When a text block is empty, and we are adding preedit text to it,
we need to merge the format of the preedit text with the current
format of the cursor, otherwise we will use a default format and
then suddenly switch to the proper one when the text is committed.
The reason this becomes a bit complex is that there are no rules
preventing someone from using several ime attributes to specify
formats for isolated parts of the text, and no rules defining the
order of such attributes. So even if the common case is one
text format attribute for the entire string, we need to make sure
we also handle the other cases gracefully, e.g. when we are setting
different formats for different substrings and then providing these
out of order. To make sure we have these corner cases covered, we
also add a set of autotests.
[ChangeLog][Qt Widgets][TextEdit] Fixed initial char format of
input method text as it is in pre-edit mode.
Task-number: QTBUG-59196
Change-Id: I1e37928e3bd1395fec1b5591908d4c69b84eb618
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
psql driver was previously using QDateTime::toString() function which is locale-depndent.
Task-number: QTBUG-59524
Change-Id: I7f50f95b5c82e66476abfee24ad28b78b3257041
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Commit e0ea0f6178 optimized QChar <->
QString(Ref) comparisons by adding more overloads to avoid creating
QStrings from QChars just to compare them.
But these new overloads made existing comparisons to QChar ambiguous.
This was known at the time for QChar/int comparisons.
It has since turned out that also comparing to '\0' is ambiguous,
ie. not comparing to int or char per se is ambiguous, but comparing to
nullptr constants is, because QString(const char*) is just as good a
candidate as QChar(char)/QChar(int).
Since we allow QString/QChar comparisons, it seems logical to solve
the problem by adding QChar<->nullptr overloads.
[ChangeLog][QtCore][QChar] Disambiguated comparisons with nullptr
constants such as '\0', which 5.8.0 broke. As a consequence,
QChar<->int comparisons are no longer deprecated, as this was a failed
attempt at fixing the ambiguity.
Change-Id: I680dd509c2286e96894e13078899dbe3b2dd83bc
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
The mmap64 functions in all Linux libc fail to properly check that the
value fits in the system call parameter. I guess the developers just
said "16 PB are enough for everyone"...
Change-Id: Ic39b2c4fd9c84522a8fafffd14ac91567ce09c09
Reviewed-by: Sami Nurmenniemi <sami.nurmenniemi@qt.io>
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
Timer IDs have been reused since Qt 4.5 or thereabouts, so just checking
if the timer ID is in the timer dictionary is an incorrect check: our
timer may have been deleted and replaced by another with the same ID.
Instead of deleting the WinTimerInfo object, let's just mark it as
unregistered by setting timerId to -1 and cooperate in deleting at the
appropriate places. Since unregisterTimer skips deleting if inTimerEvent
is true, the appropriate places are everywhere that set inTimerEvent to
true.
Change-Id: I057e93314e41372ae7a5ff93c467767c8a6d92ea
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
Reviewed-by: Oliver Wolff <oliver.wolff@qt.io>
Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
Much simpler, no memory allocation, no locking.
Change-Id: Iae839f6a131a4f0784bffffd14a9e70c7b62d9a7
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>
the View's position and metrics are in pixels, not points
Change-Id: I285d5378db98187f54019bff9b8a1cde05dc3b35
Reviewed-by: BogDan Vatra <bogdan@kdab.com>
QWindow::setVisible calls QWindowPrivate::applyCursor without checking
if screen() returns null. This patch adds a check in QWindowPrivate::applyCursor
that the screen is not null. Now that it is tested there, no need to test
it from the other caller (setCursor)
This patch should not change behavior of setCursor at all, it should only
fix the crash when coming from setVisible
Task-number: QTBUG-59528
Change-Id: I06bbdb4e04c02ac840ba637242d1f2cfde5bdd62
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
The constructor is not only disabled under QT_NO_CAST_FROM_ASCII,
but also under QT_RESTRICTED_CAST_FROM_ASCII.
Change-Id: I7bbaf2891913d5256dff7f80c49075ea3326155a
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
- recursion is not needed for walking up the parent chain;
- use camel-case, modal_window -> modalWindow;
Change-Id: I4b7697f2388fd16f11be67ba475bd63ad249d89e
Reviewed-by: Shawn Rutledge <shawn.rutledge@qt.io>
Made change to clear the hover index when the mouse leaves the widget.
This will ensure the component does not think the item still has the
mouse over it.
