QList<int> wastes 50% space on 64-bit platforms. Use a more
fitting container. Since the storage is only used temporarily,
try to allocate it on the stack with QVarLengthArray.
Also give it better name than just 'list'.
Change-Id: I3dfb1d5927ac36f4b352b5d91ce0c9401b20705e
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@theqtcompany.com>
Reviewed-by: Jan Arve Sæther <jan-arve.saether@theqtcompany.com>
If a QBrush has been created without QPixmap the texture() method will
create one. This patch avoids that in several places by checking the
type of the texture brush before accessing it, or not accessing it at
all.
Task-number: QTBUG-43766
Change-Id: If6009fe1d5bd51b239ae2c838e5c3b904b56b11a
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>
Reviewed-by: Gunnar Sletta <gunnar@sletta.org>
CppCat detected duplicate sub-expressions in the code that checked for
BottomLeftSection and BottomRightSection. It was fairly obvious to
see what the values should be.
Change-Id: Id45ca5bbd26c92b800c60867fef5170578216eee
Reviewed-by: Andreas Aardal Hanssen <andreas@hanssen.name>
There were several leaked CFStringRefs in the new OS X bundle detection
code that is part of QFileSystemEngine.
Change-Id: Id0817e9692da411c7eb8287b9bf71b99ae28f960
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Samuel Gaist <samuel.gaist@edeltech.ch>
Reviewed-by: Jake Petroules <jake.petroules@petroules.com>
According to GetStockObject() function docs at MSDN:
> It is not necessary (but it is not harmful) to delete
> stock objects by calling DeleteObject.
Change-Id: I755dc84c8b86ba2806e97ae41b3025aa3e633ae7
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@theqtcompany.com>
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
This partially reverts commit 1755038134,
which did not only fix undefined (signed left-shift), but also
implementation-defined (signed right-shift) behavior. It turned out
that code depends on a particular implementation behavior (logical
instead of arithmetic right-shift), and needs to be fixed first.
Change-Id: I9ba32d06f127d17d05e0c6f6eac3d26268587bca
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
Reviewed-by: Eike Ziller <eike.ziller@theqtcompany.com>
(this is committing the new output of generate_ui from qttools)
For QLatin1String, operator== is overloaded, even for QStringRef,
so comparing to a latin-1 (C) string literal is efficient,
since strlen() is comparatively fast.
OTOH, QStringLiteral, when not using RVO, litters the code with
QString dtor calls, which are not inline. Worse, absent lambdas,
it even allocates memory.
So, just compare using QLatin1String instead.
Change-Id: I5035d259085c21689ab0f62fd49819ab5223ffe8
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
We may receive viewWillLayoutSubviews calls for the view controller even
if QIOSScreen has released the UIWindow that retains the view controller.
Change-Id: I0cc7c50dbb5ee00224aec46d070b04efe069e85a
Reviewed-by: Richard Moe Gustavsen <richard.gustavsen@theqtcompany.com>
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@theqtcompany.com>
Now that we can rely on screen() returning 0 when the QScreen is not
available we can return early from updateProperties(). We still compute
the member variables as they may be accessed directly for the still
alive QPlatformScreen.
Change-Id: Ia7d0d29a6b50a9c932b565dc53b23d66331c275e
Reviewed-by: Richard Moe Gustavsen <richard.gustavsen@theqtcompany.com>
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@theqtcompany.com>
Change-Id: I6acac4c926fcf5459364133b6cc58baff0519074
Reviewed-by: Richard Moe Gustavsen <richard.gustavsen@theqtcompany.com>
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@theqtcompany.com>
We were only doing this for the size, which caused problems when moving
a window from one screen to another where the two screens had different
device pixel ratios.
Change-Id: If56df34677417369639ee8e4df05820fddd9198d
Reviewed-by: Richard Moe Gustavsen <richard.gustavsen@theqtcompany.com>
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@theqtcompany.com>
Add missing includes and reorder includes to avoid X defines breakage
Change-Id: Iaf95ae2488df3d3301436262ed79f7091b4be0a9
Reviewed-by: Laszlo Agocs <laszlo.agocs@theqtcompany.com>
We don't yet have API compatibility, apparently, so we need to keep up
with those changes.
