Add function returning the screen for a new geometry for geometry
change events. This ensures that the checking is done in platform
screen coordinates (which might differ from QScreen coordinates
due to high-DPI changes) and also that no screen changes are
emitted for child windows.
Change-Id: I406750f59f006f834c386d09c0c85a804014924a
Reviewed-by: Jørgen Lind <jorgen.lind@digia.com>
Reviewed-by: Laszlo Agocs <laszlo.agocs@digia.com>
This only ocurred in -opengl dynamic builds on Windows, when using
ANGLE. The !QT_OPENGL_ES_2 code path did not correctly check for ES.
We also need to avoid using REPEAT, just to be safe on ES2 implementations
(embedded devices typically) that will not do repeat on non-power-of-two textures.
Task-number: QTBUG-39638
Change-Id: I2eeab6627c17255373193c0a0fec8923f711e28c
Reviewed-by: Andrew Knight <andrew.knight@digia.com>
Reviewed-by: Jørgen Lind <jorgen.lind@digia.com>
Currently QOpenGLTextureCache::bindTexture always convert any uploaded
image to RGBA8888 before uploading. This is quite inefficient when
OpenGL natively supports uploading formats in the original format.
This patch adds support for uploading a few native QImage formats. This
also get the performance of QOpenGLTextureCache::bindTexture on par with
QGLContext::bindTexture.
The texture brush used by QOpenGLPaintEngine is also converted to QImage,
since bindTexture will convert it to QImage anyway, and going over QPixmap
may cause an unnecessary conversion.
[ChangeLog][QtGui][QOpenGLTextureCache] Support uploading common QImage
formats directly to OpenGL when supported.
Change-Id: I828a763126441a98e4547c32ef52dddf7c129a32
Reviewed-by: Gunnar Sletta <gunnar.sletta@jollamobile.com>
Adds basic support for 10-bit per color channel formats to QImage
and the XCB plugin. This will make it possible to paint to and from
these formats, but only at 8-bit per color channel accuracy.
This also fixes Qt5 applications on X11 with native 30bit depth.
[ChangeLog][QtGui][QImage] Added support for 10-bit per color channel image formats.
Task-number: QTBUG-25998
Change-Id: I93ccd3c74bfbb0bd94b352476e5fe58a94119e1f
Reviewed-by: Gunnar Sletta <gunnar.sletta@jollamobile.com>
url is set in qtbase/doc/global which is inherited by the
Qt 5 module qdocconf files.
Change-Id: Ieffa174f598f4a3b8ce8be9bfae7ca9b6981f12b
Reviewed-by: Martin Smith <martin.smith@digia.com>
This patch makes it possible to use QOpenGLFramebufferObject::toImage()
together with QOpenGLPaintDevice::setPaintFlipped where the FBO is
already mirrored.
The patch also makes checks for proper BGRA support before trying to use
it, and fixes the rare case of OpenGLES on big endian.
[ChangeLog][QtGui][QOpenGLFramebufferObject] Introduce an argument
to QOpenGLFramebufferObject::toImage() to save mirroring the result.
Change-Id: I163d802736b7059411f7dda96a31385d772823cc
Reviewed-by: Laszlo Agocs <laszlo.agocs@digia.com>
This is useful without QWidget too.
Change-Id: Ic7857e52e4a8f57c1205615a1f30323b486814af
Reviewed-by: Richard J. Moore <rich@kde.org>
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@digia.com>
Reviewed-by: Alberto Mardegan <mardy@users.sourceforge.net>
Reviewed-by: Shawn Rutledge <shawn.rutledge@digia.com>
Don't detach QImage, when it's in Format_Mono or Format_ARG32.
Use QImage::constScanLine() instead of QImage::scanLine().
Change-Id: I30fcafb576aea3189637a40fd75f77c70017ba46
Reviewed-by: John Layt <jlayt@kde.org>
Reviewed-by: Gunnar Sletta <gunnar.sletta@jollamobile.com>
The integrated assembler of clang does not understand some/all of the
ARM macro assembler syntax used in pixman-arm-neon-asm.S. By default,
this integrated assembler is used when using the "clang" command as a
driver. This patch turns off the integrated assembler of clang for that
file.
Change-Id: Ic06801266b5a4b097ca835d815bcc5d5fc672946
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@digia.com>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
There are cases where @2x images are in use but
where you don't want the associated behavior:
- You are implementing an image editor
- You are inheriting a resource set which has @2x
images, but don't want that behavior with Qt.
Add support for disabling @2x behavior with QT_HIGHDPI_DISABLE_2X_IMAGE_LOADING.
Change-Id: I206e050b6241f8935b4404c7a28b024a9e00d46c
Task-id: QTBUG-38485
Reviewed-by: Jake Petroules <jake.petroules@petroules.com>
Reviewed-by: Morten Johan Sørvig <morten.sorvig@digia.com>
So far, we had to listen to the QObject::destroyed signal from the QScreen
class to figure out whether a screen was removed. Often, this is already
too late, given that most of the QWindows have been moved by then and we
don't get to react before the windows are being set to the primary screen.
This patch introduces a new signal that will notify about a screen removal
before the screen is started to be destroyed, so that the application gets
to decide what to do with the screens before Qt decides to move things around.
[ChangeLog][QtGui][QGuiApplication] Added QGuiApplication::screenRemoved
signal to inform that a screen has been removed, before Qt reacts to it.
Change-Id: I99304179f4d345cae581a87baac6cff7b8773dea
Reviewed-by: Shawn Rutledge <shawn.rutledge@digia.com>
There is no guarantee that the OES extension is present and the standard
functions are not required to be dynamically resolvable on ES.
By performing the ARB-style lookup for the suffixless function names we
can also support ES3 compatible contexts on desktop GL.
This also fixes the problem of picking up the APPLE extension functions
instead of ARB when both are available. vaoFuncsType was set to ARB correctly
but the helper has to take the preference of ARB into account too.
Task-number: QTBUG-38168
Change-Id: If7402320e8f96891017674f3c43bc57c4b5d29f3
Reviewed-by: Sean Harmer <sean.harmer@kdab.com>
This allows QQuickSystemPalette to listen to palette changes without
installing an expensive event filter on the application object.