Task-number: QTBUG-46785
Change-Id: I34b7f0e171e9cf07ca23150af1b0e6e59a10a58a
Reviewed-by: David Faure <david.faure@kdab.com>
Cross-compilation on linux with mingw-w64 failed, since it provides
all libraries in lowercase name.
Change-Id: I8e82a4504d4ec9d3047af6b1d6a11945754f182d
Reviewed-by: Oliver Wolff <oliver.wolff@qt.io>
This corrects an issue where the file system paths returned for some
QStandardPaths values on Apple Platforms would be URL encoded, for
example having %20 instead of an actual space character.
Task-number: QTBUG-59389
Change-Id: I771a44eb20b756842c324ac6fc9bdc475ce84826
Reviewed-by: David Faure <david.faure@kdab.com>
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
Reviewed-by: Gabriel de Dietrich <gabriel.dedietrich@qt.io>
Stale Lock files in the future can happen in some situations. For exemple
two computers with different clocks access the same file system. It could
be that one of the timestamp is totaly off (several years into the future).
[ChangeLog][QtCore][QLockFile] Fixed a deadlock occurring if a corrupted
lock file's modification time is in the future.
Change-Id: I8dac98a0e898c76bcef67f8c195e126c996b6add
Reviewed-by: David Faure <david.faure@kdab.com>
Makes sure we don't use backslashes on Windows systems, which could lead
to the following warnings being emitted:
In file included from main.cpp:43:0:
repparser.h:2:10: warning: unknown escape sequence: '\.'
#line 57 "..\..\src\repparser\parser.g"
caused by lines like
#line 57 "..\..\src\\repparser\parser.g"
Change-Id: I6cfe0e39a2a58eb39f9d385ece30374bcfa09e05
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Reviewed-by: Brett Stottlemyer <bstottle@ford.com>
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@qt.io>
GCC 4.8 seems to get the failure memory order wrong when using the
overload that only accepts one memory order and produces errors such
as:
bits/atomic_base.h:577:70: error: failure memory model cannot be stronger than success memory model for '__atomic_compare_exchange'
return __atomic_compare_exchange_n(&_M_i, &__i1, __i2, 0, __m1, __m2);
^
(as seen on Android).
Fix by explicitly passing the failure orders corresponding to the
success orders, as specified by the standard:
relaxed → relaxed
release → relaxed
acquire → acquire
acq_rel → acquire
(cf. http://en.cppreference.com/w/cpp/atomic/atomic/compare_exchange).
Task-number: QTBUG-59399
Change-Id: If046e735888cf331d2d6506d8d5ca9aa7402f9ad
Reviewed-by: David Faure <david.faure@kdab.com>
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
If a later month-or-day were to have a name that's a prefix of an
earlier one's name, the code would have selected the longer name as
best match when the text matched is the shorter name, simply because
it found that one first. (Found, on Turkish Cuma(rtesi)? in Thiago's
recent new test, by reversing the loop that iterated the list.)
Make an exact match win and a match of a full name beat any prefix
match of the same length.
Change-Id: I8d954b83ccc25e4f47af2e558036d714685cef5e
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Decouple from the callers' offset into a larger list; just search for
an entry in a list, let the caller deal with the offset. Also, defer
a .tolower() to save the need to allocate a copy of each list entry.
Change-Id: I748d5214c2cc6dc592fe2bd41e3f8150f71c335b
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Since QPrinter can be used to write to PDF and this was working in
previous versions, there is no reason not to enable it here.
Task-number: QTBUG-58376
Change-Id: I5760b74881995679e8df657b7d770bba00a33551
Reviewed-by: Paul Olav Tvete <paul.tvete@qt.io>
Simplifies everything and avoids bugfixes in one not propagating to the
other.
Change-Id: I95c9e502ccc74af3bcf0fffd14a69f0cde60cc8c
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Both SPDY and HTTP/2 work with a single qhttpnetworkchannel (and this means one
socket per qhttpnetworkconnection). Normally, HTTP/1.1 connection is using up to 6
channels/sockets though. At the moment a failure to negotiate SPDY/HTTP/2 leaves us
with a downgraded HTTP/1.1 connection (with only one channel vs. default 6).
Since we initialize channels (and establish connections) in a 'lazy' manner
it's ok to pre-allocate all 6 channels and then either use 1 (if SPDY/HTTP/2
indeed was negotiated) or switch back to 6 in case of failure.
