paintEventCount() is currently flaky on macOS. It gets extra paint events after
qWaitForWindowExposed() returns, which causes the following assertions to fail.
Add the wait that was removed in 0cb940b1d3 back
to fix those failures.
Task-number: QTBUG-68032
Change-Id: I68e0b6008de40922ec740291dfdd1842e0f62f89
Reviewed-by: Sami Nurmenniemi <sami.nurmenniemi@qt.io>
Reviewed-by: Richard Moe Gustavsen <richard.gustavsen@qt.io>
Reviewed-by: Ville Voutilainen <ville.voutilainen@qt.io>
Replace the code for isRowSelected and isColumnSelected with
a much simpler algorithm for deciding if a row/column is selected.
In a model with a cross-hatch of unselectable indexes, the return values
of is(Column/Row)Selected would depend on the order in which the
selections were done.
Task-number: QTBUG-18001
Change-Id: I6aa4b1df7c07fae469a686041927fa8c42bc9b16
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@qt.io>
Reviewed-by: Richard Moe Gustavsen <richard.gustavsen@qt.io>
Previously the code for isColumnSelected and isRowSelected differed
slightly, in how unselectable indexes would be treated.
This made isColumnSelected return false for a column, which mixed
unselectable indexes and selected indexes. Thus in some situations,
the user could not deselect a column via a QTableView header.
By copying the isRowSelected code to isColumnSelected, rows and
columns behave identical.
Task-number: QTBUG-18001
Change-Id: I6ca85ac64b31a481fafeaa3bec958b18283eed8d
Reviewed-by: Richard Moe Gustavsen <richard.gustavsen@qt.io>
Home/End don't actually work on macOS. The "select all and delete" key
was not actually testing anything at all.
Change-Id: I44d3e9dd27da418afd699bf8720d5369325d20df
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
Reviewed-by: Andy Shaw <andy.shaw@qt.io>
Introduce C++11, nullptr, for, port to Qt 5 connection syntax.
Change-Id: I2d233ccd68bad533af8d4674d91236b2c049e997
Reviewed-by: Andy Shaw <andy.shaw@qt.io>
Having rows without fields is a supported use case so it shouldn't
assert. Code works quite well in release mode, but crashes in debug
mode.
Change-Id: I1c4f736318489bae09780fcdb56136181afcac17
Reviewed-by: Samuel Gaist <samuel.gaist@edeltech.ch>
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
... but only if the host it came from is an EXACT match. Also only apply
the cookie if the url is an EXACT match.
[ChangeLog][QtNetwork][QNetworkCookieJar] Cookies will no longer be
rejected when the domain matches a TLD. However (to avoid problems
with TLDs), such cookies are only accepted, or sent, when the host name
matches exactly.
Task-number: QTBUG-52040
Change-Id: Ic2ebd9211c48891beb669032591234b57713c31d
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Checking for Q_OS_LINUX, Q_PROCESSOR_ARM and use of QT_CONFIG() checks should
only happen after qglobal.h is included. Otherwise the header will be broken if
included before something that uses qglobal.h
Change-Id: I052e46784f7b174e74e8894e1b7c5b7528420f5d
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
In bootstrapped mode QDBusMessage isn't available, so looking up the type
via QMetaType wont work. QDBusMetaTypeId has this special-cased, but that
alone isn't enough for qdbuscpp2xml to produce the same result as in non-
bootstrapped mode.
The effect of this has also been described here before in detail:
http://lists.qt-project.org/pipermail/development/2017-February/028756.html
Change-Id: Id309a3a910f971c6150cdc6d06f2b48f1b95c787
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
The class is not thread safe, so one instance is not enough for whole
application. Mention that QNetworkAccessManager instance can only be
used from the thread it belongs to because it is a QObject.
Change-Id: I56184e4f8fbd36aca3f6677310431eab88346e6e
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
If the input method event contains a commit text that ends with a
newline, text, the commit string is inserted first. This changes the
current block.
This change makes sure that we apply the formatting changes (including
removing the old preedit text) to the old block in this specific case.
Task-number: QTBUG-67836
Change-Id: Ia83963780fb14b3c571dbbe3eb81fbbe20fbf412
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@qt.io>
As per today's change in the MIME spec.
https://bugs.freedesktop.org/show_bug.cgi?id=97372
Change-Id: Iba4fdd95c3ebec8a042404956db3466a46c97f1d
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
checkReason_ActiveWindow() started failing on Windows when run together with
other tests, but still passed on its own.