Change-Id: I8b693e047d993c444e393d7a714a5709692c3560
Reviewed-by: Shawn Rutledge <shawn.rutledge@digia.com>
Reviewed-by: Robin Burchell <robin+qt@viroteck.net>
1a4ff6f122 modified the features test
by checking extensions as well as the GL version. The problem was that
the GL version was wrong in the first place -- buffer textures are
natively supported since OpenGL 3.0, not 4.3.
4.3 introduced support for buffer texture ranges, i.e.
ARB_texture_buffer_range; however it's pointless to take the highest
requirement, especially considering that so far QOpenGLTexture doesn't
wrap glTexBuffer(Range) in any way. In the future, if QOpenGLTexture
will also wrap glTexBuffer, then we will also be able to introduce
a different feature flag for testing buffer texture ranges.
Change-Id: I6becbd49ac26d44ce01d088cbb7831b5cc127bdb
Reviewed-by: Sean Harmer <sean.harmer@kdab.com>
Reviewed-by: Paul Lemire <paul.lemire@kdab.com>
Conflicts:
mkspecs/qnx-x86-qcc/qplatformdefs.h
src/corelib/global/qglobal.h
src/network/socket/qnativesocketengine_winrt.cpp
src/plugins/platforms/android/androidjniaccessibility.cpp
src/plugins/platforms/windows/qwindowswindow.cpp
Manually adjusted:
mkspecs/qnx-armle-v7-qcc/qplatformdefs.h
to include 9ce697f2d5
Thanks goes to Sergio for the qnx mkspecs adjustments.
Change-Id: I53b1fd6bc5bc884e5ee2c2b84975f58171a1cb8e
The problem shows when running tst_qguiapplication.cpp and configuring
with -no-widgets
Change-Id: I9f241760953e543d488096c66a3e886a14f6ae50
Reviewed-by: David Faure <david.faure@kdab.com>
The QDrag objects were apparently leaking on all platforms
other than XCB.
Task-number: QTBUG-39651
Change-Id: I09efcd250c1f42eb385e9d5779be7af6b9b59376
Reviewed-by: Joerg Bornemann <joerg.bornemann@digia.com>
Reviewed-by: Tim Jenssen <tim.jenssen@digia.com>
QOffscreenSurface has to stay usable even after returning from app.exec().
Hence close()ing the underlying hidden window, that is used on platforms that
do not provide real offscreen surfaces, is wrong.
Normally all QWindows are closed (and thus destroy()'ed) when quitting the application,
meaning the the offscreen surface cannot be made current anymore after returning
from exec(). This is an unnecessary limitation and makes certain cleanup operations
impossible.
Task-number: QTBUG-39908
Change-Id: Iea1489378a18f29ff84ba8f13a6dad2d66d2b315
Reviewed-by: Gunnar Sletta <gunnar.sletta@jollamobile.com>
Add detection of MIPS DSPr2 at run-time in qsimd.cpp. This makes it
possible to have generic Qt builds for MIPS that can enable the fast
code paths for processors with the DSP ASE at run-time. Also, this
makes it possible to manually disable them by setting the environment
variable "QT_NO_CPU_FEATURE=dspr2". Last, but not least, functions
requiring DSPr2 are not enabled when running in CPUs with version-1
DSP.
Change-Id: Ia5a01d84119553c22ab83386c74a6cb8ba5fee53
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Qtbase contains four identical implementations of next power of two,
these should be shared and the implementation made available to other
qt modules, as it is also used many places outside of qtbase.
[ChangeLog][QtCore][QtMath] Introduced qNextPowerOfTwo methods.
Change-Id: Id23fbe5ad6bae647b30d5a4212c0330e48a50278
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
It is useful to be able to detect synthesized mouse events
in GraphicsView as well.
[ChangeLog][QtWidgets][QGraphicsSceneMouseEvent] Accessors
for Qt::MouseEventSource and Qt::MouseEventFlags were added to
QGraphicsSceneMouseEvent to enable detection of
synthesized mouse events.
Task-number: QTBUG-39814
Change-Id: Ib5835fef1f484005f9b0fc86518ed32ea79cd80f
Reviewed-by: Shawn Rutledge <shawn.rutledge@digia.com>
Reviewed-by: Laszlo Agocs <laszlo.agocs@digia.com>
Call the standard functions directly in GLES 3.0+ builds.
The catch here, just like with the mapBuffer changes, is that we could,
in theory, dynamically load a GLES3 implementation on the !QT_OPENGL_ES_3
path too. However this is limited to Windows currently and we don't have
a full GLES3 stack there (yet), and even when we do get it, the ANGLE
extensions for blit and multisampling will still work. Therefore this
isn't really an issue for now.
Task-number: QTBUG-38168
Task-number: QTBUG-39187
Change-Id: I343a737218c9fe438ee1603b37e93f0400d952a5
Reviewed-by: Andrew Knight <andrew.knight@digia.com>
Reviewed-by: Gunnar Sletta <gunnar.sletta@jollamobile.com>
The support already in place for ANGLE is now extended for NV.
On ES 2.0 the only way to get multisampled renderbuffers and blitframebuffer
is through vendor-specific extensions. QOpenGLFunctions is updated to resolve
the related functions for both ANGLE and NV, in addition to EXT.
Task-number: QTBUG-39187
Change-Id: I1aab805ced3d06dde3dc547221bbf833ff8e06c2
Reviewed-by: Andrew Knight <andrew.knight@digia.com>
Reviewed-by: Gunnar Sletta <gunnar.sletta@jollamobile.com>
qt_alphamapblit_rgba8888 is only used on little-endian systems, where
qAlpha() returns the correct value.
qdrawhelper.cpp:6256:13: error: 'void qt_alphamapblit_rgba8888(QRasterBuffer*, int, int, quint32, const uchar*, int, int, int, const QClipData*)' defined but not used [-Werror=unused-function]
Change-Id: Ibba6dd6914138f7ae5d53a8e354597f5fff65433
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@digia.com>
The config test correctly recognizes if GLES 3.0 is
available, however qopengl.h still includes the ES2
headers. This causes issues for the new GLES3 support
patches.
