Don't reuse 'w' as the loop variable for every loop in the
(very long) function. Scope the variable closer, or, where
there are no natural scopes, define a new variable.
Required turning some while loops into (more natural) for
loops.
Change-Id: I7d1a52503aeb19cd76be22153ba3d0961bd44cae
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@theqtcompany.com>
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
Later changes will subclass this class. Moving to the header now
for better readable patches.
Change-Id: If17607d69169aa5c449c36c9445308164e387f29
Reviewed-by: Andy Shaw <andy.shaw@theqtcompany.com>
If, after checking a condition, we issue a qFatal()
or a qCritical(), by definition that check is
unlikely to be true.
Tell the compiler so it can move the error handling
code out of the normal code path to increase the
effective icache size.
Moved conditional code around where possible so that
we could always use Q_UNLIKELY, instead of having to
revert to Q_LIKELY here and there.
In some cases, simplified the expressions newly wrapped
in Q_UNLIKELY as a drive-by.
Change-Id: I67537d62b04bc6977d69254690c5ebbdf98bfd6d
Reviewed-by: Konstantin Ritt <ritt.ks@gmail.com>
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
More efficient, because control block and tracked object
are co-located in a single memory allocation.
Change-Id: Id18e2d06db43568eb34c2e2d129d1b116af73acb
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
The only tricky part here is to take the size of the moved-to
member, not the moved-from parameter QString object.
To avoid accidents, give the lvalue version the same structure.
Change-Id: Ic68bb896f1d817c21d913feab43522235c51029b
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@theqtcompany.com>
Reviewed-by: Lars Knoll <lars.knoll@theqtcompany.com>
When streaming non-Q_FLAG QFlags, factor the common code
into a helper function only templated on QFlags<T>::Int
and pass what varies as parameters.
Then overload the helper function for the most common int
type (int) as an out-of-line function, implemented via the
primary template to avoid the two going out of sync.
That leaves the primary helper template only for special
cases (such as future extensions).
Change-Id: I70e0001bcfacab9f8765c9b1075fe80b596223f1
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
What triggered this change was the use of QVector::toList() in
op<<(QDebug, QVector).
Only added support for STL types of which we already include the
headers.
[ChangeLog][QtCore][QDebug] Can now output std::vector, std::list,
std::map, and std::multimap.
Change-Id: I49581e3038daa7626b00169430b72d3d5175eae7
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
Detected with a modified (deep const propagating) shared data pointer.
Most are fixed by using copy constructor and assignment operator
of the shared pointer instead of feeding it .data(). The rest
required adding const locally.
Change-Id: I51e36b9e7040dc2211ef4c5dadfbda73ad22c414
Reviewed-by: Konstantin Ritt <ritt.ks@gmail.com>
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
Leaf result classes do not need to be exposed in the headers. The
implementations were inconsistent on this point.
Change-Id: I5bd41ae9e77b932f6232218a014400a59f2ef5a0
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@theqtcompany.com>
and remove a stray one.
Change-Id: I1a10688e6cf916aa93f383ecc9e9aa8e28966e95
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@theqtcompany.com>
Reviewed-by: Andre Somers <andre@familiesomers.nl>
Reviewed-by: Mark Brand <mabrand@mabrand.nl>
Change-Id: Idfaef408536bafbb31444ec8728dbcf289abac8d
Reviewed-by: Richard Moe Gustavsen <richard.gustavsen@theqtcompany.com>
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@theqtcompany.com>
In some configurations we may end up using Mesa's EGL headers instead
of the Broadcom ones. Make this work by setting the usual define to
prevent including Xlib headers that then conflict with all sorts of
things in QtCore.
Change-Id: I4970553428e5b0e81bd76694980f3b6b194ae4c2
Reviewed-by: Andy Nichols <andy.nichols@theqtcompany.com>
Since AArch64 NEON assembly is different from Arm32 NEON we need to
write a separate version.
Assembly is used over intrinsics as the intrinsics have trouble
efficiently using the vstX and vldX instructions.
Change-Id: I5b67fc87acb2433b503e658099b742d57a9cff18
Reviewed-by: Erik Verbruggen <erik.verbruggen@theqtcompany.com>
The macro __ARM64_ARCH_8 is only set by Apple's flavor of clang. GCC and
mainline clang set __ARM_ARCH to 8, and set __ARM_ARCH_8A (when
applicable).
