With the current implementation, the QXcbConnection::m_focusWindow
was holding a dangling pointer during the focus transition from a
dying modal window to other modal window.
1) QXcbConnection::m_focusWindow holds a pointer to A;
2) A is closed;
3) relayFocusToModalWindow B;
=> m_focusWindow now points to a dead window.
4) We get a reply back from WM in a respone to
relayFocusToModalWindow (_NET_ACTIVE_WINDOW)
=> m_focusWindow now points to a valid window.
The fix is to update m_focusWindow to nullptr, when the current
focus window was destroyed and the new focus window has not been
set yet by WM.
This patch actually solves a more general case - whenever we get a
focus-out event, we should set m_focusWindow to nullptr. It is ok
for none of the windows to be in-focus while focus transition is
happening. The focusInPeeker will make sure to "optimize out" (when possible)
this no-window-in-focus state, and GUI won't be bothered by this
extra event. This is how things were working before relayFocusToModalWindow
was introduced. Having a focus-relay mechanism in-between is ok,
but it should not have changed the original behavior.
Task-number: QTBUG-48391
Task-number: QTBUG-55197
Change-Id: I6fdda9de73f999dad84000059ce4b89c0d1a964c
Reviewed-by: Laszlo Agocs <laszlo.agocs@qt.io>
When handling QActionEvent / ActionAdded, QActionEvent::before()
is 0 for the common 'append' case. Do not try to find a
QPlatformMenuItem for this.
Change-Id: I535b675fda6a2812b61eb8d7eec01da378ece413
Reviewed-by: Dmitry Shachnev <mitya57@gmail.com>
Reviewed-by: Albert Astals Cid <albert.astals@canonical.com>
Reviewed-by: Shawn Rutledge <shawn.rutledge@qt.io>
Commit 0f1b6acab7 added the detection, but
not the reporting in configure's output. This change makes them be
reported alongside F16C.
Change-Id: I27b55fdf514247549455fffd14b1711ca47b1d9a
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
Clang implements the _cvtss_sh intrinsic by way of a macro, which
uses a C99 extension and that's not allowed in C++ mode:
float16.h:119:11: error: compound literals are a C99-specific feature [-Werror,-Wc99-extensions]
/usr/bin/../lib64/clang/3.9.1/include/f16cintrin.h:76:55: note: expanded from macro '_cvtss_sh'
Reported at https://bugs.llvm.org/show_bug.cgi?id=32491.
Change-Id: I27b55fdf514247549455fffd14b170df75dd4e1f
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>
The .pro file requires the QMAKE_CFLAGS_F16C to be set to something. So
set it to AVX, as the instructions require the VEX prefix anyway (ICC
has no dedicated option for just F16C).
Change-Id: I27b55fdf514247549455fffd14b171940afd35a2
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
The new overloads do not accept parameters for the invoked function, this
use case is handled by using lambda.
Overloads for non member function pointers and functors are separated as
the return type is not retrieved in the same way.
Move QSlotObjectBase, QSlotObject and QFunctorSlotObject from
qobject_impl.h to qobjectdefs_impl.h in order to make them available in
qobjectdefs.h.
Update autotests of previous overloads because of a soft break in source
compatibility: passing null literals (0, NULL, nullptr, etc.) for the
second parameter of invokeMethod() is not supported anymore.
[ChangeLog][QtCore] Added QMetaObject::invokeMethod() overloads for function
pointers.
Task-number: QTBUG-37253
Change-Id: I6fb67e086d315ae393ce32743c4eb1abd6cc9139
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
This helps to reduce code duplication as it will be used as the basis
of other reference-counted RAII template classes in the future.
Change-Id: I6e7e32373ec738f68d960afc759213610232b9bf
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
The non-static method QProcess::startDetached now supports nativeArguments.
[ChangeLog][QtCore][QProcess] Added the ability to specify native
arguments for detached processes on Windows.
Task-number: QTBUG-52405
Change-Id: Ic97c55caac7c1156e1412322000c1e40d80bdbbb
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
QMacStyle already sets SH_Workspace_FillSpaceOnMaximize, so this will
have the same behavior when QMacStyle is in use.