Change-Id: Ia6c3061463c4d634aaed05ce0dde47bfb5e24dd8
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
Some were 5s, others 1s. Pick one (the higher).
Change-Id: I81929d4f49c2e41b4d4b75c3e2bf8ff75af868ad
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
By default, it enabled only for Gnome
Task-number: QTBUG-59298
Change-Id: Ib44f52b3175ed1904f24ec9e21f00ea3f3334287
Reviewed-by: J-P Nurmi <jpnurmi@qt.io>
When the insertion into the cache fails then it will delete the entry
for us which already calls releaseKey(). So we should not call it a
second time.
Task-number: QTBUG-58259
Change-Id: I816c6f29ef97fe3a245f145c4faf1e0649f72dc5
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@qt.io>
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
ResultStore never actually exists, only ResutStoreBase does. So casting to
ResultStore<T> and calling its member functions is UB. Put the type dependent
function as template member functions within ResultStoreBase and so we don't
need QtPrivate::ResultStore anymore.
Same goes for the iterator.
Change-Id: I739b9d234ba2238977863df77fde3a4471a9abd2
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>
For the windows file system engine, we add an extra macro to use
library loading if configured to do so, but avoid it on WinRT, as
none of the symbols would be found.
We also QT_REQUIRE_CONFIG(library) in the library headers and
exclude the sources from the build if library loading is disabled.
This, in turn, makes it necessary to clean up some header inclusions.
Change-Id: I2b152cb5b47a2658996b6f4702b038536a5704ec
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@qt.io>
On some stupid systems, execve() may clear the handler but not clear
the SA_SIGINFO flag.
This change now requires that sa_handler and sa_sigaction be in a union
together. We can't operate otherwise.
Task-number: QTBUG-59246
Change-Id: I33850dcdb2ce4a47878efffd14a84b48a8f6b1e8
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
QLocale::matchingLocales() simply created each locale using the basic
data, without (unless the matching conditions stipulated Language C)
applying number-options hacks that it applies everywhere else, when
creating the C locale. Thus the C locale in its returned list (if it
wasn't the only entry) ended up with the default number options,
without omiting separators in numbers. Thus QLocale::c() didn't
actually appear as an entry in the list. Discovered while
investigating QTBUG-58947.
Added a dumb autotest that checks various ways of getting the C locale
do actually give us equal locale objects. Fixed matchingLocales() to
apply the same hack as is used elsewhere for the C locale.
Change-Id: I263f31da623052b63171f5b5a83c65802383df21
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
On a multi-display system wide submenu might either appear on wrong
screen or not appear at all (depending on the specific display
configuration).
Task-number: QTBUG-56917
Change-Id: I40013b0bee340a01ae1c08a5e074afa63da4dbfd
Reviewed-by: Gabriel de Dietrich <gabriel.dedietrich@qt.io>
Reviewed-by: Błażej Szczygieł <spaz16@wp.pl>
This patch fixes 2 issues related to wide menus:
1) Menu took on full screen height when menu width was larger than
screen width;
2) On a multi-display system wide menu might appear on wrong monitor
(not the one where show event was triggered).
The idea is we limit parent menu and all its submenus within the screen
where it was opened.
Note that this patch fixes only geometry-related issues and there are
also some style flaws which need to be addressed (for example, currently
the text does not elide if it doesn’t fit to the menu’s width).
Task-number: QTBUG-56917
Change-Id: I7e9ff4a48bf03060d76e34d33a13ad6cc890c133
Reviewed-by: Gabriel de Dietrich <gabriel.dedietrich@qt.io>
This fixes compiling an application using QVector and -Wshorten-64-to-32
on a 64-bit system without getting this warning:
... 5.8/clang_64/lib/QtCore.framework/Headers/qvector.h:695:18:
warning: implicit conversion loses integer precision: 'typename
iterator_traits<QString *>::difference_type' (aka 'long') to 'int'
[-Wshorten-64-to-32]
int offset = std::distance(d->begin(), before);
~~~~~~ ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
... 5.8/clang_64/lib/QtCore.framework/Headers/qvector.h:731:35:
warning: implicit conversion loses integer precision: 'long' to
'const int' [-Wshorten-64-to-32]
const int itemsToErase = aend - abegin;
~~~~~~~~~~~~ ~~~~~^~~~~~~~
... 5.8/clang_64/lib/QtCore.framework/Headers/qvector.h:740:39:
warning: implicit conversion loses integer precision: 'long' to
'const int' [-Wshorten-64-to-32]
const int itemsUntouched = abegin - d->begin();
~~~~~~~~~~~~~~ ~~~~~~~^~~~~~~~~~~~
Change-Id: I52d85908f4aac20c7e9ac8063ac760ce52f85541
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
We might get there because 'wid' is already partially deleted (called
from ~QObject). In that case, it is an undefined behavior to call a
QWidget member function on it. Use QObjectPrivate::get instead.