Change-Id: Ia97f556cc207f7d828918f493fe1adab93cf31ec
Reviewed-by: Gunnar Sletta <gunnar.sletta@jollamobile.com>
This allows QQuickApplication to listen to layout direction changes
without installing an expensive event filter on the application object.
Change-Id: I2d7d8906acecbc092657c4bd918bbdc9aad9744c
Reviewed-by: Shawn Rutledge <shawn.rutledge@digia.com>
This suppresses the warning
QIODevice::seek: Cannot call seek on a sequential device
Task-number: QTBUG-39217
Change-Id: Ie7b0845c760ae6fc857d02bf9ec5c5adb24fb631
Reviewed-by: aavit <eirik.aavitsland@digia.com>
After cb8445f032, Qt no longer
considers the baseline to have a vertical size, so it does
not add an extra pixel to the font height. The documentation
needs to be updated to reflect this.
Task-number: QTBUG-39668
Change-Id: I28fc813e21d73bb03f7055b0f0843511a12d308b
Reviewed-by: Konstantin Ritt <ritt.ks@gmail.com>
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@digia.com>
Refactor the code QPixmapIconEngine::addFile() using a convenience
class for reading all images.
Special-case .ico-files: Read images into a list and replace by
higher-quality ones.
Task-number: QTBUG-39287
Change-Id: I32ab6c77a276dc5d4d9a8f7b216c81149b8772b8
Reviewed-by: aavit <eirik.aavitsland@digia.com>
The state should already be updated when fillPath() is called, so it
should use the painter's pen to fill the path with instead as this will
have been updated already.
Task-number: QTBUG-39303
Change-Id: I1cc9922d4183bd44076c26210db06ad825ebf25b
Reviewed-by: Gunnar Sletta <gunnar.sletta@jollamobile.com>
Increase and decrease actions can be generally applied to any value
interface. We therefore make them available regardless of the
existence of any action interface.
Change-Id: I82ba01965dc869439b9d741ce681e0c0687263ca
Reviewed-by: Frederik Gladhorn <frederik.gladhorn@digia.com>
QImage::fill(uint) was incorrectly performing ARGB->RGBA conversion when
called on RGBA8888 formated images.
This patch moves the color conversion to QImage::fill(QColor) where it
belongs so that fill(uint) can behave consistent with documentation and
how it treats other formats.
The fill(uint) method had no automated tests, and this patch adds one.
[ChangeLog][QtGui][QImage] QImage::fill(uint) now fills the given pixel
value unconverted when used on RGBA8888 image, making it consistent with
the documentation and treatment of all other image formats.
Change-Id: I00a9d810c61d350dbdd7c4b9ad09e5ce11896b6d
Reviewed-by: Gunnar Sletta <gunnar.sletta@jollamobile.com>
QOpenGLShaderProgram::hasOpenGLShaderPrograms tests whether
QOpenGLFunctions::Shaders is provided for the given context. As the
initialization code assumed OpenGL 2 this always was true. But
unfortunately we still cannot assume that OpenGL 2 is universally
supported. E.g. indirect rendering (no matter how bad that idea is)
does not support OpenGL 2 on all hardware and the Shader related
extensions are not available.
This change makes sure that only when OpenGL 2 is available the
features provided by OpenGL 2 are enabled. If OpenGL 2 is not
available the extensions are tested. The checks are slightly
reordered to not do the extension tests at all if OpenGL 2 is
available.
Task-number: QTBUG-39730
Change-Id: Ic775163e0dcc519925b1287f3c4ca5e8ebf101d5
Reviewed-by: Laszlo Agocs <laszlo.agocs@digia.com>
Change the semantics of QWindowPrivate::screen to contain the screen of
top level window only. Child windows always return the screen of their
toplevel window by recursing up.
The QPA plugins then no longer need to report screen changes for child
windows.
Change setScreen() accordingly, bail out for child windows, and
emit screenChanged() recursively.
Also add a check to setParent() preventing screen changes.
Task-number: QTBUG-36659
Change-Id: I19c8e12217cba1513e947a027f2492abc7b98816
Reviewed-by: Shawn Rutledge <shawn.rutledge@digia.com>
Reviewed-by: Laszlo Agocs <laszlo.agocs@digia.com>
After 717d39ac08, if HB_Face was instantiated by QFontEngineFT A
and then used by QFontEngineFT B, whilst A already destroyed,
a crash occurs in hb_getSFntTable() due to accessing a stale pointer.
Task-number: QTBUG-39278
Change-Id: I3428669a311f49cdda1725b778f45219cbcf470d
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@digia.com>
Passing in a sample count of -1 should be treated as 0. This is common
when setting up framebuffer formats from a QSurfaceFormat where the
default, unset value is indicated by a value of -1.
This broke QQuickWidget which was unaware of this limitation of
QOpenGLFramebufferObject and was passing format.samples() as the sample
count without making sure it is 0 or higher.
Task-number: QTBUG-39699
Change-Id: I324b8b006eaa992c15ae932f9df305500fefeb65
Reviewed-by: Paul Olav Tvete <paul.tvete@digia.com>
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@digia.com>
Make sure qRegisterGuiGetInterpolator is called even for static builds.
Task-number: QTBUG-37341
Change-Id: I65735a558d5bbfaa02fa4ec47d55ddf33ca1991a
Reviewed-by: Jędrzej Nowacki <jedrzej.nowacki@digia.com>
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@digia.com>
Added some checks to QRasterPlatformPixmap::createPixmapForImage() to avoid
crashes when QImage::convertToFormat() returns a null image.
Change-Id: I573505a1aff7931d9a2fb452d0a83ae93d8de7db
Reviewed-by: Gunnar Sletta <gunnar.sletta@jollamobile.com>
Reviewed-by: Morten Johan Sørvig <morten.sorvig@digia.com>
In some cases, for instance if drawEllipse() was called with a null QRect
then it would still try to draw a QVectorPath without having enough points
in the list. Therefore the point_count should be checked first before
doing any drawing.
