equivalent QDoc code, which will not be
generated correctly.
To avoid the issue, the DocBook version of "\raisedaster" is modified so
that it produces the correct DocBook code.
Change-Id: I3a37838bda885af42f8d93b86ec08126d7146ff9
Reviewed-by: Topi Reiniö <topi.reinio@qt.io>
Reviewed-by: Paul Wicking <paul.wicking@qt.io>
QDoc allows code-expanding macros to be defined so that the user can
abstract away common code in a simple to use command.
Furthermore, macros allows for a different expansion based on the
currently generated format so that format-specific requirements can be
satisfied when required.
Due to a certain bug in QDoc, when generating the DocBook format, QDoc
would expand an HTML specific macro definition when a DocBook specific
one was not provided.
As this bug is now being fixed, the DocBook format will lose some of the
output that it was previously generating.
For example, the "\beginqdoc" and "\endqdoc" macros are defined, for
HTML, to expand to the beginning of block-comment text, "/*!" and ending
of block-comment text, "*/".
To avoid losing the usage of "\beginqdoc" and "\endqdoc`" when
generating DocBook, an equivalent expansion of the macro is now provided
for the DocBook format.
Change-Id: I45fb54f1f56077771c091323a69fd63e09a910eb
Reviewed-by: Paul Wicking <paul.wicking@qt.io>
Reviewed-by: Topi Reiniö <topi.reinio@qt.io>
By default QDoc writes filePath attributes for each entry in the
.index file it generates. The attribute value is an absolute path
that's only relevant or useful for debugging purposes.
Pick-to: 6.6 6.5
Change-Id: I3bc646104af74b67c74350912359cf65ac252992
Reviewed-by: Paul Wicking <paul.wicking@qt.io>
QDoc allows code-expanding macros to be defined so that the user can
abstract away common code in a simple to use command.
Furthermore, macros allows for a different expansion based on the
currently generated format so that format-specific requirements can be
satisfied when required.
Due to a certain bug in QDoc, when generating the DocBook format, QDoc
would expand an HTML specific macro definition when a DocBook specific
one was not provided.
As this bug is now being fixed, the DocBook format will lose some of the
output that it was previously generating.
For example, the "\youtube" macro is defined, for HTML, so that a
youtube video can be embedded by reference.
To avoid losing the usage of "\youtube" when generating DocBook, an
equivalent version expansion of the macro is now provided for the
DocBook format.
Change-Id: Id74155f2c30b80b5f4490f8451cd8d00535806d6
Reviewed-by: Topi Reiniö <topi.reinio@qt.io>
Reviewed-by: Paul Wicking <paul.wicking@qt.io>
QHeaderView::mouseMoveEvent started autoscroll without propagating the
event's mouse position to QAbstractItemViewPrivate::draggedPosition.
This data member always containing QPoint() has lead to right drags not
causing an autoscroll at all. Left drags with a scroll offset just
kept scrolling until the offset was 0.
The missing propagation has been added.
As a drive by, dead code has been removed and the local variable pos
has been constified.
Fixes: QTBUG-113573
Pick-to: 6.6 6.5
Change-Id: I7b194dfc71abea6f2bbaaae18270c80eb15afb4d
Reviewed-by: Oliver Eftevaag <oliver.eftevaag@qt.io>
Fix the argument prefix that is used in the cmake_parse_arguments call.
Pick-to: 6.5 6.6
Change-Id: Ie02bdf7d2769ce084b0d173c1e8152ca6fc4fe53
Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
Most of the QWS (Qt Windowing System) classes from
the Qt 4 times are cleaned up in Qt 6. Any existing
QWS API documentation should be marked \internal.
Pick-to: 6.2 6.5 6.6
Task-number: QTBUG-35605
Change-Id: I24201406114feaf8af21403b09375a2ee9a4709e
Reviewed-by: Laszlo Agocs <laszlo.agocs@qt.io>
It's really showing how to request a resource and act on its becoming
available. The use of XML to do so is incidental; the use of
networking is central.
Pick-to: 6.6 6.5
Task-number: QTBUG-111228
Change-Id: Ibcf438c7ef3b2464ddfa8b96a79fb15523e4a468
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
We should not export non-polymorophic classes wholesale. Only export
the non-inline functions instead.
There are no implicitly-declared special member functions in this
class that could cause problems, so we don't need to delay until Qt 7.
Pick-to: 6.6
Change-Id: I204d703498bf42465ea122d1f31c443ea439bd6e
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
I don't know why the compilers didn't shout here, but what _should_
have happened is:
- qSwap<quint64>() gets instantiated
- the unqualified swap() call inside gets resolved to std::swap()
- std::swap() is not constexpr in C++17, so qSwap<quint64>() silently
gets its constexpr dropped
- error, due to the use of non-constexpr function qSwap() in constexpr
function bswap()
There's no way through the function that doesn't hit the qSwap(), so
that is also not the explanation. And, indeed, replacing qSwap() with
std::swap() gets me the expected error...
Before compilers get the idea, rewrite the code to not require
swapping.
Amends 686c02224c.
Pick-to: 6.6
Change-Id: Ie1364bb2fd148bf995a8ffd321f77a6021176928
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
GCC 11 complains:
qprocess_unix.cpp:672:12: error: ‘char* strncpy(char*, const char*, size_t)’ specified bound 508 equals destination size [-Werror=stringop-truncation]
672 | strncpy(error.function, description, sizeof(error.function));
| ~~~~~~~^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
And it's correct: if description is longer than
sizeof(error.function), then error.function will not be
NUL-terminated.
