| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
| |
The texture parameters are set in drawTexture anyways.
Reviewed-by: Gunnar Sletta
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* Use ExpandToTransparentBorderPadMode since we can use GL_CLAMP_TO_EDGE
to clamp to the texture.
* Shrink the bounding rects reported by the blur
and drop shadow filters (expanding by 2 * radius isn't needed).
* Use a single-pass blur for radii <= 3 to avoid the overhead of
rendering to an FBO.
* Made the fast blur setting generate filters for only a predefined set
of radii, and then use the actual blur radius to spread the sample
points outwards.
* Optimized the generated program to rely less on temporary variables,
as those seemed to not be handled very well by certain GLSL compilers.
Reviewed-by: Gunnar Sletta
|
|
|
|
|
|
|
|
|
|
| |
After typing a sentence in Japanese, and before committing the
sentence, you can select each component and choose a different
character representation for it. This commit fixes highlighting of the
currently selected sentence component which was broken by commit
55137901.
Reviewed-by: Marius Storm-Olsen
|
|
|
|
|
|
|
|
|
|
|
|
| |
The error message come from the QGLContextPrivate::bindTexture.
But since OpenGl errors are retains until glGetError is called the
actual error was happening somewhere else.
After adding in all the gl* call a check for a gl error, I was able
to get the place where opengl fail with a GL_INVALID_ENUM.
This happen in the call of glEnable(GL_TEXTURE_2D) in the file
qgl_x11egl.cpp. This glEnable call does not need: removed.
Reviewed-by: Tom Cooksey
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
initTexture() has explicit handling of invalid images, but when
calling adjustSpanMethods() the invalid case is not handled
for Type == Texture.
This caused two types of crashes:
* call to 0 address, because sourceFetch[] has 0 pointer
for QImage::Format_Invalid
(see https://bugs.kde.org/show_bug.cgi?id=176014)
* division by zero in tiled blend functions, because of the
" % image_size" modulo arithmetic.
(see https://bugs.kde.org/show_bug.cgi?id=203231)
Merge-request: 1213
Reviewed-by: Samuel Rødal <sroedal@trolltech.com>
|
|\ |
|
| |
| |
| |
| |
| |
| |
| | |
It is used in qwindowsurface_x11gl.cpp. The problem only showed when
building in release mode due to previous use of Q_AUTOTEST_EXPORT.
Reviewed-by: Tom Cooksey <thomas.cooksey@nokia.com>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Modern graphics drivers tend to disable shared memory pixmaps,
since rendering operations often can't be accelerated by the
GPU when the pixmap data is pinned in system RAM.
Use XShmPutImage() to flush the window surface when this is case,
instead of falling back to the slower XPutImage() method.
Merge-request: 1684
Reviewed-by: Samuel Rødal <sroedal@trolltech.com>
|
|\ \
| |/
|/| |
|
| |
| |
| |
| | |
Reviewed-by: Michael Brasser
|
| |
| |
| |
| | |
Reviewed-by: Trust Me
|
| |
| |
| |
| |
| |
| | |
Maemo5 looks very weird without these patches.
Reviewed-By: jbache
|
| |\ |
|
| | |\ |
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
The created keyed actions are passed to wrong widget if this pointer
is passed insted of itemView. Also create softkey actions in
setItemView method instead of constructor in order that custom items
views are also working correctly.
Task-number:: QTBUG-4702
AutoTest: All QComboBox tests passed
Reviewed-by: Jason Barron
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
This was achieved by putting back the old trap handler after the S60
framework construction has finished.
Task: QTBUG-4960
AutoTest: Included and passed
RevBy: Shane Kearns
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
If there is no focusWidget at the time of the event delivery, we must
ensure that we don't dereference a null pointer.
RevBy: Jason Barron
RevBy: Liang Qi
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
QS60Style causes a compilation warning due to using incorrect parameter
when adjusting rect size in adjusted() call. The problematic line is
using twice same parameter and thus leaves one pre-set parameter unused,
which causes a compilation warning. Fixed by using the correct
parameter.
Task-number: N/A
Reviewed-by: Trust Me
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
Looks like the originals may have had some extra exports from udeb in
them
Reviewed-by: TrustMe
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
Note: No autotest exports, no EGL exports from QtGui
Reviewed-by: Jason Barron
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
This reverts commit f4abf627a8d097e095022d2709718a681b54bd7e.
DEF file was unconditionally enabled for Webkit, ignoring setting in
qtbase.pri, which was supposed to be the global place to enable/disable
DEF file usage. Remove this workaround since we still haven't got
DEF files switched on by default.
|
| | | |
| | | |
| | | |
| | | | |
Reviewed-by: TrustMe
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
The idea is that softkeys and pop-ups need to take care of setting all
softkeys by themselves.