Change-Id: I9b8dbb87c73b74e9df9eb10ec790a484d05d4c46
Reviewed-by: Gunnar Sletta <gunnar.sletta@jollamobile.com>
This lays the foundation for iOS accessibility.
The approach is slightly different from other a11y bridges in
that we completely flaten the hierarchy of wigets/quick items to
a list. This works well with VoiceOver since there are comparatively
few elements. The cache implementation for OS X is re-used.
With this patch VoiceOver on iOS works on many applications out of the box.
For now it sends the screen changed notfification somewhat overzealous,
that will need revisiting and potentially new API in QAccessible.
Device orientation changes are not yet supported.
[ChangeLog][iOS] Accessibility was added to the iOS platform port.
This enables Qt applications to be read by VoiceOver on iOS devices.
Task-number: QTBUG-39097
Change-Id: I441e844652d528cc2fdcc444f43b54ed6fa04f0c
Reviewed-by: Richard Moe Gustavsen <richard.gustavsen@digia.com>
A vectorized codepath for rotating transforms were added in 5.3.0, but
was only properly tested for SSE2. The NEON version remains unfinished.
Since it was never working, this patch reverts the NEON version.
Task-number: QTBUG-39445
Change-Id: Ifbce0e03781d217ad976c6b18ac88381055cba66
Reviewed-by: Gunnar Sletta <gunnar.sletta@jollamobile.com>
Prevents a shutdown crash in the Qt Mfc migration solution.
Task-number: QTBUG-39317
Change-Id: I7f0aa40715baed57ff59682a477084347dfd45bc
Reviewed-by: Andy Shaw <andy.shaw@digia.com>
Some drivers are reported to get confused when passing different
sample counts (requested vs. actual) to the color and depth/stencil
attachments. To overcome this, pass the requested sample count to all
the attachments.
Task-number: QTBUG-33406
Change-Id: I17b0e3dbbd78de2ab0f45e95164b4f326d47aeff
Reviewed-by: Paul Olav Tvete <paul.tvete@digia.com>
Reviewed-by: Kimmo Leppälä <kimmo.leppala@digia.com>
Now that the current fbo is not tracked anymore, toImage() and
blitFramebuffer() will not restore the previously bound (or
whichever Qt thinks was current before) fbo. toImage() needs to
accommodate this since the isBound() check was going wrong now that
the blit always makes the fbo non-current.
The current fbo is now queried via glGet both in toImage() and
blitFramebuffer().
This will fix the recently introduced Lancelot failures with -glbuffer.
Change-Id: I43a780beaeac4697d92cb0ebda7d14ca28a3924f
Reviewed-by: aavit <eirik.aavitsland@digia.com>
Reviewed-by: Gunnar Sletta <gunnar.sletta@jollamobile.com>
When binding an FBO directly via glBindFramebuffer, the QOpenGLContext's
internal current_fbo, that is maintained by QOpenGLFramebufferObject,
becomes out of sync. This will lead to QOpenGLFramebufferObjects thinking
they are still bound.
Such state tracking should be avoided since it is becoming increasingly
difficult to keep it consistent between the various OpenGL API wrappers
and will never be robust enough when the application changes the state by
directly calling OpenGL functions.
current_fbo is now removed in QtGui. QtOpenGL is not touched.
Change-Id: Id809aab1306c9486d1e2ba3bb5aa93593659e920
Reviewed-by: Gunnar Sletta <gunnar.sletta@jollamobile.com>
The selectable/selected states refer to items in a list and similar,
do not interpret them as text selection states.
Without this change NVDA for example announces text edits as selected
which makes no sense and which it doesn't do for native text items.
Change-Id: Ib1d109523bd4cc2b9b40ace8a8c3d7d3a7f9f25c
Reviewed-by: Jan Arve Sæther <jan-arve.saether@digia.com>
Call clear() in the destructor of QOpenGLTextureGlyphCache class to
prevent QOpenGLGlyphTexture object leak.
Change-Id: I290b09b0786d30603391e6855a21e9232b112739
Reviewed-by: Gunnar Sletta <gunnar.sletta@jollamobile.com>
QRasterPaintEngine::alphaPenBlt makes the assumption that a device with
bit depth 32 has optimized alphamapBlit and alphaRGBBlit routines.
This will fail on RGBA8888 format resulting in some text not being
rendered.
Change-Id: Ia7d88bb0e3094894affceda1acc42396b67b3677
Reviewed-by: Gunnar Sletta <gunnar.sletta@jollamobile.com>
As ANGLE doesn't support GL_RGBA as the internal format for its
multisampled framebuffer extension, use GL_RGB8_OES to ensure a working
out-of-the-box experience.
Task-number: QTBUG-39283
Change-Id: Icb364225e74e5d3480a8617131a56e9f65f04ba5
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@digia.com>
Reviewed-by: Andrew Knight <andrew.knight@digia.com>
When allocating (mutable) storage for a texture, OpenGL mandates that
the client-side format and pixel type "match" the internal format.
That needs to happen even if we are not actually uploading anything
(because we're passing NULL as the data parameter and no PBO is bound).
This means that we need to pick a compatible format/type and not just
pass GL_RGBA / GL_UNSIGNED_INT. In turn, it implies adding new
enum values to the various format/type enums.
Change-Id: If40c63b1d44764b3be131dd1b41d13983a19ae45
Reviewed-by: Sean Harmer <sean.harmer@kdab.com>
Reviewed-by: Paul Lemire <paul.lemire@kdab.com>
Don't use only the GL version, as vendors expose many many extensions
on viable hardware.
For instance, I have a NVIDIA G210 which supports up to GL3.3, but
which features immutable storage, immutable multisampled storage,
texture buffers and ranges, stencil texturing, and cubemap arrays.
Change-Id: Ie6023ee854b679737fca982578cb2093e10d083f
Reviewed-by: Paul Lemire <paul.lemire@kdab.com>
Reviewed-by: Sean Harmer <sean.harmer@kdab.com>
To extract the target of a given texture object. Somehow this
accessor was missing.