Change-Id: Ibe9f93b2ca94d84b0ce2d3a3078b7db6d5906042
Reviewed-by: Gabriel de Dietrich <gabriel.dedietrich@qt.io>
We still support buffering via stdio.h (FILE*), but QFSFileEngine itself
is not responsible for the buffering. Opening by file name (not FILE*
and not file descriptor) already added the Unbuffered flag.
Change-Id: I2bc52f3c7a574209b213fffd1498d4e3f78fe1bf
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
The configure-time detection (cxx11default) isn't enough if the compiler
can be changed. This is especially necessary if Qt is compiled with a
compiler that defaults to >= C++11 (e.g., GCC 6) and then the user
selects a compiler another compiler (e.g., Clang) via -spec option. In
that case, we'd miss adding the -std=c++11 or -std=gnu++11 option to the
command-line, causing the compilation to fail.
As a nice side-effect, even moc without moc_predefs.h will now get the
__cplusplus setting.
Task-number: QTBUG-58321
Change-Id: I74966ed02f674a7295f8fffd14a8be35da9640e1
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@qt.io>
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>
Some compilers are known to complain about this with a warning. GCC
complains about const on return values on -Wignored-qualifiers (enabled
at -Wextra), so it's not too much of a jump to assume that others do
too. Besides, this is not Qt Library API policy. As maintainer for
QtCore, I'm exercising my prerrogative in specifying certain unspecified
parts of the coding style, like I've done for constructor initializer
lists.
Since all the classes involved are exported (including QVector, through
derived classes), we can't remove the qualifier until Qt 6, since there
are compilers known to encode the qualifier in the mangled name
(suncc). I'm not introducing #ifdef to silence unknown compilers unless
we get an actual complaint.
Change-Id: I33850dcdb2ce4a47878efffd14a876edef843c46
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Implicit conversion from QByteArray to const char* works for most gtk
functions. But gtk_file_chooser_dialog_new() uses varargs and passing
the non-POD QByteArray through varargs does not work (it's UB).
Task-number: QTBUG-59763
Change-Id: I85f9323d99342896e6921cdeb85f5a1af7377b4f
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>
Starting a detached process with a custom process environment can now be
achieved by:
QProcess p;
p.setProgram("foo");
p.setProcessEnvironment(myEnv);
p.startDetached();
[ChangeLog][QtCore][QProcess] Added the ability to set a custom process
environment for detached processes.
Task-number: QTBUG-2284
Change-Id: I49406dffb64fa2aed41ea05cb271bd42eeabb729
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
The three static QProcess::startDetached overloads support a limited number of
features: program, arguments and working directory. To support more features of
QProcess (without adding a plethora of overloads) we add a non-static method
startDetached that can be used as follows:
QProcess p;
p.setProgram("cat");
p.setArguments("meow");
p.setWorkingDirectory("/tmp");
if (!p.startDetached())
qWarning("Cannot start process.");
We plan to add support for nativeArguments, processEnvironment,
standard{Output|Error}File and maybe more in subsequent commits.
[ChangeLog][QtCore][QProcess] Added non-static QProcess::startDetached
to support more features for detached processes.
Task-number: QTBUG-2058
Task-number: QTBUG-2284
Task-number: QTBUG-37656
Task-number: QTBUG-52405
Task-number: QTBUG-57687
Change-Id: If6fdd57ecb28cd13aa5fff566216a4177f81d339
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Rafael Roquetto <rafael.roquetto@kdab.com>
This change fixes several common qdoc warnings in this new software.
However, because vulkan/vulkan.h is not always available, a few typedefs
were also added to define some standard vulkan types for clang to see
when vulkan.h is not available.
Change-Id: Id3db8c0a59b0c415829c1873a868aba0e84fe21c
Reviewed-by: Laszlo Agocs <laszlo.agocs@qt.io>
clangqdoc must see the declaration in vulkan
regardless of whether Qt is configured with vulkan.
Change-Id: I91fcaa4c39121b70e10c4fc3e854e49d959a6c02
Reviewed-by: Martin Smith <martin.smith@qt.io>
This update removes the ! charaqcter from a few qdoc comment
markers for functions that should not have been documented.