Change-Id: I6da314bf8385684d1332aa031a2d92012941303b
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>
As the destroyed() signal is emitted from ~QObject, it is not allowed
to use static_cast to a QAbstractButton on that pointer anymore.
And the qobject_cast will also fail which will keep a dangling pointer
in the hash.
Change-Id: If0d22fcc30cde87e771e70914c3afb04ea207289
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>
In fe97ecf408 we added support for
arbitrarily sized glyphs, since there is no guarantee that a glyph
will fit inside the em square. There was, however, a hardcoded max
size in the distance field generator, so for very tall glyphs we
could potentially get a crash.
[ChangeLog][QtGui][Text] Fixed crash for very tall glyphs
Task-number: QTBUG-57241
Change-Id: Id95c0f10d82a1294f7e7a51ac32e88a5a2e0a790
Reviewed-by: Yoann Lopes <yoann.lopes@qt.io>
Reviewed-by: Konstantin Ritt <ritt.ks@gmail.com>
This fixes the following Valgrind warning:
"Syscall param writev(vector[...]) points to uninitialised byte(s)
Uninitialised value was created by a stack allocation"
The xcb_send_event() requires all events to have 32 bytes.
It calls memcpy() on the passed in event. If the passed in
event is less than 32 bytes, memcpy() reaches into unrelated
memory. And as it turns out, this behavior is actually
described in the xcb_send_event function's documentation.
This patch adds a macro that declares an event for safe
usage with xcb_send_event.
Change-Id: Ifcaab5e9a3b52b7f64ac930b423e0c7798bbfedb
Done-with: Uli Schlachter
Task-number: QTBUG-56518
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>
This fixes compiling an application using QList and -Wshorten-64-to-32
on a 64-bit system without getting this warning:
... 5.8/clang_64/lib/QtCore.framework/Headers/qlist.h:897:26:
warning: implicit conversion loses integer precision: 'long' to 'int'
[-Wshorten-64-to-32]
int removedCount = e - n;
~~~~~~~~~~~~ ~~^~~
Change-Id: I688ed086805c431821c2ee6078fa5aeb631e7a07
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>
This fixes compiling an application using QTimer and -Wshorten-64-to-32
on a 64-bit system without getting this warning:
... 5.8/clang_64/lib/QtCore.framework/Headers/qtimer.h:171:21:
warning: implicit conversion loses integer precision: 'rep'
(aka 'long long') to 'int' [-Wshorten-64-to-32]
setInterval(value.count());
~~~~~~~~~~~ ^~~~~~~~~~~~~
Change-Id: I3e0407a7193c841308f7271c41a8dd5a2eb2a534
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>
After commit d207738245 we unconditionally
write version two, but it seems useful to allow users to specify the
version explicitly.
Change-Id: I81d3de3d7f87318653f89bf10e3618becd8329d6
Task-number: QTBUG-58769
Reviewed-by: hjk <hjk@qt.io>
Since MySQL now properly supports UTF-8 through utf8mb4 option, it
caused regression on systems which did not use UTF-8 encoding by default
which caused queries to fail as they were converted into system codec.
To fix this, simply use UTF-8 encoding for queries when MySQL supports
it.
Task-number: QTBUG-59176
Change-Id: I21cc9102b15df15a31bc7c74469321c44a257946
Reviewed-by: Andy Shaw <andy.shaw@qt.io>
Unconditional cast to a QMouseEvent while the event might also be a
QKeyEvent.
Change-Id: If5eb6fbad6e4440c167ff95298f51efde1834217
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>
Replace all QT_NO_PROCESS with QT_CONFIG(process), define it in
qconfig-bootstrapped.h, add QT_REQUIRE_CONFIG(process) to the qprocess
headers, exclude the sources from compilation when switched off, guard
header inclusions in places where compilation without QProcess seems
supported, drop some unused includes, and fix some tests that were
apparently designed to work with QT_NO_PROCESS but failed to.
