This patch optimizes the unpremultiply method further by using a
lookup table to avoid any divisions at all.
The opportunity is taken to export both premultiply and unpremultiply
since they are commonly used methods relevant to the exported QRgb type
that can be both premultiplied and unpremultipled ARGB.
[ChangeLog][QtGui][QColor] Exported highly optimized methods for
premultiply and unpremultiply of QRgb values.
Change-Id: I658bcf57b0bc73c34c1765b64617d43b63ae820b
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Gunnar Sletta <gunnar.sletta@jollamobile.com>
Don't delete the thread object without detaching it, use detach()
instead of CloseHandle(), and avoid a double-delete.
Change-Id: Ia169a96fb32805e06abe099c3c35e97ce485f088
Reviewed-by: Oliver Wolff <oliver.wolff@digia.com>
private api, but useful for 2d gui code that suddenly finds itself
needing to get a set of textures onto some fbo
I didn't want to include ARB_copy_image since it looks like its from
texture/renderbuffer -> texture/renderbuffer while this class implies
texture -> write fbo. We could wrap ARB_copy_image in QOpenGLTexture or
some other class or we can add it later.
I have not added any QOpenGLTexture functions since this class opperates
on the GLuint identifier. We can add overloads later.
Change-Id: I3e565b33466c1c183a249a33c3e82c6786debd55
Reviewed-by: Paul Olav Tvete <paul.tvete@digia.com>
Instead of "blacklisting" Unix platforms that don't have Linux headers,
do the opposite and check for Q_OS_LINUX in the framebuffer code that
has Linux-specific code.
This should help fix the build on other Unices, such as the BSDs.
Change-Id: Icb6edf34bb20c9a98843b6dc0e2a87ee71fcd046
Reviewed-by: Laszlo Agocs <laszlo.agocs@digia.com>
This change does not modify the actual algorithm implemented. It only
updates the source code so that the code generation is more optimal:
- change only one variable per loop (the "offset" variable)
- unroll the tail expansion of the last 15 characters
The Neon code for the toLatin1 codec most likely benefits from the
unrolling of the tail too, but I can't verify that I haven't broken
anything.
Change-Id: I8a92fd3c1aa700e6f8b0c8ebdb1978ade394757f
Reviewed-by: Olivier Goffart <ogoffart@woboq.com>
In case the window manager supports _NET_ACTIVE_WINDOW a client message
is sent to the root window and the window manager is expected to either
activate the window or mark it with demands attention.
In addition the code unconditionally also called xcb_set_input_focus
which breaks the intention of sending the client message. Thus it is
now only done if the WM doesn't support the protocol. This was also the
logic in Qt 4.x in QWidget::activateWindow (qwidget_x11.cpp).
In addition set_input_focus is only called if the window is a toplevel.
Change-Id: Iec4c02f0ea62db30a4b347d474fdfd2f05b8895b
Reviewed-by: Jørgen Lind <jorgen.lind@digia.com>
On platforms like the Raspberry Pi the EGL implementation is not
compatible with X. This means that while EGL and Open GL ES can be enabled,
and will work fine with platform plugins like eglfs, EGL (and thus GL) should
be automatically disabled in xcb, otherwise a compilation error will occur
since xcb assumes the native handle types are the corresponding Xlib types.
Task-number: QTBUG-36551
Change-Id: I2cc4c558abb4b25d422a2c01da9b75b865ace402
Reviewed-by: Jørgen Lind <jorgen.lind@digia.com>
...due to a typo introduced in d8225fab8f and stricter checks
introduced in 4e319ca4c4.
This uses recently added QStringIterator to decrease the code duplication.
Change-Id: Ib1c2519db1c4d65f3ac14e6f9ae3fa272860b1dd
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
Use NSWindow and NSView accessors on QCocoaWindow
instead of QGuiApplication and nativeResourceForWindow.