Change-Id: Ie43366bed3627a20204600e68e426b55abf37af6
Reviewed-by: Paul Lemire <paul.lemire@kdab.com>
Reviewed-by: Sean Harmer <sean.harmer@kdab.com>
Reviewed-by: Laszlo Agocs <laszlo.agocs@digia.com>
Factor out common code paths from the QPixmap/QImage conversion
code and use the same algorithm for copying the image data with
alpha correction as does qt_pixmapFromWinHBITMAP().
Rename the previous version of qt_imageFromWinHBITMAP() to
qt_imageFromWinIconHBITMAP() since it is used for HICON conversion.
Task-number: QTBUG-39084
Change-Id: Ia4042c33db485c3604461a5eafd6282968b36e3b
Reviewed-by: Andy Shaw <andy.shaw@digia.com>
...and check the returned value where it may cause undefined behavior
(i.e. negative amount of items or iteration from -1 to n).
Change-Id: Ib7bd9ab178526df45b792ad48b91ebbab6be861a
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
Doing makeCurrent() followed by destroying the surface (and no
doneCurrent) results in still having the pointer stored internally
in the QOpenGLContext. If then a function like defaultFramebufferObject()
is called, the pointer is dereferenced.
To fix this, the doneCurrent() has to be called when the surface is destroyed
before the context without doneCurrent(). This is pretty much what the
user would expect anyhow.
Task-number: QTBUG-38994
Change-Id: Ibd4083d9291c7fd39b38ce81a988a8e0c9d55d60
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@digia.com>
Reviewed-by: Jørgen Lind <jorgen.lind@digia.com>
The dynamic builds (-opengl dynamic) are now functional on Windows.
In such a build no components in Qt link to any OpenGL libraries directly
and qmake will not automatically add any such libraries to the
applications' makefiles. Instead, the libraries are chosen and loaded
during runtime and applications are expected to use QOpenGLFunctions
instead of direct OpenGLfunction calls.
Set the environment variable QT_OPENGL to desktop or angle to skip testing
and force the given implementation. The application attributes (AA_UseOpenGLES
and such) are also taken into account.
The testing logic is same as before: We try to load opengl32 and
resolve a shader related function. If this fails, ANGLE is chosen. This
allows utilizing full desktop OpenGL on systems that have proper drivers,
while a transparent fallback to ANGLE will be done automatically for
systems that don't. The latter includes also remote desktop connections.
Software rendering via Mesa llvmpipe is supported too. The fallback is
automatic on systems where the desktop test fails and ANGLE fails to load
or initialize (e.g. due to missing libs like d3dcompiler), as long as a
suitable patched build of Mesa is available.
[ChangeLog][QtGui] Dynamic OpenGL implementation loading is now supported
on Windows. This requires Qt to be configured with -opengl dynamic.
Task-number: QTBUG-36483
Change-Id: Ie8bb25a6d55b3a1609b00150aeccd909aec27313
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@digia.com>
When these extensions aren't available, use ANGLE's versions of them.
Task-number: QTBUG-31010
Change-Id: I9a85b9f4d2bb60bdb1d79c92edf241b95d0627bf
Reviewed-by: Gunnar Sletta <gunnar.sletta@jollamobile.com>
Reviewed-by: Laszlo Agocs <laszlo.agocs@digia.com>
QOpenGLContext::versionFunctions returns 0, if a QOpenGLFunctions object
for a legacy OpenGL version is requested while using the core profile.
This leads to a crash QOpenGLContextPrivate::maxTextureSize()
Change-Id: I32845643094336cebcc666806a411524fe3e869b
Reviewed-by: Laszlo Agocs <laszlo.agocs@digia.com>
Reviewed-by: Sean Harmer <sean.harmer@kdab.com>
Once LayoutData::items gets cleared, accessing (or re-using)
SpecialData::resolvedFormats may lead to a undefined behavior,
so clear SpecialData::resolvedFormats right after LayoutData::items.
Change-Id: Ib389f0029a0562f5d2837f62f76197510a0db099
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
formats() -> formatCollection()
additionalFormats -> formats
QTextEngine has three different code paths: in context of QTextDocument,
additionalFormats are currently _additional_ formats;
though in QTextLayout, they are the only formats in use;
and the QRawFont-related path shares the QTextLayout's behavior.
This is a preparation step to consolidating these three into a single one.
Change-Id: I427ccc3c2f672ce090899bb0a0995972315daafa
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@digia.com>
For Windows builds, the necessary libraries are not found
in the sdk if the wrong environment is used, which is quite
common.
Task-number: QTBUG-34940
Change-Id: I7d844649790cbfacab3154a717d318fd570c4149
Reviewed-by: Brad King <brad.king@kitware.com>
Reviewed-by: Stephen Kelly <stephen.kelly@kdab.com>
It had no effect because of an explicit check for NoBrush.
However the default in QBitmap is (unfortunately) QBrush(color0), rather
than NoBrush, so the brush must be updated when calling setBrush(NoBrush).
I suppose the real issue is that lastBrush is default-constructed in
QRasterPaintEngine, rather than starting with the brush from QPainter,
which is QBrush(color0) for the case of the bitmap. But no reason to
special case NoBrush here anyway.
Task-Number: QTBUG-38781
Change-Id: I9996ac12bf628920cfaf0de9c886f637a336028b
Reviewed-by: Gunnar Sletta <gunnar.sletta@jollamobile.com>
Don't store unused values (pos_x and levels) and re-use
already calculated ones (itemStart, itemEnd, and itemLength).
Also const-ify some members to make the code a bit more clear.
Change-Id: Ied80ebf9e4e7e8a1d057e413a9bd24f84b8aaf92
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
For webfonts, we can't rely on the fontDef (and hence the cache Key) alone,
since the resulting fontcache key might be strictly identical.
Instead we have to check if the cached engine doesn't belong to a webfont
when we're looking for a "local" font.
