QTBUG_10735_crashWithDialog started to show flakyness recently - it crashes,
but not every time (which fits the definition of UB perfectly).
While the test itself is doing weird things and puts our event dispatcher
into a weird state, our API allows to:
1. Using QDialog to enter event loop (with runModalSession under the hood), then ...
2. to call from a slot (e.g. timer-attached) QApplication::closeAllWindows() while ...
3. we are still inside that special loop and using the 'session' object, thus ...
4. on the next iteration with [NSApp runModalSession:session] we'll re-use already released
session (released by endModalSession which in turn was called indirectly by closeAllWindows).
And Cocoa gives us a warning/hint: "Use of freed session detected. Do not call
runModalSession: after calling endModalSesion:."
Task-number: QTBUG-62589
Change-Id: Ie651cee1fba43cfd2b0fc44af5eddc5fd52e2907
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
Reviewed-by: Morten Johan Sørvig <morten.sorvig@qt.io>
Reviewed-by: Gabriel de Dietrich <gabriel.dedietrich@qt.io>
This reverts commit 3d5bf00f18.
This change needs to be reverted because Windows 10 Creator's
Update doesn't fail on this test anymore during CI runs.
Reason for this is unknown.
Change-Id: I9f1c88606c97afc5952af34e04310612b783a9c2
Reviewed-by: Tony Sarajärvi <tony.sarajarvi@qt.io>
Delegating the border painting to qDrawPlainRect ensures that there are
no off-by-one pixel issues.
Task-number: QTBUG-61849
Change-Id: I56dc849da80fad00d02a0f9c60dbb621e6de7c48
Reviewed-by: Michael Winkelmann <michael.winkelmann@qt.io>
Reviewed-by: J-P Nurmi <jpnurmi@qt.io>
Reviewed-by: Alessandro Portale <alessandro.portale@qt.io>
Use QImageReader::supportsOption() instead of ::supportsAnimation(),
since the former checks what the handler supports in general, not just
the particular device.
Task-number: QTBUG-61642
Change-Id: I57db24425b4fd8821446659936e6a8ca55008921
Reviewed-by: J-P Nurmi <jpnurmi@qt.io>
If the style changes after the browser already jumped to an anchor,
the resulting browser position will be messed up.
So, after we changed the style we need to make sure that we jump
to the anchor again.
Since browsers do not jump to anchors that they already jumped to,
that means first jumping to the top, then to the actual anchor.
Task-number: QTCREATORBUG-18448
Change-Id: I86c736adab6940903276f8a896b4054ddae11ebe
Reviewed-by: Topi Reiniö <topi.reinio@qt.io>
Dialogs and Sheets by default have the WindowsContextHelpButtonHint
set, which adds a question mark button to dialogs on Windows. This
button then triggers the 'What's this' mode by changing the cursor,
and letting the user explore the UI by showing whatsThis tooltips.
Anyhow, the paradigm is little used today and a lot of applications
do not set any whatsThis properties, leaving the mode pretty
non-functional. It's therefore common to explicitly remove the
WindowsContextHelpButtonHint from dialogs. However, this has to
be done for _every_ dialog.
Instead, this patch adds a global application flag to not set the
WindowsContextHelpButtonHint by default. This allows developers to
already buy into the Qt 6 behavior, where the flag will not be set
anymore by default.
[ChangeLog][QtWidgets] Added AA_DisableWindowContextHelpButton
attribute. Setting this attribute globally prevents the automatic
"What's this" button on dialogs on Windows
(WindowsContextHelpButtonHint).
Change-Id: I497a79575f222c78b2d5d051a6de346b231f72d3
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
It's true that formattedDataSize is useful for formatting file sizes
from QFileInfo, but these links don't make sense here, and also cause
qdoc warnings due to lack of class scoping.
Change-Id: I9dd28200aa9d0da048db0c02ac66dc20c1b42e5c
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
nullptr can be used directly in the Qt code since Qt 5.7.
Use it in generated code for consistency.
Change-Id: I249aeaf0a39b46ce1106b29d3ea4569a399908b7
Reviewed-by: Jarek Kobus <jaroslaw.kobus@qt.io>
CID 158429 (#1 of 1): Dereference before null check (REVERSE_INULL)
check_after_deref: Null-checking ctx suggests that it may be null, but it has already been dereferenced on all paths leading to the check.
