Profiling shows that the cost of QAcccessible::
queryAccessibleInterface is dominated by plugin loading.
(json parsing etc.)
Cache QAccessiblePlugin per class. Also cache the fact
that no plugin is found for a certain class. This speeds
up the average queryAccessibleInterface call by a factor
of 10X
Change-Id: Iab6d052dec499a2203d1dcc4672a8a543b279239
Reviewed-by: Frederik Gladhorn <frederik.gladhorn@digia.com>
For qwineventnotifier.h, just wrap the code around #ifdef Q_OS_WIN. This
has the added benefit of fixing the current qt_no_master_include problem
(that is, even if you #include <QtCore> on Windows, you wouldn't get
it).
For qtypetraits.h, it requires qglobal.h first.
Change-Id: If1ba09a0a29de429a5b87e9878c8ac6a62a443c4
Reviewed-by: Laszlo Papp <lpapp@kde.org>
Reviewed-by: Olivier Goffart <ogoffart@woboq.com>
kernel/qwidgetsfunctions_wince.h:61:34: error: no newline at end of file [-Werror,-Wnewline-eof]
Change-Id: Icbc35227946652db53fc8454d1d42043aa7c15b9
Reviewed-by: Andreas Holzammer <andreas.holzammer@kdab.com>
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@digia.com>
Going through QLocale and QString is not really needed.
This also makes the result of the conversion of negative numbers
in bases other than 10 independent of the architecture and
implements the documented behavior of treating them as
unsigned types.
Change-Id: Ibc231dc5241deb5cbadd9796484a8b5f79c29410
Reviewed-by: Olivier Goffart <ogoffart@woboq.com>
Clang reports:
graphicsview/qgraphicsitem.cpp:1779:13: error: function '_q_qgraphicsItemSetFlag' is not needed and will not be emitted [-Werror,-Wunneeded-internal-declaration]
static void _q_qgraphicsItemSetFlag(QGraphicsItem *item, QGraphicsItem::GraphicsItemFlag flag,
^
According to the public Git history, in Qt 4.5.1 this function was
already unused. The only reason it wasn't caught so far is that it is
recursive: it calls itself. So it is used... by itself.
Change-Id: I6fc6b33cb314b845525dc9315d0ad742e113d5cd
Reviewed-by: Andreas Aardal Hanssen <andreas@hanssen.name>
Reviewed-by: Olivier Goffart <ogoffart@woboq.com>
qsql_psql.cpp:774:12: error: enumeration value 'CancelQuery' not handled in switch [-Werror=switch]
qsql_mysql.cpp:1163:12: error: enumeration value 'CancelQuery' not handled in switch [-Werror=switch]
qsql_sqlite.cpp:527:12: error: enumeration value 'CancelQuery' not handled in switch [-Werror=switch]
qsql_odbc.cpp:88:97: error: format '%d' expects argument of type 'int', but argument 3 has type 'long unsigned int' [-Werror=format]
qsql_odbc.cpp:706:76: error: cast to pointer from integer of different size [-Werror=int-to-pointer-cast]
Change-Id: I79965283057e92a44a0c8375530cfb30107c891c
Reviewed-by: Mark Brand <mabrand@mabrand.nl>
It's deprecated and it's not available on iOS anyway. The recommended
way of getting the number of processors online is via sysctl or
sysconf (both of which are just slightly below).
qthread_unix.cpp:397:13: error: 'MPProcessorsScheduled' is deprecated: first deprecated in Mac OS X 10.7 [-Werror,-Wdeprecated-declarations]
Change-Id: I4bf60985fbde155b78b840f3de3ff0a142b78b19
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@digia.com>
qbenchmarkvalgrind.cpp:229:5: error: variable '_qzz_res' set but not used [-Werror=unused-but-set-variable]
Change-Id: I3c5896659105650d6d824b10ff3beffbdf494e24
Reviewed-by: Jason McDonald <macadder1@gmail.com>
The Linux futex implementation had a Q_ASSERT for positive values, but
the documentation says that negative values should be interpreted as
infinite (equal to lock()).
Test that too.
Change-Id: I2f96a502d672732781e88e49797756ca9a809121
Reviewed-by: David Faure (KDE) <faure@kde.org>
Reviewed-by: Olivier Goffart <ogoffart@woboq.com>
qtconcurrentfunctionwrappers.h:277:22: error: 'QStringList' was not declared in this scope
Change-Id: I54c48386d90146c872679672a1d8cc3675d49c39
Reviewed-by: Olivier Goffart <ogoffart@woboq.com>
Requires C++11 rvalue references and variadic templates so we can
implement perfect forwarding.
Change-Id: I62e47d1ffd0c61e8386f9f246aa79031b7430b46
Reviewed-by: Olivier Goffart <ogoffart@woboq.com>
QUrl::fromUserInput("http://") was invalid, which doesn't make sense
since QUrl("http://") is valid. Same for "smb:" which is actually
even more a valid URL from a user's point of view.
Change-Id: I371ac393d61b49499edf5adbbc2a90b426fe9e5d
Reviewed-by: Marco Martin <mart@kde.org>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Some things needed to be corrected for testing with PSQL, this was
checked against the the PostgreSQL documentation to confirm that the
exepected behaviour is correct.
Change-Id: I45a6b343e9eb920fcae2a62910ecc956abcac0f0
Reviewed-by: Mark Brand <mabrand@mabrand.nl>
GCC was complaining:
qevdevtouch.cpp:475:13: error: 'maxId' may be used uninitialized in this function [-Werror=maybe-uninitialized]
Which got me scratching my head: maxId was unconditionally initialised. How
could GCC be complaining about it being uninitialised? Well, turns out that
bestId could be uninitialised and the code does:
if (bestId > maxId)
maxId = bestId;
Of course, if bestId was uninitialised, the warning should have been in the
"if" line first.