Dropping support for older versions is not yet acceptable since some distros
(in particular current version of some embedded ones) may ship these versions.
Change-Id: Ibea780abd76c4b89661012dfea46868b432ded42
Reviewed-by: Shawn Rutledge <shawn.rutledge@digia.com>
Applications have no public API to check for the ThreadedOpenGL flag
reported by the platform plugins. Add a static function to QOpenGLContext
to give interested applications a way to decide if they should do
rendering on separate threads.
Task-number: QTBUG-44208
Change-Id: I629332512ab31fe49f842065a911270cd6cea7aa
Reviewed-by: Gunnar Sletta <gunnar@sletta.org>
Use it also in the context initialization code which also relied on pbuffers
until now.
This is safe to do since QOffscreenSurface is deliberately kept minimal: we only
support it for rendering to FBOs or uploading textures, it is quite useless for
any rendering type of operation. Therefore simply not creating a surface when
the extension is available is viable and should not break anything.
While we should have done this a long time ago, the real inspiration now is
to avoid crashes with some bad EGL implementations.
Mesa in particular tends to crash in its interal configuration handling when
trying to create a pbuffer surface with more exotic attributes, for example with
multisampling enabled. Just getting rid of those properties is not an option
since the resulting pbuffer would not be compatible with the existing context.
Therefore our only option is to avoid the pbuffer altogether.
There's a so-far-ignored patch for Mesa http://patchwork.freedesktop.org/patch/38100/
which tries to solve some of this but it turns out it is not sufficient to avoid problems
with the multisampled case. So to avoid more headache, we just stop using pbuffers
there.
[ChangeLog][QtGui] QOffscreenSurface is now relying on EGL_KHR_surfaceless_context when
available, and avoids creating a pbuffer surface when the extension is present.
Change-Id: Id18742768b8e66c8d92ce65a9bf64b0296e14db7
Reviewed-by: Sean Harmer <sean.harmer@kdab.com>
Add support for SSL on iOS/OS X by adding a SecureTransport based
backend.
[ChangeLog][QtNetwork][QSslSocket] A new SSL backend for iOS and OS X,
implemented with Apple's Secure Transport (Security Framework).
Change-Id: I7466db471be2a8a2170f9af9d6ad4c7b6425738b
Reviewed-by: Richard J. Moore <rich@kde.org>
it's debugging code which is used only once (if even uncommented).
Change-Id: Ie57347017dd24f4acecff2a7132f82898dea3122
Reviewed-by: Joerg Bornemann <joerg.bornemann@theqtcompany.com>
more efficient use of string functions.
Change-Id: I3d95d6379eaab025b18449b706f93631a2132aad
Reviewed-by: Joerg Bornemann <joerg.bornemann@theqtcompany.com>
the path is processed, and afterwards fixForOutput()ed again. the first
call makes no sense (even if it registered some variables that are gone
in the second call, that would be pointless exactly because they are gone).
Change-Id: I251f1e4858bec36f3a7a9427c2ba78031b35a2d3
Reviewed-by: Andy Shaw <andy.shaw@digia.com>
Reviewed-by: Joerg Bornemann <joerg.bornemann@theqtcompany.com>
commands come already fully expanded and quoted from the project.
Change-Id: I239d5c305f5f65d32c832bc09bfd1c322051e149
Reviewed-by: Joerg Bornemann <joerg.bornemann@theqtcompany.com>
no other generator does it. if it actually buys anything, it should be
re-instantiated differently.
Change-Id: I8431702ac7d558d65fd28a7f9e36bb49db2eb253
Reviewed-by: Joerg Bornemann <joerg.bornemann@theqtcompany.com>
the logic was such that if the bundle name already had the specified
extension which was not the default extension, the default extension
would be appended, too. i don't think that was the intention ...
now we simply put the default into QMAKE_BUNDLE_EXTENSION if its empty -
the variable is not used anywhere else where it would be expected to
preserve its emptiness, so this is safe.