Task-number: QTBUG-4916
Reviewed-by: Jason Barron
|
| | |\ \ |
|
| | | | |
| | | | |
| | | | |
| | | | | |
Two new APIs in QDesktopWidget
|
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
Problem caused by freezing with wrong openC version in environment
Reviewed-by: TrustMe
|
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
One private export has been changed from non-const to const pointer
parameter
Reviewed-by: TrustMe
|
| | |\ \ \ |
|
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
Reviewed-by: TrustMe
|
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
QtGui had some debug functions only exported in the debug build.
Now these are exported in release mode as well, but as stubs (i.e. no
debug output is generated).
Reviewed-by: Thiago Macieira
|
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
Since raising toplevel widget nowdays brings the whole app to top,
logically lowering toplevel widget should put the app to background.
Reviewed-by: axis
|
|\ \ \ \ \ \
| |/ / / / / |
|
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
Task: 262677
Reviewed-by: joao
|
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
After closing a window we used to go through the list of
kDocumentWindowClass windows to pick the next one to pop
to front. This patch will search the kMoveableWindowClass
list of windows first, and as such, pop to front any
modal window first
Rev-By: MortenS
|
| | |_|_|/
| |/| | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
At the moment, Qt, in many places, does not really understand that
a mouse wheel, or touch pad, might operate on a much higher
granularity than 15 degrees (that is, a delta of 120). This is clear
disadvantage on mac, since the mighty mouse, and track pad, got a
resolution that is close to 1 degree. This is called pixel scrolling.
This patch first and formost changes the implementation of
QAbstractSlider::wheelEvent to _really_ understand what to do when
delta is less than 120. Rather than accumulate delta until 120
is reached, then scroll with a value equal to:
offset * step * QApplication::wheelScrollLines (default = 3), we
multiply offset directly, before waiting for 120. This means that
event tough offset is below 120, multiplying it with wheelScrollLines
and step will very often give a value over 120, menaing we can scroll
much earlier and _much more_ fined grained. This also fixes some
auto tests that was ifdeffed out because of specialised mac code
written inside this function from before.
(NB: we still plan to introduce a new event for pixel scrolling,
perhaps for Qt-4.7)
Rev-By: Andreas
Rev-By: denis
|
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
Task-number: QT-2322
Reviewed-by: Sarah Smith
|
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
This fixes an "off by 1" bug in screen compositing with OpenVG
that left lines all over the background when windows were moved.
Task-number: QT-2322
Reviewed-by: Sarah Smith
|
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
reserve() affects capacity(), not length().
Task-number: QTBUG-551
|
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
This reverts commit f4abf627a8d097e095022d2709718a681b54bd7e.
DEF file was unconditionally enabled for Webkit, ignoring setting in
qtbase.pri, which was supposed to be the global place to enable/disable
DEF file usage. Remove this workaround since we still haven't got
DEF files switched on by default.
(cherry picked from commit 3b7f570e6f296ef0a5c9c581ed06cb19986164a0)
|
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
Reviewed-by: Trust Me
To-be-completed-by: QtWebKit developers
|
| | | | |
| | | | |
| | | | |
| | | | | |
Reviewed-by: Trust Me
|
| | | | |
| | | | |
| | | | |
| | | | | |
Reviewed-by: Trust Me
|
| | | | |
| | | | |
| | | | |
| | | | | |
Reviewed-by: Thiago
|
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
Task-number: 262636
Reviewed-by: gunnar
|
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
These functions are implemented directly in assembly, so they need the
proper directives to enable/disable visibility. On ELF systems, it's
.hidden, whereas on Mach-O systems (Mac) it's .private_extern. On
Windows, it's not necessary since you have to explicitly export. I
also implemented the AIX idiom, though it's unlikely anyone will
implement AIX/POWER JIT.
That leaves only HP-UX on PA-RISC unimplemented, from the platforms
that Qt supports. It's also unlikely that we'll imlpement JIT for it.
Reviewed-by: Kent Hansen
(this commit was 26d0990c66068bfc92a2ec77512b26d4a0c11b02, but was
lost during a WebKit update)
|
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
It was missing the ".text" directive at the top of the file,
indicating that code would follow. Without it, the assembler created
"NOTYPE" symbols, which would result in linker errors.
|
| | | | |
| | | | |
| | | | |
| | | | | |
Reviewed-by: Samuel
|
|\ \ \ \ \ |
|
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
There's no reason to lock ourselves to int in the API when some of the
backend could handle floating point blur radii.
Reviewed-by: Bjørn Erik Nilsen
|
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
Usable for future optimizations.
Reviewed-by: Samuel
|