The offending tests was checkReason_focusWidget(), which showed a window but did
not wait for it to be active.
After adding this wait the whole test executable passes on Windows as well.
Amends fd87c8da82.
Change-Id: I384bc45176fcd7bf6f491a4f39b46464ba45693b
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
We need to make sure we don't emit CMake declarations for private
headers if those headers are absent. However, most of the time we have
private headers and should add them.
Task-number: QTBUG-37417
Change-Id: I639eb93d008de27928dedac540894af70c1883b9
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@qt.io>
This way the platform window is destroyed in a timely manner, preventing
redundant close events from the window system.
Task-number: QTBUG-43344
Change-Id: Ifdfca59ceacef54405f1c227c493dc514a1b27ea
Reviewed-by: Morten Johan Sørvig <morten.sorvig@qt.io>
Instead of a local wrapper for it.
Change-Id: I0708dfad44b3db0c7a13e75ba5b4193ab50ac315
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Gatis Paeglis <gatis.paeglis@qt.io>
GCC with LTO sees through our hack:
qprintdialog_unix.cpp:212:7: warning: type ‘struct QPrintDialogPrivate’ violates the C++ One Definition Rule [-Wodr]
qabstractprintdialog.cpp:49:7: note: a different type is defined in another translation unit
This hack was there so that the QPrintDialog functions in
qabstractprintdialog.cpp could use the d pointer. So instead of hacking
around the issue, just use the class that this file has access to:
QAbstractPrintDialogPrivate.
Change-Id: I3840d727dee443318644fffd1528e2e8b814e983
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
Placing a boolean between two pointers means we'll have a 7-byte padding
hole (64-bit). So move it to the bottom of the class, consuming the tail
padding that needs to be there anyway on 64-bit systems.
On 32-bit Unix systems, the better place would be at the top, as the
parent class (QDialogPrivate) has a 3-byte tail padding. But that's
fragile, as QDialogPrivate can change, doesn't apply to MSVC's ABI and
doesn't gain us anything on 64-bit.
Change-Id: I3840d727dee443318644fffd1528e4f05f4142bd
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
The output directory may be the same as the source directory in
the case of an in-source build, but Xcode treats the SYMROOT as
a build directory, and automatically excludes it from Time Machine
backups, which may result in not backing up sources.
Instead we map SYMROOT to an .xcode subdirectory of the output
directory, and then use CONFIGURATION_BUILD_DIR to make sure
the final build targets end up where they used to.
Task-number: QTBUG-52474
Change-Id: I3852ca9088e75ca62fca4c1217b5485175d9436f
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@qt.io>
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
Set the devicePixelRatio to 1 on the scaled down pixmap.
A scaled down version of the high DPI pixmap is added to the NSCursor in
addition to the high DPI one, but the devicePixelRatio must be set
correctly on the smaller of the two for macOS to pick the right image
to use on a high resolution display (retina).
This change also fixes the problem that only a high DPI custom cursor
with a hotspot in the upper left quadrant is applied by macOS. I suspect
that the NSCursor was discarded by macOS, because the hotspot was
outside the device independent bounds of the smaller scaled image.
Task-number: QTBUG-52211
Change-Id: I7e552e8f62f5255dd3786da44b2f619f6790c37a
Reviewed-by: Morten Johan Sørvig <morten.sorvig@qt.io>
the dependency paths are fixified against the output directory, so we
must resolve them accordingly.
Change-Id: Id92750aad358153bd2db5daca3194c54eda58dbb
Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
this variable hasn't been used in any syncqt.profile for many years, and
if it would, it would cause quite some side effects.
Change-Id: Ia936eff8f6a2a801fc644eee991821165d51dc1f
Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
Add calls to changeKeyboard() to the QWindowsKeyMapper constructor
and the handling of WM_INPUTLANGCHANGE so that the locale is
correctly initialized and changes are processed.
Change-Id: Ia30d8c6434ca85165e4882240ae16f9a75dcf4ff
Reviewed-by: Oliver Wolff <oliver.wolff@qt.io>
Remove the hardcoded mapping from the key table for fallback keys
since the keys are locale-dependent.
Task-number: QTBUG-57992
Change-Id: I016ab5f7f7e8abfd30f6416d2e7597db7deecb9b
Reviewed-by: Oliver Wolff <oliver.wolff@qt.io>
A <br> has a new-line but no text, so be able to handle no lines.