Doing so eliminates 4 qdoc warnings.
Note that this update deliberately ignores a clangqdoc bug that
might eventually require teaching clangqdoc to store template
information with classes and functions so that qdoc can match
function specializations correctly and document many different
specializations when there is only one declaration. Currently,
it can't handle function specializations.
Change-Id: I5c1db98dabc802cce18f2fc3e64b82e15165ad4d
Reviewed-by: Martin Smith <martin.smith@qt.io>
Change-Id: If1d2cf175d51b3c02881e21937b0a2d33b78aadd
Reviewed-by: Anton Kudryavtsev <antkudr@mail.ru>
Reviewed-by: Ville Voutilainen <ville.voutilainen@qt.io>
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Remove most of the std::equal() tests that were used to determine
equality in pre-relational-operator-times again.
Amends a1421e4787.
Change-Id: Iff64808f5ac60861caee899d594b512b58046636
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Reviewed-by: Anton Kudryavtsev <antkudr@mail.ru>
Reviewed-by: Sean Harmer <sean.harmer@kdab.com>
Reviewed-by: Giuseppe D'Angelo <giuseppe.dangelo@kdab.com>
Introduce convenience function sessionManagerInteractionBlocked()
that is also present when session manager is disabled and
reduce #ifdef-ery accordingly.
Change-Id: Ifa211a3bf19901db669743a16acde6e6dc455f75
Reviewed-by: Oliver Wolff <oliver.wolff@qt.io>
This amends commit c20cd4d98e
Change-Id: I2fa903209694cd05e0dcf31e6f92d96ed934c91d
Reviewed-by: Stephan Binner <stephan.binner@basyskom.com>
Reviewed-by: Ulf Hermann <ulf.hermann@qt.io>
Makes the code slightly less awkward to deal with.
Change-Id: Ie98a04d812beecfbf0d19e77b8e784db783a19d6
Reviewed-by: Konstantin Ritt <ritt.ks@gmail.com>
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
The SHA3 family is a modified version of Keccak. We were
incorrectly calculating Keccak (and even *testing* Keccak!),
but claiming it was SHA3.
To actually calculate SHA3, we need invoke Keccak on the original
message followed by the two bits sequence 0b01, cf. §6.1 [1].
[1] http://dx.doi.org/10.6028/NIST.FIPS.202
[ChangeLog][QtCore][QCryptographicHash] QCryptographicHash now
properly calculates SHA3 message digests. Before, when asked
to calculate a SHA3 digest, it calculated a Keccak digest instead.
Task-number: QTBUG-59770
Change-Id: Iae694d1a1668aa676922e3e00a292cddc30d3e0d
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
When using anything but the mouse, there is no release event sent.
Instead one needs to subscribe to the ICorePointerRedirector.
Task-number: QTBUG-58781
Change-Id: I664ba24bd89ea9f513f8f11b385e2f06ece42fd0
Reviewed-by: Oliver Wolff <oliver.wolff@qt.io>
... by delegating to QConcatenable<const char[N]>.
The only thing that varied was the nested type alias 'type', which
therefore got retained.
Change-Id: I202f899034e1ddd23c6d1978a31be5eb7c195697
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Two changes are needed to pass tst_Selftest on QEMU
1. Pass QEMU specific env variables to the subtests
2. Ignore output on stderr on some tests when running on QEMU
Change-Id: Ie1f722fd183aac5973e87d408005e06cbafcde17
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
Unix mmap(2) system calls do allow for mapping beyond the end of the
file, though what happens after you try to dereference the pointers it
gives is unspecified. POSIX[1] says that implementations shouldn't allow
it:
The system shall always zero-fill any partial page at the end of an
object. Further, the system shall never write out any modified portions
of the last page of an object which are beyond its end. References
within the address range starting at pa and continuing for len bytes to
whole pages following the end of an object shall result in delivery of
a SIGBUS signal.
However, Linux allows this in read-write mode and extends the file
(depending on the filesystem).
Windows MapViewOfFile never allows mapping beyond the end.