Coverity-Id: 158429
Change-Id: I6d202599e962dae5ea4a45401f34237dd496d38b
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
This amends patch f27d1ccbb2.
Change-Id: I4c7a390a5f2cdd3307007c7b6708692c36f861b4
Task-number: QTBUG-62396
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
Define the lib dependencies for corelib in corelib.pro, where they
belong.
Change-Id: I973d3b0c571782d869b27dea243e899db4dddc43
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@qt.io>
This was added in c6612de3 for WEC7 which we do not support anymore.
Change-Id: I329374bb8375d629a6f7619236371c0fc953792d
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@qt.io>
It is currently impossible to get output from autotests in Visual Studio
or Qt Creator when running under the debugger. Qt Creator's cdb
integration cannot distinguish between the inferior's console output and
cdb's console output. If the inferior's output came from
OutputDebugString we'd be able to catch and display it.
Pave a way to force QTestLib's logging facility use OutputDebugString.
Task-number: QTCREATORBUG-16161
Change-Id: Iccd69c283626266ee4384a6163a8b72bb0e7df27
Reviewed-by: hjk <hjk@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
It was already done correctly in the GCC generators, but lacked in MSVC.
Task-number: QTBUG-11117
Change-Id: I5e6c2e4802dbe33c0f15c46a227a08c3f0cc5707
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@qt.io>
Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
This prevents unnecessarily removing and reinstalling the completer
as event filter on the same widget.
This does not prevent what's going on in QComboBox::focusInEvent(),
where we'd set the line edit as widget just to immediately override
it with combo box itself.
Change-Id: I70c081a920f4daf4d7560e5cd7158e4070042d42
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
With MSG_TRUNC on Linux, recv/recvfrom/recvmsg will return the full size
of the datagram, even if it won't fit the buffer you passed. On Darwin,
we have a getsockopt() option to get that value.
Change-Id: I6e9274c1e7444ad48c81fffd14da7d5b93815f90
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
I'm getting crashes in Akonadi processes due to libproxy. I don't have
direct evidence that this was caused by a threading condition, but it's
clear from the source code of libproxy that the plugins it runs for
expanding PAC scripts are not thread-safe. To overcome this problem, we
only run libproxy functions in one thread only.
#0 0x00007f745f0ac1d8 in JSC::HeapTimer::timerDidFire() () at /usr/lib64/libjavascriptcoregtk-4.0.so.18
#1 0x00007f745f0ac287 in () at /usr/lib64/libjavascriptcoregtk-4.0.so.18
#2 0x00007f748e5ae9c5 in g_main_context_dispatch () at /usr/lib64/libglib-2.0.so.0
#3 0x00007f748e5aed88 in () at /usr/lib64/libglib-2.0.so.0
#4 0x00007f748e5aee1c in g_main_context_iteration () at /usr/lib64/libglib-2.0.so.0
#5 0x00007f7494f4268f in QEventDispatcherGlib::processEvents(QFlags<QEventLoop::ProcessEventsFlag>) () at /usr/lib64/libQt5Core.so.5
#6 0x00007f7494eeb35a in QEventLoop::exec(QFlags<QEventLoop::ProcessEventsFlag>) () at /usr/lib64/libQt5Core.so.5
#7 0x00007f7494d1b31a in QThread::exec() () at /usr/lib64/libQt5Core.so.5
#8 0x00007f7494d1fd2e in () at /usr/lib64/libQt5Core.so.5
#9 0x00007f74913174e7 in start_thread () at /lib64/libpthread.so.0
The pacrunner implementation of libproxy uses libdbus-1 which
(officially) is thread-safe, but experience tells that it has
problems. Since it is not running a JS engine, we don't need a thread,
but we do need to lock around it.
Change-Id: I84e45059a888497fb55ffffd14d2f638f21e807d
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
This patch avoids intermediate copying of the data by using a
new QIODevice::skip() API.
Change-Id: I358a83b21e2a58b59481745162fce1894549c0ac
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Reviewed-by: Markus Goetz (Woboq GmbH) <markus@woboq.com>
Without TLS (and thus ALPN/NPN negotiation) HTTP/2 requires
a protocol upgrade procedure, as described in RFC 7540, 3.2.