Change-Id: I356b785ffdbfedf8f1ed682840db431db2779ba5
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
As these are strongly typed, they won't implicitly convert to int, so make sure
to cast explicitly.
(cherry picked from commit 9626baaea9, auto test
is not included)
Task-number: QTBUG-49597
Change-Id: I29c4331a9b0c61f2e60c2bcab5a99f65daa7060f
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
Speeds up application startup by a few percent, even with just a normal
amount of fonts installed.
Change-Id: I4c3d87119ddbc53e66166f21cb72946cdf7e4a41
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
Reviewed-by: Konstantin Ritt <ritt.ks@gmail.com>
QSystemTrayIcon's window should send QEvent::ToolTip to QSystemTrayIcon
main class under X11. This patch fixes regression inroduced in Qt 5.0.
Change-Id: I81f6d85e13f492e5e7d13dacc44185a511e5085d
Task-number: QTBUG-46130
Reviewed-by: Błażej Szczygieł <spaz16@wp.pl>
Reviewed-by: Laszlo Agocs <laszlo.agocs@theqtcompany.com>
The __ARM_NEON is the standard define for NEON instructions support
__ARM_NEON__ is only legacy, and specifically not defined in
AArch64 builds, which causes us not to detect NEON support there.
The NEON assembler files doesn't build with AArch64, so the NEON
drawhelper methods must be excluded for now.
Change-Id: Ie32f855bde94ee7efd8a8ddb7766c931778e729b
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
In a Chrome Remote Desktop session the htotal and/or vtotal timings
can be zero and lead to a SIGFPE exception.
Task-number: QTBUG-49322
Change-Id: Id530335cc760d1938ed888ad095427fcf32c651d
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@theqtcompany.com>
Reviewed-by: Nicolas Capens <nicolas.capens@gmail.com>
Reviewed-by: Gatis Paeglis <gatis.paeglis@theqtcompany.com>
The scaling bit was being set incorrectly in two ways. First, the two
flags SourceOverPixmapCapability and SourceOverScaledPixmapCapability
where being confused. Second, the scaling bit was being set on the wrong
mask.
Change-Id: I4bed44f8bb82f153bc35ad2b580d33b8b6f11f89
Reviewed-by: Laszlo Agocs <laszlo.agocs@theqtcompany.com>
Fix syncqt-warning:
QtWidgets: qtbase/src/widgets/styles/qpixmapstyle_p.h does not have the "We mean it." warning
Change-Id: I66ab3a12e53fe91254d539e236a58ada16bb2cd2
Reviewed-by: Rafael Roquetto <rafael.roquetto@kdab.com>
As 3D touch can be disabled/enabled at runtime on those devices, we need
to watch for changes to the relevant settings and update the touch device
capabilities that we report to the user.
Note that iOS will deliver touchesBegan with a touch force of 0, which
we will reflect/propagate as a 0 pressure, but there is no clear
alternative, as we don't want to wait for a touchedMoved before
sending a touch press event to Qt, just to have a valid pressure.
Change-Id: I47fb8a9f98ab3244e16a337bbfcf1fe24e4c7aa2
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@theqtcompany.com>
We test that inline conversion between same image depths always succeed
inline, but that requires that any direct conversions exists both in
non-inline and inline versions.
This patch adds a non-sse2 inline conversion from ARGB32 to ARGB32PM
which was missing.
Change-Id: I71937cd4b77fb41fe2064da937f6dcbf2a6534e6
Reviewed-by: Gunnar Sletta <gunnar@sletta.org>
As a follow-up to 352c357e6f enabling
support for multiple evdevtablet plugins at runtime (one per device),
we also need to adjust the way QGuiApplication handles the events
received from those plugins, in particular when multiple devices
are sending tablet events concurrently.
Replace the static members in QGuiApplication by a vector storing
the same data per-device, so tablet press/release events can be
recognized independently.
Change-Id: Ie0975cdb03a8f6d05903e2e2e57ceb9de73a74a4
Reviewed-by: Jake Petroules <jake.petroules@theqtcompany.com>
Don't know how to handle 'hint', use qRegisterMetaType to register it.
becomes:
QSignalSpy: Unable to handle parameter 'hint' of type 'QAbstractItemModel::LayoutChangeHint' of method 'layoutChanged', use qRegisterMetaType to register it.