Change-Id: Iabee1e42dd119726d932f89c66a90810463d70db
Reviewed-by: Gabriel de Dietrich <gabriel.dedietrich@digia.com>
Fixes an issue introduced in b8a38a6737. We forgot
to check for category validness, and printed qCDebug(c, ""...)
unconditionally for compilers not supporting variadic
macros.
Change-Id: I779a97dacd7ea559f0d91473aad2c29e73ea2f89
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Konstantin Ritt <ritt.ks@gmail.com>
This should fix the build on platforms like the BSDs, where linux/fb.h does
not exist.
Change-Id: Ia06403d71c3aaa0e177c85dcefa7986647777ae0
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Add the required printsupport plugins to the QTPLUGIN variable
as is done for the QPA plugin.
[ChangeLog][QtPrintSupport] Made the Qt buildsystem automatically include the
necessary plugins so that static applications can print.
Task-number: QTBUG-29663
Change-Id: I0e2e3b0f25dd5714bd187711c85893926b0c4e85
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Remove findDragOverWindow(); no longer try to find the window under cursor
since that would require synthesizing drag leave events. This is left to the
DnD handling of QWidgetWindow, which searches the child which accepts drops.
Task-number: QTBUG-33635
Change-Id: I844041bb862bb0382147be4e28f880a26843c453
Reviewed-by: Joerg Bornemann <joerg.bornemann@digia.com>
Reviewed-by: Oliver Wolff <oliver.wolff@digia.com>
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@digia.com>
Only top level windows of certain types should be drop sites. Add code to
update drop site registration to setParent().
This ensures the drop site is unregistered for example when winId()
is invoked on a parentless widgets which is later inserted into a layout.
Task-number: QTBUG-33635
Change-Id: Ia3161ecd59753f3f39d2a5967e9cb922cf0b9203
Reviewed-by: Oliver Wolff <oliver.wolff@digia.com>
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@digia.com>
Makes it easier to add code in the future too.
No-op change.
Change-Id: I228c36813ccf8ee95ed4b6cbbc20af3178d1b84a
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@digia.com>
In addition to being more common and consistent with QCommandLineParser, this
will make it possible to add the documentation for these options
in the QCommandLineParser-generated help output.
[ChangeLog][General] Builtin command-line options such as -reverse,
-session, -style etc. now all support double dash, e.g. --reverse,
--session, --style...
Change-Id: Ia2e22c854ccc6a9d7b863b1234317005bc822191
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@digia.com>
Add qt_mac_resolveOption().
Support setting options using either an environment
variable or a QWindow property, with a default fallback
value. The options are resolved with precedence in
the mentioned order.
Adds options for the following:
NSView setWantsBestResolutionOPenGLSurface
NSVIew setWantsLayer
NSOpenGLContext NSOpenGLCPSurfaceOrder
The window properties use the _q_mac_camelCase format.
The environment variables use the QT_MAC_ALL_CAPS format.
Change-Id: I1978a02d62e107a2120d81ffdd0f7b32f7731644
Reviewed-by: Gabriel de Dietrich <gabriel.dedietrich@digia.com>
Disassembly shows that the compiler does not perform the zero-expansion
by itself. It always opts to copy byte-by-byte, which is not very
performant.
Change-Id: I08780902461d9e3e6b7b54298f41d1eca61339c4
Reviewed-by: Olivier Goffart <ogoffart@woboq.com>
QScreen::grabWindow() is not always reliable because it grabs from the
framebuffer. (The window might then be covered by other windows, e.g.
"Stays on top"-Windows, popups etc).
If QScreen::grabWindow() fails we therefore fallback to
QWidget::grab(). This will not grab from the frame buffer, but it will
ask the widget to render itself (with its current state) to a pixmap
and return it.
QWidget::grab() should usually return the expected pixmap, and the
pixmap it gives is not subject to the state of the window manager.
This means that both QScreen::grabWindow() *and* QWidget::grab()
must produce an unexpected pixmap in order for the test to fail.