Change-Id: I2de11c8fdbef53362b66674d3429a042e1835757
Reviewed-by: Pierre Rossi <pierre.rossi@gmail.com>
This allows to easily create a matrix that performs the transformation
used by OpenGL fixed function to go from normalized device coordinates
to window coordinates.
This comes in useful if you need to perform the NDC->window coordinate
conversion inside a shader.
Change-Id: I183b3545bfb3eb1e8b13fc3172911b46926fcbb7
Reviewed-by: Paul Lemire <paul.lemire@kdab.com>
Reviewed-by: Laszlo Agocs <laszlo.agocs@digia.com>
QFontEngineQPA was really QFontEngineQPF2, and has been renamed. The
multi font engine in qfontengine_qpa.cpp was really a base implementation
of a multi font engine, used by other multi font engines, and has been
renamed and moved accordingly into qfontengine_p.h/cpp.
Change-Id: Iac7409c4dbf0fdc3ee993ce4f7dc96cb00a422e6
Reviewed-by: Konstantin Ritt <ritt.ks@gmail.com>
Reviewed-by: Paul Olav Tvete <paul.tvete@digia.com>
This old workaround is not necessary and could not be used anyhow since
it would affect a way too wide range of drivers. Modern drivers should
be able to honor the default GL_UNPACK_ALIGNMENT of 4 even for 1 byte
per pixel formats like GL_ALPHA. Instead, document why the behavior
is correct.
Change-Id: I1687448ba92875c8ff772ccc371894e88ff64096
Reviewed-by: Gunnar Sletta <gunnar.sletta@jollamobile.com>
Change-Id: I8552199c8b3e365adefbc2bb096c8153e222cec8
Reviewed-by: Konstantin Ritt <ritt.ks@gmail.com>
Reviewed-by: Paul Olav Tvete <paul.tvete@digia.com>
Change-Id: I5c3564aa9c10833957de14a4c28dbd2fb7eb11e4
Reviewed-by: Konstantin Ritt <ritt.ks@gmail.com>
Reviewed-by: Paul Olav Tvete <paul.tvete@digia.com>
Change-Id: I9730247371ccf0e83579b71e13a9f3d8adcf0fff
Reviewed-by: Konstantin Ritt <ritt.ks@gmail.com>
Reviewed-by: Paul Olav Tvete <paul.tvete@digia.com>
The event handler would ignore the close if the window was hidden
while the close() function would not. In addition, the close()
function would force removal of focus etc while the event handler
did not. Make it consistent by calling close() from the event
handler.
Change-Id: Ia05b08299f218620fb40a8cad5d3771158c4701e
Reviewed-by: Morten Johan Sørvig <morten.sorvig@digia.com>
Reviewed-by: Jørgen Lind <jorgen.lind@digia.com>
Due excess 'break' in loop, function xToCursor() with "CursorOnCharacter" option
for BiDI text returns wrong cursor position (start glyph position) all time.
Task-number: QTBUG-38846
Change-Id: Iba6671905e0785da6f343db19d6c3bb3e2cf5e8a
Reviewed-by: Andrey Volykhin <andrey.volykhin@lge.com>
Reviewed-by: Konstantin Ritt <ritt.ks@gmail.com>
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@digia.com>
In particular, if you have a <pre> tag in your HTML, this will
become a QFont with family "Courier New" and fixedPitch==true.
On Android, there's no "Courier New" font, and since the
style hint is AnyStyle, we will just return Roboto, which is
a proportional font.
Note that this exactly matches the condition when fetching the
fallback families for the font in the loadEngine() function,
which was introduced by 06568ff89c48dee8aab278b8b0538c331aa84595
in Qt 4.
[ChangeLog][Text] Respect QFont::fixedPitch() for fallbacks
when font family cannot be matched.
Task-number: QTBUG-36083
Change-Id: I64787c547dc492b9dd3c49f1edf0d9626d198260
Reviewed-by: Jørgen Lind <jorgen.lind@digia.com>
Reviewed-by: Konstantin Ritt <ritt.ks@gmail.com>
The former was documented twice and the latter was missing the qdoc
formatting.
Change-Id: Id8dfb21a0c2fd26134b5738448971fe2627a12d6
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@digia.com>
Reviewed-by: Leena Miettinen <riitta-leena.miettinen@digia.com>
Reviewed-by: David Faure <david.faure@kdab.com>
You should never force that option. Just let Qt automatically detect
from the environment.
Change-Id: I43ae4951969d2067cc111eff6302921a0af82658
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@digia.com>
Reviewed-by: Leena Miettinen <riitta-leena.miettinen@digia.com>
Using Q(Open)GLFunctions concurrently on multiple threads had some issues
due to incorrect synchronization: The m_groups list in the
QOpenGLMultiGroupSharedResource, in which the Q(Open)GLFunctions instance
is stored, became corrupted under certain scenarios, for example in the
tst_qglthreads autotest and any two threads that happen to enter an
initializeOpenGLFunctions() or QOpenGLContext::functions() call concurrently.
Locking in value() has been introduced in 666c25c089
to fix such issues, however using the context group's mutex is not enough: that still
allows two threads using two contexts with a different context group to concurrently
enter insert(). Instead, the MultiGroupSharedResource has to have its own mutex
to protect its own member variables.
Task-number: QTBUG-38771
Change-Id: If01c44c2084b95e487bc9146576ca180ed8044da
Reviewed-by: Jørgen Lind <jorgen.lind@digia.com>
This involves exposing a new function in the QPlatformNativeInterface
which gets a public function for QGuiApplication
Proof of concept is done through implementing _NET_WM_WINDOW_TYPE
setters for xcb
Change-Id: Ic9544e775fb71cc9b30273595ec41b1cdb1c9d64
Reviewed-by: Laszlo Agocs <laszlo.agocs@digia.com>
When the start point is not on the curve, the end point will be
used as starting point for the QPainterPath, if it is on the curve -
if it is not on the curve, the mid point between start and end point
will be used. The origianl start point will still be used as a control
point for the first bezier curve.