Task-number: QTBUG-60853
Change-Id: I3d4dbd529114bbe8afe760c3622b52446202ec7c
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@qt.io>
Some of our examples, and perhaps also some applications,
call requestUpdate() immediately after producing a
frame. This can cause Cocoa to immediately start
(trying to) draw a new frame without processing e.g.
input events.
This should (and will) be handled by rate limiting
updates with CVDisplayLink. In the mean time fall back
to using the base class QPlatformWindow implementation,
which is implemented using a timer, which will allow
for input event processing.
Change-Id: Ic2541f344b2f4018d785404a06274959a7bad2df
Reviewed-by: Gabriel de Dietrich <gabriel.dedietrich@qt.io>
According to Apple's documentation, there's no delegate
in NSFontManager. We set its target instead. The action
is changeFont: by default.
Change-Id: I8c01bfa97c78dd8097f38c27353748d13f51489f
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
We also ask the parent style to draw the header label
instead of the base style which is more likely to respect
other text related parameters.
Change-Id: I6dd658fa4d016a76d7c450478dc42f07e4b807c4
Task-number: QTBUG-33855
Task-number: QTBUG-37153
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
The NEON implementation uses rsqrt and thus can not be taken on 0, so
replace the minimum with something close to zero instead of zero.
Task-number: QTBUG-59961
Change-Id: Ia39e45be675b056c1e22900495ce9ba4e8b70e5f
Reviewed-by: Erik Verbruggen <erik.verbruggen@qt.io>
This was originally added so that you could replace a T with
QAtomicInteger<T> in the same class and still keep ABI. However, for
legacy reasons, on 32-bit x86, types larger than 4 bytes keep an old
1990s alignment of only 4 bytes, but modern std::atomic<T> for those 8-
byte types enforces an alignment of 8 bytes. Therefore, the requirement
to keep alignment is not possible to guarantee.
In other words: you may not replace T with QAtomicInteger<T> or
std::atomic<T> and assume no ABI breakages in all platforms.
This is a requirement to implement atomicity. An 8-byte type aligned to
only a 4-byte boundary could cross a 16-byte boundary or, worse, cross a
cacheline boundary. Crossing the 16-byte boundary could be bad on some
processors, but crossing the cacheline boundary (addresses ending in
0x3C, 0x7C, 0xCC and 0xFC, or 4 out of 64 possible addresses or 6.25%)
is always bad: the CPUs cannot guarantee an atomic load or store
operation.
See also <https://gcc.gnu.org/bugzilla/show_bug.cgi?id=71660>.
Task-number: QTBUG-67858
Change-Id: If90a92b041d3442fa0a4fffd15283e4615474582
Reviewed-by: Frederik Gladhorn <frederik.gladhorn@qt.io>
We need to mark the object as deleted before destroying
it's declarative data, otherwise all sorts of bad things
can happen.
This fixes the qwidgetsinqml autotest in qtdeclarative.
Change-Id: I05a645ebe1ca7a50c8927e3dbd9ebb5aaf369a71
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
(cherry picked from commit 3e91625b58)
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
There had been a fake declaration for qEnvironmentVariable() in qglobal.h
thaqt was only visible to QDoc. It was removed in favor of documenting
both the actual declarations of qEnvironmentVariable(), one with a 2nd
parameter for passing a defualt value and one without that parameter.
But the one without the default value parameter was marked internal, so
it didn't appear in the docs.
When both functions were documented with a shared comment, a bug in
QDoc was revealed, because these functions are global, while the shared
comment functionality had only been implemented for class member functions.
Now the shared comment functionality has been implemented for global
functions, so these two functions are now documented with a shared
comment. We can, of course, reintroduce the #ifdef QCLANG_QDOC trick, if
that is pre3ferred.
Change-Id: I41d85def5daa3215a995d7697d064dfae37e8b2a
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
That check is flaky on Windows. It doesn't seem to be testing Qt functionality.
I also don't see CreateFile2() documentation mentioning any guarantees that
opening the same file twice would give the same HANDLE each time.
Change-Id: Ica2e60571ae9fc39bf822803a2a9dd6add8323d7
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
This fixes qdoc warnings introduced by 67d5f79fe6.
Change-Id: I4b199e6243d9a7706befe4bc9549c78c11026d9e
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
We're relying on C++11 since a while, so lets not advertise creating
lists of strings with operator<<() anymore.
Change-Id: I14a3442ff852ac2c106d90c63504eb9ebb737609
Reviewed-by: Leena Miettinen <riitta-leena.miettinen@qt.io>
Reviewed-by: hjk <hjk@qt.io>