Change-Id: I276554155bb1e5b510d2a2d43628d91669464fe2
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@digia.com>
Reviewed-by: Frederik Gladhorn <frederik.gladhorn@digia.com>
Do not use deleteLater() to delete the thread.
Task-number: QTBUG-36357
Change-Id: Ie7c87b92a7c73d5fbac01d4951d387ee2facd05c
Reviewed-by: Shawn Rutledge <shawn.rutledge@digia.com>
Reviewed-by: Joerg Bornemann <joerg.bornemann@digia.com>
Follow-up to 91d3298: qt_timezone() expects the number of seconds west
of UTC, whereas tm_gmtoff returns the number of seconds east of UTC, and
contrary to the timezone variable it is not oblivious to DST.
We have to account for those two facts and make sure we return a value
compatible with what timezone would have.
Change-Id: Iacb9077f50d4c847ac09e5a7e952d0e4cd22da1b
Reviewed-by: Gabriel de Dietrich <gabriel.dedietrich@digia.com>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
In the future, we might want to refactor this code to inconditionally
use the common Desktop / ES2 subset (like the rest of Qt does),
and then resolve only the functions actually available at runtime.
For now, state why we're doing that on Windows.
Change-Id: Ic21035bcd88ddc1d9274fd90a146c2824d783b25
Reviewed-by: Gunnar Sletta <gunnar.sletta@jollamobile.com>
Reviewed-by: Sean Harmer <sean.harmer@kdab.com>
OpenGL ES 2 doesn't support 1D textures. So introduce a proper
feature flag and warn if we try to allocate one there.
Change-Id: I73cf58c1f257d2472564f45bff222231e39aca52
Reviewed-by: James Turner <james.turner@kdab.com>
Reviewed-by: Sean Harmer <sean.harmer@kdab.com>
The code first tried to check for the GL_OES_texture_3D extension,
and if present it resolved the pointers to the gl*Tex*3DOES functions.
But then it overwrote those pointers by attempting to resolve the
Desktop GL gl*Tex*3D functions, thus making them unusable.
So, if the extension is found, don't overwrite the pointers.
If the extension is NOT found, keep the general behavior of still
trying to resolve those functions.
(That is going to fail, but refactoring the general behavior
belongs to a separate commit.)
Change-Id: Idaba122cf9500f136e3f79284d3c82284257036d
Reviewed-by: James Turner <james.turner@kdab.com>
Reviewed-by: Sean Harmer <sean.harmer@kdab.com>
The QT_XCB_FORCE_SOFTWARE_OPENGL environment variable is equivalent to
LIBGL_ALWAYS_SOFTWARE for Qt 5 applications only. This is most useful
with drivers that only support OpenGL 1. We need OpenGL 2, but the user
probably doesn't want LIBGL_ALWAYS_SOFTWARE in OpenGL 1 apps.
Together with
http://pkgs.fedoraproject.org/cgit/qt5-qtbase.git/tree/10-qt5-check-opengl2.sh
which goes into /etc/X11/xinit/xinitrc.d, it makes QML 2 just work on
old hardware that supports only OpenGL 1.x in hardware. The scriptlet
checks the glxinfo output for the OpenGL version and sets
QT_XCB_FORCE_SOFTWARE_OPENGL if the major version is < 2. (The scriptlet
requires xorg-x11-xinit and glx-utils.)
Tested on a Radeon 9200 SE (RV280) that supports only OpenGL 1.3 in
hardware.
Change-Id: Ief80d283820d6336052b8f390a0030ba9b687492
Reviewed-by: Jørgen Lind <jorgen.lind@digia.com>
CMAP guarantees there is 1:1 mapping between the Unicode code points
array and the glyph indexes array, which means the QString length
always greater than or equal to a sufficient glyph indexes array length.
Simply add some asserts and improve the memory consumption
and the performance where possible.
Change-Id: I0bc8cdba5f86c1c0ba9e79c8d5f96cbe56ec463e
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
Using the same architecture value in VC Project and manifest
files only makes sense for x64. Instead of doing magic we
just set the correct values inside the mkspecs. VCPROJ_ARCH
is used for Visual Studio, while the manifests use
WINRT_MANIFEST.architecture.