Change-Id: Ieceea2504dea6fdf43b81c7c6b65c547b01b9714
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@qt.io>
the library has a dependency on libdl.
Task-number: QTBUG-58301
Change-Id: I36567ded32980b241ff2f01cfdec044510405a75
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
instead of having a library and a test, use a library with two sources,
the first being empty (i.e., just libc). this allows us doing away with
the "libdl" feature, and using just the "dlopen" one.
subsequently, replace all LIBS+=$$QMAKE_LIBS_DYNLOAD with
QMAKE_USE+=libdl.
the definitions of QMAKE_LIBS_DYNLOAD remain in the qmakespecs for
backwards compat only. n.b.: the only specs where it is not empty or
"-ldl" (i.e., what we support now) are the hpux ones, where the library
is called 'dld'.
technically, the "library" feature should depend on '!unix || dlopen', but
that's for a later patch.
Change-Id: Ib8546affc4b7bc757f1a76729573ddd00e152176
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
Reviewed-by: Ulf Hermann <ulf.hermann@qt.io>
This existed in QProcess before forkfd, but was lost in the port to it
(commit 1814142b7a). The original QProcess
fix was done in 97279d0582.
Task-number: QTBUG-57584
Change-Id: Ibc5c715fda334a75bd2efffd14a425871f3162b5
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
This patch fixes 2 simple typos in QGraphicsItem and QPainter
documentation and a copy/paste error between QAbstractItemModel's
beginRemoveColumns and beginRemoveRows documentation.
Change-Id: I32bdc4dc69154a40fe30a5b8c08d0c3a001853f8
Reviewed-by: Harri Porten <porten@froglogic.com>
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>
We do use xcb-xrender, and it has its own detection
logic in configure.json.
Change-Id: I20bbc1ddf5dd0c32e93ef2c12c7b0eda3f96f4f4
Reviewed-by: Alexander Volkov <a.volkov@rusbitech.ru>
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@qt.io>
Unlike setTimeSpec, this forgot to clear the bit when detaching. So it's
possible that some further use of the flags could incorrectly conclude
that the data was short and then proceed to corrupt the pointer.
The example from QTBUG-59061 caused this because toUTC() -> toTimeSpec()
calls setMSecsSinceEpoch which left the bit set; then addDays() calls
setDateTime(), which calls checkValidDateTime() and that corrupted the
pointer. This problem was more visible on 32-bit systems because no
QDateTime was short (except for default constructed ones), but it
can happen on 64-bit with sufficiently large dates.
Task-number: QTBUG-59061
Change-Id: Ibc5c715fda334a75bd2efffd14a562a375a4e69b
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
A QDockWidgetItem will be leaked if a QDockWidget is dragged out of a
floating tab window, and then plugged back somewhere.
The problem is that QMainWindowLayout::unplug was not returning the
QDockWidgetItem* from the floating tab's layout. When that's the case,
a new QDockWidgetItem is created in QDockWidgetPrivate::startDrag
and will be put into the layout, leaking the old QDockWidgetItem.
Change-Id: Ifb9c1c562cb74383ebff1df0f91ee225c5cdb296
Reviewed-by: Sérgio Martins <sergio.martins@kdab.com>
Fix a couple incorrect references to the integer precision classes.
Update snippet to use floating point precision classes.
Task-number: QTBUG-51630
Change-Id: I9b08cfb68937a8e1179ee414d7981956ef7bc106
Reviewed-by: Martin Koller <kollix@aon.at>
Reviewed-by: Michael Brüning <michael.bruning@qt.io>
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>
Reviewed-by: Topi Reiniö <topi.reinio@qt.io>
The call to trimmed() makes sense for URLs typed in a browser's location bar,
but its use in every code path made it impossible to open a file with a trailing
space in command-line tools that uses fromUserInput(cwd) to handle command-line
arguments, as recommended. For instance kde-open5 "file.txt " would fail.
Change-Id: Ie61182684521d91f077d3e76f95b7240965ab405
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
operator=, which was only used for clearing, wasn't clearing the hash.
This led to a mismatch between the vector and the hash (given that the hash
points into the vector).
Spotted by interrupting kmail in gdb, and it was in this code
iterating over a 2000 entries hash (the first vector entries not matching
the hash, this code keep appending new entries for the same formats).
This fixes QTBUG-8862 again, the initial fix having been accidentally
reverted in 467b15a.