[1] http://pubs.opengroup.org/onlinepubs/9699919799/functions/mmap.html
Change-Id: Ie67d35dff21147e99ad9fffd14acc8d9a1a0c38d
Reviewed-by: Sami Nurmenniemi <sami.nurmenniemi@qt.io>
Reviewed-by: Teemu Holappa <teemu.holappa@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
tst_QWidget::updateWhileMinimized has been failing on Ubuntu 14.04
and was already blacklisted there. Now we extend it to cover Ubuntu
16.04.
Task-number: QTBUG-46116
Change-Id: I6758657cca46bb4c76cddb0298f9b87b8a43655b
Reviewed-by: Liang Qi <liang.qi@qt.io>
The font database uses UIKit symbols on all UIKit platforms and thus it
must be linked to. Also, limit the missing CoreText.h workaround so
that it is not applied for versions of Xcode where the bug is fixed.
Change-Id: Ia2a94baba6b0df618604a0bc3092fbd2a48d5aea
Reviewed-by: Jake Petroules <jake.petroules@qt.io>
Mixing different protocols on client-server sockets works differently
on 10.11, making previously successful handshakes failing now.
Failure is specific to 10.11 with SecureTransport.
Change-Id: I35374b40fa3d167802775b526cf6465ae78749cf
Task-number: QTBUG-48860
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
The code here performs a std::mismatch() on the normalized parts of
'root' and 'path', returning whether 'path' is a prefix of 'root'. If
'path' is a proper prefix of 'root', returns the first 'root' part
after 'path', in '*match'.
Instead of using QString::splitRef(), perform the splitting as we go,
using this class' own tool, QStringSplitter.
The goal here is to avoid the heap allocation caused by the QVector
return of splitRef(), of course. But even though using QStringSplitter
means the splitting is now purely inline, text size in optimized GCC 7
Linux AMD64 builds increases by only 24B.
Change-Id: I396cc6a30306940220c9e783c31506e6ac17b784
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
TLD lookup is performed in up to three steps, which are decribed in a
code comment. For the purposes of this patch, it suffices to mention
that the last two steps involved a QStringBuilder expression.
Avoid having to build these strings by distinguishing between the
lookups using an enum instead of a string prefix. Since qt_hash()
calls can now be chained, form the hash value by hashing the prefix
character and the domain (part) separately.
Also perform the expensive fromUtf8() call only when the string prefix
(if any) matched. Once we have nothrow UTF-8/UTF-16 comparison
operators, qIsEffectiveTLD() will no longer allocate memory, and run
at full speed. Not there, yet, though.
I have full confidence that the optimizer will have field day with
this, inline all three containsTLDEntry() calls, and constant-fold the
checks inside the loop. GCC does, so I didn't try to do it by hand.
Change-Id: I1a7647ff52b6742310928ad859d851d1f0c6bdc3
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
There are some callers of qt_hash that first build a string just to hash it.
By allowing to pass an initial value for 'h', we can chain qt_hash() calls
to avoid having to allocate memory just to hash a two-part string.
Change-Id: Ifaca82d47b2fb8c707912342c3ddd84f91e70267
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
The fix for dangling pointers in 524f39 caused some problems when the
QByteArray was implicitly converted to const gchar*. This is fixed
by wrapping the QString in question in qUtf8Printable where possible
and removing the former convenience method.
Task-number: QTBUG-59692
Change-Id: I5abcf42e1c23b12c7a5c4c195d801f377fe9d138
Reviewed-by: J-P Nurmi <jpnurmi@qt.io>
Apply white text only for macOS style.
Amends 2c0033983b
Task-number: QTBUG-59784
Change-Id: I9e66e929699efd715ed4565394f1aba763aeb32a
Reviewed-by: Jake Petroules <jake.petroules@qt.io>
GCC bug 68949 causes tst_QGraphicsGridLayout and tst_QGraphicsLinearLayout
to fail on 5.2.x/5.3.x: https://gcc.gnu.org/bugzilla/show_bug.cgi?id=68949.
This change adds aggregate initialization to QSizeF arrays to work around
the bug. The bug was discovered when compiling and running tests on ARM
with GCC 5.3.0.
Change-Id: I9ecf7b032b6ca1477c29dca3bd7d0ec8d69a0454
Reviewed-by: Ville Voutilainen <ville.voutilainen@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