Change-Id: Ied34d10f9fe60756bddc0037dcb2f1d3bbfd3e12
Reviewed-by: Joerg Bornemann <joerg.bornemann@theqtcompany.com>
there is no reference to it anywhere.
Change-Id: I72403be6c8294d9b2e64075ebd428eba24d97097
Reviewed-by: Joerg Bornemann <joerg.bornemann@theqtcompany.com>
this isn't some fuzzy logic, the call sequence is well determined.
Change-Id: I1696b49ed687da83d2969efcfe23ac6565630020
Reviewed-by: Joerg Bornemann <joerg.bornemann@theqtcompany.com>
it never left the rudimentary stage. should it ever be re-added, it
needs to be done basically from scratch anyway.
Change-Id: I76858c8a2c90235f228f7a6e5a178a10a2669d37
Reviewed-by: Rolland Dudemaine <rolland@ghs.com>
Reviewed-by: Joerg Bornemann <joerg.bornemann@theqtcompany.com>
suppose we queried QT_INSTALL_DOCS/get. if no [EffectivePaths] existed,
we'd try [Paths]. if that didn't exist, either, we'd use the built-in,
which is relative. so we'd query QT_INSTALL_PREFIX ... without the /get,
which is technically wrong.
this doesn't currently matter, as all groups have the same built-in
defaults anyway, but it may (hint hint) matter later on.
Change-Id: I5a3746e80ca1bfbf2f37db3604cc351bddeacd94
Reviewed-by: Joerg Bornemann <joerg.bornemann@theqtcompany.com>
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@theqtcompany.com>
it's not worth making things more complicated to remove the duplication of
one real LOC.
Change-Id: Iddbd0db7d0f81b80192b3980dbe2316b246a3b57
Reviewed-by: Joerg Bornemann <joerg.bornemann@theqtcompany.com>
With auto-vectorization enabled in QtGui, the 32bit version of
qPremultiply is faster than the 64bit version since it can be vectorized
wider (4x on 128bit as opposed to 2x). Since all our important 64bit
targets have SIMD, that makes the 64bit version pointless.
Change-Id: I4e9070a3a3c8e2b54f17a95ba0aee0405cbb8ec9
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>
On a GNOME or UNITY desktop, Qt will currently read font-settings from
Xft and use those ignoring any fontconfig instructions. This patch
changes the behavior so the Xft settings are only used as default, but
any explicit overrides by fontconfig will take precedence.
Task-number: QTBUG-43660
Change-Id: Ie10d5828cbfdd95fe5364c63a625d455d9213936
Reviewed-by: Konstantin Ritt <ritt.ks@gmail.com>
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@theqtcompany.com>
There are currently only two users, but I have patches in
the pipeline which mark some other function noinline, so
proavtively centralize the ifdef'ery involved.
Change-Id: I1f02351fdc903d4e026089e12b8a976ed6a8d603
Reviewed-by: Kai Koehne <kai.koehne@theqtcompany.com>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
The function used a QMap<int,int> to map the %n's to the index
in the args array.
By way of construction, the key was sorted in ascending order
while the values, initially all -1, were later reassigned to be
0...map.size()-1, ie. std::iota().
The only information this data structure stores is therefore
the sorted %n's. For that, a sorted vector is a vastly superior
data structure.
Go one step further and use QVarLengthArray to avoid
allocating any memory for the common case of just a few
placeholders.
As an artifact of the underlying refactoring, the management
of that data structure has been moved to a separate class,
ArgMapper.
Runtime for the following test from tst_qstring:
QString str("%1 %2 %3 %4 %5 %6 %7 %8 %9 foo %10 %11 bar"); // not timed
str = str.arg("one", "2", "3", "4", "5", "6", "7", "8", "9");
str = str.arg("ahoy", "there");
went down from 2.2us to 1.5us.
Change-Id: Ic552615fbac646b78ba05eb4e3215e63d202fd94
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