Change-Id: Ia34b3d002a0199e1930431a4bbdb2ec981ed4ffc
Task-number: QTBUG-8862
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
The QSystemLocale constructor remembers the most-recently-constructed
instance - a dodgy enough proposition at the best of times - and
shares it with much of the rest of QLocale. There is a global static
instance, actually of a derived singleton class, to which it is
usually set on program start-up. However, the constructor deleted the
remembered instance before remembering any new instances; there was no
way this could not lead to bad consequences. So let's not do that.
Change-Id: Ie8f3d655c9d4f75f6ec00a5861d98d6020ecc633
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
In 0ae4b948515af904dba397448839056d9b7965af, we added a work-around
specifically for the Android emulator, causing it to override the
precision qualifiers in shader code, since the emulator would send
the shader code directly to the host OpenGL driver and thus there
was no guarantee that the qualifiers would be supported.
Since then, the original issue with the emulator has been fixed,
so now Qt applications are failing because of the work-around.
[ChangeLog][Android] Removed old work-around which was causing
OpenGL shader compilation to fail on updated Android emulators.
Task-number: QTBUG-44697
Change-Id: I0da879b3a8fbe7cb2d0969cdf45664d0b3499891
Reviewed-by: Laszlo Agocs <laszlo.agocs@qt.io>
Reviewed-by: BogDan Vatra <bogdan@kdab.com>
A window can resize while dragging, this happens on Windows when dragging
across screens, to a screen with a bigger scale factor. When that occurs
it might lead to the press pos being outside of the window.
Change-Id: Ic61ec7088c8fa81395d43ce665952dbd2eecba39
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
RunLoopModeTracker as one of Qt Cocoa classes was not wrapped
in namespace which limited its use in Objective-C single symbol space.
Change-Id: Ida2c62c6f543a3bf5107f28c78d27435bcb3470d
Reviewed-by: Jake Petroules <jake.petroules@qt.io>
This is still not a great place for it, but this way we can
Q_REQUIRE_CONFIG(library) in qlibrary_p.h and qfactoryloader_p.h is the
only private header that doesn't require library support in plugin.pri.
Change-Id: Ia6c1ac8799822eca1e0cbeca98d1c1b8de74beae
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@qt.io>
Make it iterate the addresses available looking for an IPv4 address,
when that's what it needs, instead of just assuming the first entry in
the list (when non-empty) is IPv4.
Based on a suggestion by Dmitry Pankratov.
Task-number: QTBUG-27641
Change-Id: I1920f68ade44a996ea5c2ed691a87ff3e686f35a
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Pass classname 'exefile' to ShellExecuteEx() for UAC prompt.
This allows running executable files that do not have a .exe extension.
Task-number: QTBUG-59008
Change-Id: I88d669481e893db50edccd7b30259e5366477556
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@qt.io>
Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
now that it works nicely, make use of it for the zlib deps, which are
not transitive when the detected library is built statically.
Change-Id: Iaed87a37b36f714f0b919244cd84809650102ba9
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
the outdated ones remain for backwards compatibility; some remain
unchanged.
Task-number: QTBUG-30083
Change-Id: Ia596b854d26b00fcb4f48df0da7ad893650ac1c8
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
also actually deletes qfeatures.txt, which was already claimed by
a668c6a6, but not actually done.
Task-number: QTBUG-58411
Change-Id: I686760632fee7c10b01bd2e83f2481b01bc2b774
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
don't attempt to hand-craft a library export any more. instead, use the
configure system's built-in mechanism, and refer to it via QMAKE_USE.
this also allows us to rely on transitive dependencies in the autotest.
as a side effect, this makes the openssl-linked feature imply the
openssl one.
Change-Id: I5dd209b63bc8fbbc62852f6ffc472d4452ea2e68
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
The this pointer cannot be null, so we can't do d->staticFunction while
d is a null pointer. This was caught by Clang 3.8's ubsan.
Change-Id: I3c0d39b88cca83d827a69ed1544a4412b18ac907
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
The problem was caused by commit 0feeb6f6d2
which fixed QTBUG-58036. It reparented widget that used to be skiped.
In particular, floating widgets are skiped. But seting the parent of a
floating widget docks it. And so as a result it would not be skiped anymore.
This has two side effect: This breaks the animation (as the widget is
docked too early; and cause crash when QDockWidgetGroupWindow get
reparented as this breaks invariant that these are always floating.
So restore the skip from before commit 0feeb6f6d2,
and explicitly set the parent in all cases when the animation finishes.