Task-number: QTBUG-38601
Change-Id: I90e07f1141288873263f1f12fa51e6a248d0804f
Reviewed-by: Fabian Bumberger <fbumberger@rim.com>
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
GetGlyphOutline fails for ttf fonts with embedded bitmaps. This results in
distancefield rendering not rendering glyphs (for the failed paths). We
need to fall back to texture rendering if this is the case.
Change-Id: Ibdf7dc5c1d34f513c436f88fabbdcc4089bb6fef
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@digia.com>
The image format returned by QOpenGLFramebufferObject::toImage() on
OpenGL ES/2 builds changed in 5.2.0 but was not documented.
Change-Id: I2379c62c7f715803192ffa9bb8ae1171ef4b2314
Reviewed-by: Gunnar Sletta <gunnar.sletta@jollamobile.com>
QGLWidget has a paint device type of Widget, yet drawing on it
from outside the gui thread is fine as long as ThreadedOpenGL is
supported. It has probably been overlooked that the device type
OpenGL applies to QGLFramebufferObject and such but not QGLWidget.
This will fix the problem of flooding the output with warnings in
the tst_qglthreads autotest and potentially elsewhere too.
Task-number: QTBUG-38771
Change-Id: Ie014a2610a0db41b31c30c404e286d4997aecdc3
Reviewed-by: Giuseppe D'Angelo <giuseppe.dangelo@kdab.com>
Reviewed-by: Gunnar Sletta <gunnar.sletta@jollamobile.com>
This patch reuses the code and pattern used to vectorize scaling, to
also vectorize rotating transforms in fetchTransformedBilinearARGB32PM.
This provides significant improvements in QtWebKit on benchmarks using
rotating transforms.
Change-Id: If250e0f5dae1ff0f954301f96cc8970a99ae3e9d
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Gunnar Sletta <gunnar.sletta@jollamobile.com>
When using directfb platform the drawing of some widgets (e.g. Toolbar)
would crash the application.
Pixmap device pixel ratio can be used in QCommonStyle::drawControl() as
a denominator to calculate pixmap width or size. The drawing of some
widgets (e.g. Toolbar) would fail the assertion of a non-zero
denominator.
This commit initializes the device pixel ratio when generating
BlitterClass pixmaps and makes QBlittablePlatformPixmap::metric(
PdmDevicePixelRatio) return devicePixelRatio() instead of 1.
Task-number: QTBUG-38709
Change-Id: Ia8b96c5341b5ac1297c483deed1c695846398b88
Reviewed-by: Gunnar Sletta <gunnar.sletta@jollamobile.com>
Reviewed-by: Laszlo Agocs <laszlo.agocs@digia.com>
The desktop doesn't belong to this application, we shouldn't be setting
it (assuming it is even possible). On X, for example, it's possible to
set the name and you end up with the root window having _NET_WM_NAME
pointing to the last application that created a QDesktopWidget.
Change-Id: Ib62d0a4d56a4d2a74afc6b33fa607867343e7aba
Reviewed-by: Jørgen Lind <jorgen.lind@digia.com>
If the override cursor was set before
QtQuick2ApplicationViewer was shown then the cursor
was the default cursor and not the override cursor.
The new cursor was applied only if the window was
visible and the cursor has been set. Now the cursor
is applied also if override cursor is set.
Change-Id: I88618c719e43b2802e209bef440461c532d19f48
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@digia.com>
Reviewed-by: Laszlo Agocs <laszlo.agocs@digia.com>
gcc doesn't yet warn when a non-POD return datatype is unused,
but clang does, at least.
Change-Id: Ia4c39001994ac09cf20e458bc5a66710bd11780c
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>
Reviewed-by: Gunnar Sletta <gunnar.sletta@jollamobile.com>
[ChangeLog][QtGui][QGuiApplication] Add support for -icon command line
argument on X11, add -qwindowicon on all platforms.
Change-Id: Iacc602466699bf634d8b34aab7ed73c83fd9844f
Reviewed-by: Frederik Gladhorn <frederik.gladhorn@digia.com>
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@digia.com>
As it's meant for QML anyways, this allows it to coexist peacefully with
existing code.
Task-number: QTBUG-29806
Change-Id: Ib04993f47eb2f9f7fc49c4a5400f18f9682a7aaa
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
In the justification code, we unconditionally subtracted one
from the line_length, but then compared the result to 0 afterwards,
so we did not support when the line_length is 0 initially, which
can happen if it only consists of spaces (in which case
trailingSpaces will be non-zero and line_length will be zero.)
The fix is to bail out for both strings of length 1 and length 0.
[ChangeLog][Text] Fixed an assert when justifying a QTextLine
which only contains spaces.
Task-number: QTBUG-38520
Change-Id: Ib04993f47eb2f9f7fc49c4a5400f18f9682a72f2
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
Reviewed-by: Konstantin Ritt <ritt.ks@gmail.com>
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
The plain text edit's smart repaint logic in
QPlainTextDocumentLayout::documentChanged assumes that during a change inside
just one block, the block is in the state before the edit and a call to
layoutBlock() is going to bring it up-to-date. Then only a comparison of the
bounding rect - before and after - is going to allow for smart repaints.
The assumption of the layout being in the same state as before the edit got
broken by commit cc57a2e90f, which introduced
code to use a QTextCursor within a slot connected to QTextDocument's
contentsChange signal. The usage of the QTextCursor there ends up updating the
layout of the block ahead of time, breaking the assumption and therefore the
optimization, in the sense that during changes in the preedit that cause a
change of height / line count, the old bounding rect in
QPlainTextDocumentLayout::documentChanged and the new bounding rect will be the
same. This causes a repaint of only the edited block, missing repaints of the
following blocks, even though the line count effectively changed.
So what's causing QTextCursor to mess with the layout is the attempt of
updating the vertical movement x property. This patch inhibits the update,
marking it as dirty for initialization later. This means that slots connected
to this low-level signal cannot rely on the cursor's visual x position, but
that doesn't seem useful anyway and isn't required for commit
cc57a2e90f.