Task-number: QTBUG-49623
Change-Id: I5020bb5b6f4ba87438d0f862279bed1ceb203d12
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
0e99f3c broke tst_qquicktextinput::horizontalAlignment_RightToLeft()
and tst_qquicktextedit::hAlign_RightToLeft(). This fix was proposed
by Konstantin.
Change-Id: I602b7301d415f266224ae2c1ffd81244e9565862
Reviewed-by: Simon Hausmann <simon.hausmann@theqtcompany.com>
Reviewed-by: Lars Knoll <lars.knoll@theqtcompany.com>
On Windows, the wrong value was used to calculate the
design-to-device scale. The assumption has been that tmHeight
in the TEXTMETRIC is the pixel size of the em square, but
it is not, it's the height of the font (ascent + descent).
The pixel size of the font is defined to be the em square size
in pixels.
On OS X, the kerning data was never actually read from the
font. I've added a lazy initialization for this similar to
the one in the FT engine.
This was discovered when investigating QTBUG-48546, as it turned
out that the kerning information extracted by Qt in this case was
different from the one used by Harfbuzz.
I've changed testfont.ttf to kern "_2" so that the digit is positioned
directly on top of the underscore and constructed a test.
[ChangeLog][QRawFont] Fixed kerning on advances in QRawFont for
OS X and Windows.
Change-Id: Ic9a321ad119ea880cef89b861c75a820ab8d3182
Reviewed-by: Konstantin Ritt <ritt.ks@gmail.com>
If, after checking a condition, we issue a qWarning(),
by definition that check is unlikely to be true.
Tell the compiler so it can move the error handling
code out of the normal code path to increase the
effective icache size.
This change contains the changes to the accessible/,
effects/, kernel/, styles/ and itemviews/ subdirs.
Moved conditional code around where possible so that
we could always use Q_UNLIKELY, instead of having to
revert to Q_LIKELY here and there.
In QWidgetPrivate::setWindowModified_helper(), as a
drive-by, I swapped the evaluation order of an
&&-expression (newly wrapped in Q_UNLIKELY) to be
more readable and more efficient (cheaper check
first) at the same time.
In qDraw* (qdrawutil.cpp), simplified boolean
expressions (sometimes by skipping re-checking
conditions already checked in a previous guard clause).
Change-Id: I58be22be0a33522c2629a66c2f6c795771a99f3f
Reviewed-by: Lars Knoll <lars.knoll@theqtcompany.com>
Before, you had to first hide a window, then set the flags, then show
it again as the flags were only applied when showing the window. This
is unintuitive.
Task-number: QTBUG-49628
Change-Id: I240e633ac2581c0ff0e4f35dead1b79e15e15350
Reviewed-by: Laszlo Agocs <laszlo.agocs@theqtcompany.com>
SecTrustSettingsDomain is now (in the latest SDK) an enum, not a typedef
for integer type.
Change-Id: I1ad891190116cb27e10e38167fc15b4ee16a28f4
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@theqtcompany.com>
Things I noticed while reading - so I fixed them.
Change-Id: I48f6f2eef7ac3cf901e2f1305c503fb18f5ab2ae
Reviewed-by: Topi Reiniö <topi.reinio@theqtcompany.com>
This naming pattern is consistent with the other drivers and prepares
for subsequent changes.
Change-Id: I77d1edcfd52727cd1126adbf70a808dfbf3255be
Reviewed-by: Andy Shaw <andy.shaw@theqtcompany.com>
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@theqtcompany.com>
This sample will help user needing to react on QFileOpenEvent get
started faster.
Change-Id: I7def292894fc39d803f70cbf0453f6791b7aea15
Reviewed-by: Topi Reiniö <topi.reinio@theqtcompany.com>
Changes to other widgets in the window, especially special cases like
moving dock widgets around, trigger an unfortunately high number of
paint events and calls to paintGL(). Let's try to avoid this.
There is no need to send out a paint event to a texture-backed widget
when it was not explicitly dirtied. Overlaps won't matter since such
widgets are not part of the backingstore. Everything else has to work
like ordinary widgets, though, it is only the QPaintEvent sending we
can optimize away, nothing else.
Task-number: QTBUG-49466
Change-Id: I8ef294ba0a6c305d0002a80e85c06db2c2501cf8
Reviewed-by: Paul Olav Tvete <paul.tvete@theqtcompany.com>
This will make sure that certs in the domainUser (login),
and domainAdmin (per machine) keychain are being picked up
in systemCaCertificates() in addition to the (usually immutable)
DomainSystem keychain.