We start as HTTP/1.1 (and thus we create QHttpProtocolHandler first),
augmenting the headers we send with 'Upgrade: h2c'. In case
we receive HTTP/1.1 response with status code 101 ('Switching
Protocols'), we continue as HTTP/2 session, creating QHttp2ProtocolHandler
and pretending the first request we sent was HTTP/2 request
on a real HTTP/2 stream. If the first response is something different
from 101, we continue as HTTP/1.1. This change also required
auto-test update: our toy-server now has to respond to
the initial HTTP/1.1 request on a platform without ALPN/NPN.
As a bonus a subtle flakyness in 'goaway' auto-test went
away (well, it was fixed).
[ChangeLog][QtNetwork][HTTP/2] In case of clear text HTTP/2 we
now initiate a required protocol upgrade procedure instead of
'H2Direct' connection.
Task-number: QTBUG-61397
Change-Id: I573fa304fdaf661490159037dc47775d97c8ea5b
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
Consider the following:
/root/target - a file
/root/path/link -> ../target
/root/path/other/exe - executable
Running from /root/path/other.
exe is:
#include <QDebug>
#include <QFileInfo>
int main()
{
qDebug() << QFileInfo("../link").symLinkTarget()
return 0;
}
The link references /root/target, but the current output is
/root/path/target.
The link doesn't depend on the PWD. It depends on its own directory.
Change-Id: I61e95018154a75e0e0d795ee801068e18870a5df
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Also add import script.
The remaining diff to clean 1.6.32 is archived in the qtpatches.diff file.
[ChangeLog][Third-Party Code] libpng was updated to version 1.6.32
Change-Id: I1b4b78e39a6eb098d1b66c2528e47544bd9e6713
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
the replacement value may well constitute the whole output string - this
is in fact common, given this rather typical usage pattern:
BAR = $$replace(FOO, -flag, -otherflag)
this must be considered when constructing the return value.
compare 3c8134958c.
as of now, this is irrelevant, as QString::replace(QRegExp, QString) will
always memcpy the replacement into a detached copy of the target, but one
never knows.
Change-Id: Ia1f271f45023746040fc28ce6d88a6609e05e5c2
Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
technically, we should not rely on the i/o classes not storing the
strings beyond the instantiated object's life time.
Change-Id: I0990769b3cf86860184869036c096c531160e9be
Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
property values are de-facto guaranteed to be backed by full QStrings,
so there is nothing to be gained from using the raw data optimization,
while doing so risks raw data leaks.
Change-Id: I3d43da9aaadd4d5811c4b1a9d7ac734049da423c
Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
it's not clear why detaching would be necessary; there is no danger of a
raw data leak here.
concatenating a QStringRef with a non-empty QLatin1String (the only
expected use of this overload) will yield a detached QString anyway, so
this makes little difference in practice.
amends f137957e08.
Change-Id: I521c0e89a8b0c1ae62b1450e81b0ae91a931bcfa
Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
no m_tmp is involved any more in this code path; it uses QStringRef.
amends 11d957d043.
Change-Id: Ib272d61edfb150a549c5e6a9a60d53502702e802
Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
the m_tmp array is a member, so the index toggle for accessing it also
needs to be one - otherwise, odd iteration counts will defeat the
mechanism.
Change-Id: If7a800ed5a4b4168625daf1ebbd5d2d164569d8e
Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
... in $$basename(), $$dirname(), and contains(). the latter case is
marginal, as it only applies to mutuals which are regexes, which i don't
remember ever seeing used.
QRegExp saves a copy of the matched string, so it's necessary to
alternate between two temporaries to avoid detaching. we already
did that in most places.
Change-Id: I97b8294585c17c76d1756f83971f42cb88353af0
Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
in most cases, the main advantage is not using toQString(m_tmp), which
reduces the possibility of raw data leaks. in cases where we used
toQString() without temporary, this is a slight optimization.
Change-Id: Ib343acffd383aa2c4fefab75fb52762fb534dfc6
Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
The code was factored out in 0058f00b6 to be shared with Wayland, but
the Wayland platform plugin now lives in its own repository (as of
75a66c3d), and doesn't support running on macOS anymore.