While a quick check suggests that the user of the field performs a
qstrnlen(), thus avoiding falling off the unterminated end of
error.function, it's safer to always NUL-terminate. A single added
qDebug() << error.function would already be UB.
Fix by using _q_strncpy(), which is also more efficient, as it doesn't
write 0.5KiB of NULs in the likely case that description is short.
Amends 90bc0ad41f.
Change-Id: If5c2cb80fc4a3c92b8e78b680a635045bb14a30d
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Apparently it is not enabled since the TLS backends became plugins,
which is unfortunate I wrote code in here that assumed it was.
Change-Id: I765435ce7d2c6fafa452d0453c705c27ff477be6
Reviewed-by: Alexey Edelev <alexey.edelev@qt.io>
It's a RAII class, and RAII classes should be marked [[nodiscard]] at
the class as well as the ctor level.
Pick-to: 6.6
Task-number: QTBUG-104164
Change-Id: Ie88023ba7c57dad7c2116c1c19a80b908b3a9f4d
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Having the generic qt_itemFired: as action would result in the whole
submenu tree closing if an item with a sub-menu was clicked on. This
is not how native applications behave. They respond by immediately
opening the submenu, or do nothing if the menu is already open.
By using submenuAction: as the selector we achieve the same behavior.
A complication here is that for some reason we defer associating the
submenu NSMenu to an NSMenuItem until QCocoaMenu::setAttachedItem(),
instead of doing it in QCocoaMenuItem::setMenu(), or even as part of
QCocoaMenuItem::sync().
As a result, AppKit's NSMenuValidation logic will conclude that the
item does neither have a submenu, nor a valid target/selector combo
to be validated, and will explicitly disable the item. This can be
debugged by passing -NSTrackMenuValidation YES to the application.
To work around this we explicitly enable the item once we have set
a valid submenu for the item.
Pick-to: 6.5 6.6
Fixes: QTBUG-114199
Change-Id: I7178e7687066b3fe082454c512ec9c7eab3bded4
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
Declaring it as an explicit dependency should hopefully make
the compilers order the object-files to be linked in the
correct order.
Task-number: QTBUG-114243
Pick-to: 6.6 6.5
Change-Id: I723fee468e58786f66d1bba50bc4086beb50adb1
Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
Reviewed-by: Alexey Edelev <alexey.edelev@qt.io>
The idea is to store a concatenable in a QStringBuilder object by value
or by reference, depending on how it was originally passed into the
concatenation operator. So if it was passed by r-value, we treat it as
a temporary object and hold it by value (and use move-semantic if
available), otherwise we hold it by reference (as before).
To achieve this we first change concatenation operators '%' and '+'
to take their arguments by universal reference. Next we instantiate
QStringBuilder object with deduced types of the arguments, which will
be a "value type" or a "reference type" according to "universal
reference deduction rules".
Further we use perfect forwarding to pass arguments to QStringBuilder's
constructor. Thus arguments, initially passed by r-value reference
and which are move-constructible, will be "moved" to corresponding
QStringBuilder member variables.
So, to summarize:
1. Arguments passed by l-value reference - stored in QStringBuilder
object by reference (as before).
2. Temporary objects passed by r-value reference - stored in
QStringBuilder object by value. If a type is move-constructible
(QSting, QByteArray, etc), the object will be "moved" accordingly.
Special thanks to Giuseppe D'Angelo for the tests.
Fixes: QTBUG-99291
Fixes: QTBUG-87603
Fixes: QTBUG-47066
Task-number: QTBUG-74873
Task-number: QTBUG-103090
Task-number: QTBUG-104354
Change-Id: I64f417be0de0815ec5ae7e35a1cc6cef6d887933
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Vladimir Belyavsky <belyavskyv@gmail.com>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Marc Mutz <marc.mutz@qt.io>
In case the current GTK3 theme has no monospace font defined, a
monospace font requested by a Qt application can trigger a GTK warning
"Theme parsing error: <data>:1:0: Expected a valid selector".
The warning is triggered by Qt requesting "{font-family: monospace;}".
In this case:
=> ensure fallback to GTK standard monospace font
=> request "* {font-family: monospace;}" to avoid the warning
Task-number: QTBUG-112896
Pick-to: 6.6 6.5
Change-Id: I24a8da62908af9b153245f53026af60e63a600d7
Reviewed-by: Oliver Eftevaag <oliver.eftevaag@qt.io>
As Eddy pointed out in review, the existing overload API docs says using
this method is discouraged; but adding a QDeadlineTimer overload is more
about preventing overflow, as it can handle qint64 or
chrono::milliseconds.
So it's either add this new overload or change the existing one to take
a qint64.
[ChangeLog][QtCore][QCoreApplication] Added processEvents() overload
that takes a QDeadlineTimer.
Task-number: QTBUG-110059
Change-Id: I02f938ee8243c09e493bd88ed496b862d87910f7
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Marc Mutz <marc.mutz@qt.io>
It's a RAII class, and RAII classes should be marked [[nodiscard]] at
the class as well as the ctor level.
Pick-to: 6.6
Task-number: QTBUG-104164
Change-Id: I77260dc00d51d62ed1064e13f566d04e88d28ef9
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
It's a RAII class, and RAII classes should be marked [[nodiscard]] at
the class as well as the ctor level.
Pick-to: 6.6
Task-number: QTBUG-104164
Change-Id: Ie877e261cfe602410d9d9bb3acc658d0bb7c4e72
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