Change-Id: I0e3e29ad22d6ffe4d62242d48a18dadb916fc14f
Reviewed-by: Sérgio Martins <sergio.martins@kdab.com>
The 'rules' vector is made up of all the individual {env,config,...}Rules
vectors under mutex protection whenever init() is called (only from the
QCoreApplication ctor) or, at any time, by a call to QLoggingCategory::
setFilterRules().
Yet, the writes to the individual *Rules vectors were never protected by
registryMutex, racing against the reads of the same vectors in the
updateRules() function.
Fix by protecting all access of all member variables with registryMutex.
Add some strategic comments to make analysis easier for the next guy.
Change-Id: If68d15a553ec7038693574a34f10a39f4cd480e8
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
In this case we can safely return the same QSize which we accept since
it is already contains the right size for the given menu and this size
will be bounded to screen geometry before QMenu will be displayed
anyway. We also get rid of one dependency on HITheme.
Change-Id: I7502a96d180fc4a41ce3dfabe8a200b886016348
Reviewed-by: Gabriel de Dietrich <gabriel.dedietrich@qt.io>
AİİA.pdf takes 8 QChars, but after lowercasing it takes 10, so the code cannot
assume the length to be the same.
Task-number: QTBUG-58822
Change-Id: Id6fbb99f6afd08ee420099cd66372732d7598d9e
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
An Acquire barrier in QReadWriteLocker::unlock was missing to synchronize
with the testAndSetOrdered on d_ptr in the lock functions.
The race is between the write of d->writerCount in tryLockForWrite,
and the read in unlock. The acquire on d->mutex is not enough because
it is not on the same object. While that race could be fixed by taking
the newly-allocate()ed d's mutex before publishing it with testAndSet,
there's another race, on 'recursive', between a newly-minted Private*
with recursive == false in tryLockForWrite(), and the read of 'recursive'
in unlock().
Task-number: QTBUG-58917
Change-Id: I10ba36573c0e57468d11e9b77d85045711feaea1
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>
see 5a1b4832a2 for more detail
Task-number: QTBUG-58142
Change-Id: I51851ea9b4fe7b8eeadc452bc3dbb1ea00026d29
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
The _next variable need the acquire and release fence in next()
to synchronize with the equivalent operations in release() (which
already have the them)
The ordering on the _v[block] is not enough as this does not
synchronize the same object.
Task-number: QTBUG-58917
Change-Id: I17cc39e6791433348b6227363dbea92bcf03700d
Reviewed-by: David Faure <david.faure@kdab.com>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Like the lvalue QVector::append() overload, when we reallocate,
we need to take a copy of the function's argument because the
reference will get stale upon reallocation.
Add a test.
[ChangeLog][QtCore][QVarLengthArray] Fixed a bug involving
appending an item already in the container to the container
again.
Change-Id: I06eeed6cb383dd5924e47a302bb3d1666d04c8e8
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
In QCocoaDrag::dragPixmap, it treats QDrag::source as a QWindow, but it
is not - it's just a generic QObject* of some kind (which QQuickDrag
sets to the originating QQuickItem, and the widgets stack sets to a
QWidget). This failure means that dpr stayed at 1.0.
Unfortunately it’s not possible to receive a pointer on QWindow directly
from QQuickItem because QtWidgets and QtQuick do not share the sources,
but we can use the same dpr as current focused window has because drag
can only start from active window - press on a window which is not
focused should activate it first.
Task-number: QTBUG-57942
Change-Id: Id358c181d03d519188caaa83fb4226033b8ed1ea
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
Reviewed-by: Robin Burchell <robin.burchell@crimson.no>
Automount filesystems like /home on many operating systems (QNX and
OpenIndiana, at least) don't like if you try to mkdir in them, even if
the file path already exists. OpenIndiana even gives you an ENOSYS
error.
So instead, let's try to mkdir our target, if we fail because of ENOENT,
we try to create the parent, then try again.
Task-number: QTBUG-58390
Change-Id: Ibe5b1b60c6ea47e19612fffd149cce81589b0acd
Reviewed-by: James McDonnell <jmcdonnell@blackberry.com>
Reviewed-by: David Faure <david.faure@kdab.com>
QNetworkSession's ctor has a race condition: it 1) connects signals/slots
and also 2) registers meta-types (but after these connects). Our users apparently
have a lot of per-thread QNAMs in multiple threads (and implicitly
many QNetworkSessions on different threads too). From error logs it appears
that while one thread tries to connect signals/slots and evaluates the types
of signal and slot arguments, another thread can register this type.