Change-Id: Ied54f5f3e0a3e6fdaaedfcd456a140b46706f33c
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
QtWidgets in particular expects a show event before any resize or expose
events. QWindow::setVisible creates the platform window before sending
the show event, so we block sending any resize events during platform
window creation of QCocoaWindow, which means we miss out on the resize
event that indicates that the window is maximized.
We should probably teach QWindow::setVisible to send the show event
before creating the window, but until then we work around it by sending
an explicit resize event in QCocoaWindow::setVisible.
Task-number: QTBUG-62774
Change-Id: I6d512c9775986bcbcbf5c02a9e49984015fa9782
Reviewed-by: Morten Johan Sørvig <morten.sorvig@qt.io>
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
When a ObjC++ QObject subclass is listed in the regular HEADERS, qmake
creates a .cpp file. The moc file will then fail to compile, as it
requries ObjC++ headers. Using Q_FORWARD_DECLARE_OBJC_CLASS() can be
used to let the class be parsed by The compiler, but link will still
fail, as the generated methods (e.g. signals) must be built with ObjC++
compiler, in case they have ObjC parameters:
Q_FORWARD_DECLARE_OBJC_CLASS(NSString);
class MyClass: public QObject {
Q_OBJECT
signals:
void objcSignal(NSString * myObj);
};
The canonical workaround for that is including the .cpp file into the
corresponding .mm file. This also offers a compilation speed advantage,
but is somewhat counter-intuitive.
Therefore, we introduce a separate variable which instructs moc to create
.mm files directly.
Task-number: QTBUG-1581
Change-Id: Ia98af58006efd168ea37f3a63c396979e7e81baa
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@qt.io>
The definition of isTouchScreen() is protected with XCB_USE_XINPUT22 so
the implementation needs to have this too.
Task-number: QTBUG-62226
Change-Id: Icc3de01a6cb1299b43e56fc9f77833764131ca4b
Reviewed-by: Gatis Paeglis <gatis.paeglis@qt.io>
When QMAKE_* variable assignments were passed to the configure line
they would cause the current contents of the private pro output to be
overwritten. This would cause anything added to it before the QMAKE_*
variable assignments to be parsed to be lost.
Change-Id: Idcb8cad5f07cbb96b4da204384f5618b95b375b0
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@qt.io>
Originally when the QPrinter was created it would create the engine with
the default printer and then change it afterwards even though the
desired printer may already be known here. So by passing the printer
name we ensure that it is initialized with the desired one right away.
Task-number: QTBUG-62221
Change-Id: Iaa90243708b57bf89354a527a982ac45c991f603
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
By sharing the implementations of the methods between QNSWindow and
QNSPanel we don't need a helper, and can remove duplicated code. This
duplication would expand in the future, as for each method added to
the QNSWindowProtocol, we would have to add forwarding functions
in both QNSWindow and QNSPanel, forwarding to QNSWindowHelper, and
then two more functions in QNSWindow and QNSPanel in case we wanted
to call super from the helper, similar to [QNSWindow superSendEvent].
The only snag is that calls to super are hard-coded to a specific
superclass during complication, so we provide our wrapper for
objc_msgSendSuper that resolves the superclass at runtime.
The helper class QSendSuperHelper provides compile time implicit
instantiation of the right template without having to provide
the return type as a template argument, via operator T and
a fallback for the case of no return type via the destructor.
Change-Id: Iaf13f27675d90f884470f5005270ea0d9d0316f3
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
The default on iOS has been raster for two years now, as of 3e892e4a97,
and we haven't seen any major performance regressions that would warrant
keeping the OpenGL based code-path alive.
This includes the default surface format, which was ony set so that
QPainter clip regions would work when using the GL backed backing store.
Change-Id: I37b880a758b9c3fad1f23ae60268629ffbe9bc3e
Reviewed-by: Richard Moe Gustavsen <richard.gustavsen@qt.io>
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
Dongmei Wang
Timur Pocheptsov
Joni Jäntti
Alessandro Portale
Eirik Aavitsland
Eike Ziller
Friedemann Kleint
Kai Koehne
Shawn Rutledge
Thiago Macieira
Alexander Volkov
Robert Loehning
Jesus Fernandez
Gabriel de Dietrich
Joerg Bornemann
Orgad Shaneh
Simo Fält
Alex Trotsenko
André Klitzing
Oswald Buddenhagen
Tor Arne Vestbø
Francois Ferrand
Andy Shaw