Also consider the trust settings on OS X: If a certificate
is either fully trusted or trusted for the purpose of SSL,
it will be accepted.
[ChangeLog][Platform Specific Changes] OS X now accepts trusted
certificates from the login and system keychains.
Task-number: QTBUG-32898
Change-Id: Ia23083d5af74388eeee31ba07239735cbbe64368
Reviewed-by: Markus Goetz (Woboq GmbH) <markus@woboq.com>
This avoids manual symbol lookups and makes the code more readable.
Mark identical code.
Also use smart pointers instead of manual memory management.
Change-Id: I62820313dce87de6623cdc87b6e1361200ed7822
Reviewed-by: Markus Goetz (Woboq GmbH) <markus@woboq.com>
This reverts commit 5b62a5e7aa.
This commit is reverted due to two reasons:
1) It was written incorrectly and does not work as is. The
ifdefs should be ifndefs. In its current state, it does
the exact opposite of what it is supposed to be doing.
2) There is another environment access inside qsimd.cpp
(which checks QT_NO_CPU_FEATURE). This access causes the
app to hang.
All in all that approach is not sustainable as we might get
bitten by environment access again and again. Instead we should
use another environment container or use a recursive mutex for
WinRT and Windows CE.
Task-number: QTBUG-49529
Change-Id: Iaca76404dc1023551a7c25489a609681135765fd
Reviewed-by: Andrew Knight <andrew.knight@intopalo.com>
Reviewed-by: Samuel Nevala <samuel.nevala@intopalo.com>
We assumed that creating a window would always result in a configure
notify event, so we delayed sending the initial expose event until
receiving the configure notify.
That strategy fails in cases where the window does not need a reconfigure
after being created, such as when not running a window manager, or when
creating child windows. In those cases the window is just mapped, and
we ended up never sending an expose event.
The problem was masked by sometimes receiving an explicit expose event
from X, just after the mapped notification, which we then sent without
waiting for configure. Unfortunately we can't rely on this behavior
and need to remove the deferred expose event logic, so that we always
ensure that at least one expose event is sent to Qt.
Change-Id: I702be7f24de2a1e89c085fb6bd95bb8ff7792a27
Reviewed-by: Laszlo Agocs <laszlo.agocs@theqtcompany.com>
Reviewed-by: Uli Schlachter <psychon@znc.in>
Reviewed-by: Frederik Gladhorn <frederik.gladhorn@theqtcompany.com>
Reviewed-by: Jędrzej Nowacki <jedrzej.nowacki@theqtcompany.com>
This is a partial revert of aecf3006bd.
The DPI needs to be set inside the constructor to have the text
rendering initialize properly. Landscape orientation fix is still valid
and the dpi change was unrelated to resolve QTBUG-49470.
Task-number: QTBUG-49610
Task-number: QTBUG-49470
Change-Id: I928b8d291b65cd744731c009917804b96253c276
Reviewed-by: Samuel Nevala <samuel.nevala@intopalo.com>
Reviewed-by: Oliver Wolff <oliver.wolff@theqtcompany.com>
We're never interested in trailing zeroes, unless the number is exactly
0. qdtoa would return an empty string if the result was exactly '0',
which is also fixed by this change.
Change-Id: I3ba2f7e835b92d54d9008ad03fdf6ce5fb3af8a4
Reviewed-by: Lars Knoll <lars.knoll@theqtcompany.com>
When bootstrapping we might rely on sscanf(3) for parsing doubles from
strings. We don't want this to be localized.
Task-number: QTBUG-49616
Change-Id: I1607bb750b479fd7007dd0d875d600be59caa859
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@theqtcompany.com>
Reviewed-by: Lars Knoll <lars.knoll@theqtcompany.com>
Marc Mutz
Mark Brand
Sérgio Martins
Alex Trotsenko
Simon Hausmann
Jake Petroules
Laszlo Agocs
Allan Sandfeld Jensen
Erik Verbruggen
Robin Burchell
Volker Krause
Ilya Kotov
Nicolas Capens
Louai Al-Khanji
Friedemann Kleint
Tor Arne Vestbø
Romain Pokrzywka
J-P Nurmi
Eskil Abrahamsen Blomfeldt
Ulf Hermann
Timur Pocheptsov
Edward Welbourne
Samuel Gaist
Daniel Molkentin
Oliver Wolff
Maurice Kalinowski