If the first thread extracted signal argument types, then the second
registered this type, we can end up in a 'type mismatch' error on the first
thread with seemingly the same types in a debug message (something like
"type mismatch A::Some <-> A::Some") - they have the same name, but one
has type() == 0 and another - some non-zero type().
Now we call qRegisterMetaType before connect.
Task-number: QTBUG-50901
Change-Id: Idbb9515573e174bbc5e5bf55dc3a7911a81646ea
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
The library feature already covers this. As library depends on the
dlopen compile check, we can assume dlopen to be available, also on
vxworks.
Change-Id: Idcdb07ab4688c6158651d9a5ad5e2ba126bd7d9e
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@qt.io>
It is never written after initialization, and accessed concurrently
outside mutex protection, so make sure that it never changes by
marking it as const.
Change-Id: Ib28b2e8b3067d596a95d1699165a6ebfb15fe674
Reviewed-by: David Faure <david.faure@kdab.com>
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
Adding 0 each time will obviously not produce a new identifier each
time...
Also use static initialization for QBasicAtomicInt.
A default-constructed static QBasicAtomicInt at function scope
will be dynamically initialized. It will still be zero-initialized,
but at least GCC adds guard variables for such objects.
When using aggregate initialization, the guard disappears.
Amends 265db5ad9b.
Change-Id: Ia71290cf26c486dcbcc74381f12cd0c4712d6019
Reviewed-by: David Faure <david.faure@kdab.com>
Use of XInput 2 may result in regressions. For example we call
XISelectEvents for each window every time a device plugs/unplugs.
It causes significant delays when there are many native windows.
Change-Id: I60b799bb667d0e4bca1f9c52cdaa07b04bcc749f
Task-number: QTBUG-57013
Reviewed-by: Gatis Paeglis <gatis.paeglis@qt.io>
QDeadlineTimer will eventually support CLOCK_MONOTONIC_COARSE on Linux,
so let's be safe before that change appears.
Change-Id: I9b423f1b71194205b27afffd14a383a7e9f66a9d
Reviewed-by: Stéphane Fabry, Cutesoft <stephane.fabry@gmail.com>
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
It was unnecessary, since we only cached the static types, which are all
generated by the macro anyway. The way it was implemented, this produced
data races that are strictly-speaking UB, even if all the threads were
writing the same values to the same data locations.
This commit changes a little the code to simplify, since we're changing
those lines anyway.
Task-number: QTBUG-58851
Change-Id: Idc5061f7145940f987dffffd14a30047846e3113
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
When the QFileDialog receives a language change event, and native
dialogs are used, this code path dereference a nullptr in the qFileDialogUi member.
Protect by checking the return value of QFileDialogPrivate::usingWidgets() as
done in other places deferencing qFileDialogUi.
Fixes a crash in GammaRay
Change-Id: I3e9eb81c75d9fc983519a58c5ec06e316ac80263
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>
If the QSemaphore::tryAcquire() call times out, we mustn't
touch *res, because there was no happens-before relation
established between *res = result in the lambda and our
returning *res;
Fix by returning a default-constructed hash in that case.
Add a strategic std::move().
The same problem exists in runOnAndroidThreadSync(), but
I have no idea how to solve it, because there the shared
object is the runnable itself.
Change-Id: I9a2c431144c169fbd545763555d96153143a11bf
Reviewed-by: BogDan Vatra <bogdan@kdab.com>
Eskil Abrahamsen Blomfeldt
Michael Jabbour
Marc Mutz
Thiago Macieira
Daiwei Li
Peter Seiderer
Konstantin Ritt
Olivier Goffart
Gatis Paeglis
Dan Cape
Samuli Piippo
Jake Petroules
Kevin Funk
Edward Welbourne
Andy Shaw
Timur Pocheptsov
Nikita Krupenko
Oswald Buddenhagen
Ulf Hermann
Oleg Yadrov
Stig Bjørlykke
Jason Erb
Jesus Fernandez
Nico Vertriest
Simon Hausmann
Joni Poikelin
Christian Gagneraud
Aaron Linville
David Faure
Sergio Martins
Pavol Markovic
Stephan Reiter
Shawn Rutledge
Eric Lemanissier
Florian Bruhin
Alexander Volkov
Dmitry Shachnev
Samuel Gaist
Frederik Schwarzer