Task-number: QTBUG-38536
Change-Id: I5fae12d646a4b2d2cc22b9f2d021e5dc8cfdda94
Reviewed-by: Paul Olav Tvete <paul.tvete@digia.com>
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@digia.com>
Reviewed-by: Konstantin Ritt <ritt.ks@gmail.com>
On Windows, DataLocation currently returns the value obtained by
passing CSIDL_COMMON_APPDATA to SHGetSpecialFolderPath(). This is
the local non-roaming path. For actually storing settings, the roaming
path should be used (CSIDL_APPDATA). Introduce new AppDataLocation to
return that path and AppLocalDataLocation for the local path and deprecate
DataLocation.
[ChangeLog][QtCore][QStandardPaths] QStandardPaths now has new
enumeration values AppDataLocation, AppLocalDataLocation to be able
to differentiate between roaming and local paths on the Windows
operating system. DataLocation is deprecated in favor of AppDataLocation.
Task-number: QTBUG-38483
Change-Id: Ib1de8c7031a863ed5eac10c747de6f7ff1a090c7
Reviewed-by: Joerg Bornemann <joerg.bornemann@digia.com>
Reviewed-by: David Faure <david.faure@kdab.com>
isES() becomes isOpenGLES(). The library type enums are changed
DesktopGL -> LibGL and GLES2 -> LibGLES. This removes the now
unnecessary version number, the confusing "desktop" term and provides
better readability.
The old function/values are kept until the related qtdeclarative
changes are integrated.
Task-number: QTBUG-38564
Change-Id: Ibb0a1209985f1ce4bb9451f9b7b093c2b68a6505
Reviewed-by: Sean Harmer <sean.harmer@kdab.com>
Use axis aligned clips when possible instead of layer-clipping. This can
be much faster when a lot of clipping operations take place.
Change-Id: I6865d69fc917a7da858033b4c362b307724d9006
Reviewed-by: Risto Avila <risto.avila@digia.com>
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@digia.com>
Reviewed-by: Andrew Knight <andrew.knight@digia.com>
Broken GIF images could set invalid width and height
values inside the image, leading to Qt creating a null
QImage for it. In that case we need to abort decoding
the image and return an error.
Initial patch by Rich Moore.
Task-number: QTBUG-38367
Change-Id: Id82a4036f478bd6e49c402d6598f57e7e5bb5e1e
Security-advisory: CVE-2014-0190
Reviewed-by: Richard J. Moore <rich@kde.org>
The 10.6 OpenGL headers have "GLenum" as the type
for the 3rd argument. The OpenGL standard has "Glint",
which is also what Qt expects.
Work around this by casting the pointer type.
Task-id: QTBUG-38406
Change-Id: I6d820f41e0d14cbc2d50d91997b6c40b626b159f
Reviewed-by: Laszlo Agocs <laszlo.agocs@digia.com>
they have been fully superseded by 4255ba40ab.
Change-Id: If7ac14c8b7d3cf00fb0cb916036b62eb86c9cee0
Reviewed-by: Joerg Bornemann <joerg.bornemann@digia.com>
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@digia.com>
In ES builds configure will now check for OpenGL ES 3.0 and 3.1
availability. This allows qopengl.h to include the correct header
and, by defining QT_OPENGL_ES_3 and QT_OPENGL_ES_3_1, the OpenGL
wrappers can provide support for MapBuffer, VAOs, etc. on GLES 3.0+
too. Right now this is not possible since the only standard way to
use an ES function specific to a given version is to use the function
directly. The extension mechanism (eglGetProcAddress and friends), that
is often used in desktop GL, is not available for such functions.
[ChangeLog][QtGui] QtGui's OpenGL headers are now automatically including
the highest available header (gl31.h, gl3.h or gl2.h) in OpenGL ES builds.
Task-number: QTBUG-38168
Change-Id: Ib857c58fe1696a9546fdd1aa143a9237e80325a5
Reviewed-by: Jørgen Lind <jorgen.lind@digia.com>
Reviewed-by: Sean Harmer <sean.harmer@kdab.com>
For now only xcb on GLX is supported. Other platforms will follow later.
Add also some missing documentation for the platform OpenGL context factory
functions.
[ChangeLog] QOpenGLContext is now able to adopt existing native contexts.
Task-number: QTBUG-37552
Change-Id: I5dd959f102df178f646b2df5989203b5dc6de376
Reviewed-by: Jørgen Lind <jorgen.lind@digia.com>
Currently a click on e.g. the dock icon is not propagated to the
application so if for example the main widget is hidden, it can't be
brought back. Also neither applicationDidBecomeActive nor
applicationDidResignActive do anything. This patch fixes it
[ChangeLog][QPA][OS X] Add support for ApplicationState capability.
Application can now detect when an application states has changed
as well when the dock icon has been clicked.
Task-number: QTBUG-10899
Change-Id: I53d3e6eed4adc62b343e7aa3e3d8068d3248e7df
Reviewed-by: Jake Petroules <jake.petroules@petroules.com>
Reviewed-by: Gabriel de Dietrich <gabriel.dedietrich@digia.com>
Friedemann Kleint
Frederik Gladhorn
Laszlo Agocs
Allan Sandfeld Jensen
Jerome Pasion
Liang Qi
Michael Zanetti
Martin Pley
Erik Verbruggen
Morten Johan Sørvig
Aleix Pol
Tim Blechmann
J-P Nurmi
Giuseppe D'Angelo
Adrian Perez de Castro
Andrew Knight
Thiago Macieira
Shawn Rutledge
Eskil Abrahamsen Blomfeldt
Andy Shaw
Jan Arve Saether
Martin Gräßlin
Konstantin Ritt
Kai Koehne
Sergio Ahumada
David Faure
Liang Jian
Oswald Buddenhagen
Tasuku Suzuki
Berthold Krevert
Simon Hausmann
Stephen Kelly
Jorgen Lind
Sean Harmer
Tor Arne Vestbø
Gunnar Sletta
Andrey Volykhin
Bernd Weimer
Lucile Quirion
Kati Kankaanpaa
Alan Alpert
Ivan Komissarov
Jani Heikkinen
Louai Al-Khanji
Lars Knoll
Albert Astals Cid
Samuel Gaist