Change-Id: I5e174ab2957d76ad040c14fa6ef8535129b6dce3
Reviewed-by: Laszlo Agocs <lagocs83@gmail.com>
The code was copied from the main parser there, so remove the
duplication.
Change-Id: I85748f6f76b3097ff22958d9de67cfa27061a72b
Reviewed-by: David Faure (KDE) <faure@kde.org>
These cases weren't handled before.
The validateComponent function is copied from QUrlPrivate::parse, with
the added modification that it now needs to check the gen-delims for
the userinfo.
Change-Id: I055167b977199fa86b56a3a7259a7445585129c6
Reviewed-by: David Faure (KDE) <faure@kde.org>
this is much more elegant than the so far propagated !isEmpty(QT.foo.name).
also replace feature-specific tests (no-gui and no-widgets) and the
obsolete contains(QT_CONFIG, foo) syntax.
Change-Id: Ia4b3c8febcabf9eeca67b1f9173a523820b1038b
Reviewed-by: Sergio Ahumada <sergio.ahumada@digia.com>
Reviewed-by: Tasuku Suzuki <stasuku@gmail.com>
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@digia.com>
...and remove the outdated QUnicodeTables::Script enum.
QFontEngineData now has one extra slot that never used
(engines[QChar::Script_Inherited]). engines[QChar::Script_Unknown],
if accessed, would be set with a Box engine instance, and could be used
as a minor optimization some time later.
In order to preserve the existing behavior, we map all scripts up to Latin to Common.
Change-Id: Ide4182a0f8447b4bf25713ecc3fe8097b8fed040
Reviewed-by: Pierre Rossi <pierre.rossi@gmail.com>
Reviewed-by: Konstantin Ritt <ritt.ks@gmail.com>
gcc is apparently tolerant against this abuse. icc is not.
Task-number: QTBUG-28775
Change-Id: I872d32177562f076ca8aecd263bc4bb3c7e8edc0
Reviewed-by: Kai Koehne <kai.koehne@digia.com>
If the font has a CFF table, GDI will not label it as
TMPF_TRUETYPE, however, we can still use GetFontData to get
the SFNT tables. This is required to get the maxp table which
contains the glyph count, which is required to use the font
with the distance-field renderer.
Task-number: QTBUG-28746
Change-Id: I3ca1e3d96ea53c453e6fa422b33d1f1f5050a82c
Reviewed-by: Konstantin Ritt <ritt.ks@gmail.com>
Add a command line program allowing to test rename, copy, etc.
Task-number: QTBUG-28246
Change-Id: Ie9667f03b65a874475700ec9ecd91ca2ed32ed97
Reviewed-by: David Faure (KDE) <faure@kde.org>
Fix MinGW build errors by using the standard main signature.
Change-Id: I0ebe7307a825a7ec50e654f163fbf8fe7060a478
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@digia.com>
Use QWidget::isVisibleTo(parent) instead of isVisible(), just like
QAbstractScrollArea::replaceScrollBar() does. This removes the need
of using QCoreApplication::processEvents() to deliver the actual
hide event just for testing if the scrollbar was requested to be
hidden as it should.
Change-Id: Ie9a816e7b871d280a4b3d9d76adb10601915bd56
Reviewed-by: J-P Nurmi <jpnurmi@digia.com>
Reviewed-by: Caroline Chao <caroline.chao@digia.com>
As recommended by Giuseppe, don't rely on the env var, but use the
internal but exported seed atomic int. This way, the compiler will
detect breakages, rather than runtime.
Change-Id: Iec2bc88c53532d3463d2dc5c73631fc9bc34747b
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Especially so that subclasses remember to emit reset.
Change-Id: Iadcae3fc5fe72584465d4134f385ed0a1d77bfcd
Reviewed-by: Stephen Kelly <stephen.kelly@kdab.com>
QDoc supports \relates in the context of a function that
relates to an existing class or a namespace, but not in the context
of a class. We can use \sa to list the related classes or namespaces
for a class instead.
This change ensures that QDoc generates documentation for QMessageLogger
and QMessageLogContext classes.
Task-number: QTBUG-28468
Change-Id: I2242ab730fe5e3acf54b6fa65774e751d2daa7a4
Reviewed-by: Jerome Pasion <jerome.pasion@digia.com>
Reviewed-by: Martin Smith <martin.smith@digia.com>
Reviewed-by: Kai Koehne <kai.koehne@digia.com>
Explicitly delete animations, and remove the check for stopped state
as that is already done by QAbstractAnimation::stop() and we want to
delete the animation regardless of the current state.
Task-number: QTBUG-28506
Change-Id: I3e34316e5077a8627ff5e6d3babd1873bbbaa774
Reviewed-by: Jens Bache-Wiig <jens.bache-wiig@digia.com>
A pointer to the data of one qimage is used, with an offset, in a copy.
In the mono case, that could lead to overflow: the last row would
of the copy would stretch 1 byte beyond the end of the allocated
area. Fix by reducing the height of the copy, so that it keeps
within the allocated memory.
Task-number: QTBUG-28322
Change-Id: I09abfc83f738f8af000fc50f8c94f63dba3a6cfe
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@digia.com>
... with respect to empty and null strings.
Change-Id: Ic107d5bcc8b659497a567b75a7244caceba5a715
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