WINRT_MANIFEST.architecture was added to x64 mkspecs for
consistency and phone mkspecs do not use WINRT_MANIFEST.architecture
so it does not have to be set there.
Change-Id: I009473104875b4add8c0530dc6f51177919e997b
Reviewed-by: Andrew Knight <andrew.knight@digia.com>
If VCPROJ_ARCH is not recognized or unset, make "arch" default to x86,
or link won't find the libs.
Change-Id: If2cbda37a80c0fa43e1464775c036cebf10f931a
Reviewed-by: Oliver Wolff <oliver.wolff@digia.com>
Reviewed-by: Joerg Bornemann <joerg.bornemann@digia.com>
The code didn't check for malformed surrogate pairs. That means that
- high surrogates followed by *anything* were decoded as they formed
a valid surrogate pair;
- stray low surrogates were returned as-is.
We can't return surrogate values in UCS-4, so properly detect these
cases and return U+FFFD instead.
[ChangeLog][QtCore][QTextCodec] Encoding a QString in UTF-32 will now
replace malformed UTF-16 subsequences in the string with the Unicode
replacement character (U+FFFD).
Change-Id: I5cd771d6aa21ffeff4dd9d9e5a7961cf692dc457
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Konstantin Ritt <ritt.ks@gmail.com>
Code units 0xD800 .. 0xDFFF are not UCS-4, so we can't happily return them.
Instead, if we encounter a stray surrogate, replace it with 0xFFFD, which
is what Unicode recommends anyhow.
References:
§3.9 Unicode Encoding Forms
D76: Unicode scalar value: Any Unicode code point except high-surrogate
and low surrogate code points.
As a result of this definition, the set of Unicode scalar values consists
of the ranges 0 to D7FF_16 and E000_16 to 10FFFF_16, inclusive.
[...]
UTF-32 encoding form: The Unicode encoding form that assigns each Unicode
scalar value to a single unsigned 32-bit code unit with the same numeric
value as the Unicode scalar value.
§ C.2 Encoding Forms in ISO/IEC 10646
UCS-4. UCS-4 stands for “Universal Character Set coded in 4 octets.” It is
now treated simply as a synonym for UTF-32, and is considered the canonical
form for representation of characters in 10646.
§ 3.9 Unicode Encoding Forms (Best Practices for Using U+FFFD)
and
§ 5.22 Best Practice for U+FFFD Substitution
Whenever an unconvertible offset is reached during conversion of a code
unit sequence:
1. The maximal subpart at that offset should be replaced by a single
U+FFFD.
2. The conversion should proceed at the offset immediately after the
maximal subpart.
[...]
Whenever an unconvertible offset is reached during conversion of a code
unit sequence to Unicode:
1. Find the longest code unit sequence that is the initial subsequence of
some sequence that could be converted. If there is such a sequence, replace
it with a single U+FFFD; otherwise replace a single code unit with a single
U+FFFD.
2. The conversion should proceed at the offset immediately after the
subsequence which has been replaced.
[ChangeLog][QtCore][QString] QString::toUcs4 now does not return invalid
UCS-4 code units belonging to the surrogate range (U+D800 to U+DFFF)
when the QString contains malformed UTF-16 data. Instead, U+FFFD
is returned in place of the malformed subsequence.
Change-Id: I19d7af03e749fea680fd5d9635439bc9d56558a9
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
Reviewed-by: Konstantin Ritt <ritt.ks@gmail.com>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
This helps with a flood of dbus messages due to properties.
Change-Id: I7aa9d36a077d84a88dab561d007d597b0780e096
Reviewed-by: Aaron McCarthy <mccarthy.aaron@gmail.com>
That commit made QString::toXxx (8-bit) functions use C++11 ref
qualifiers, so we need to match it here.
Change-Id: I45b50464d36f858d012b12e0cb511aae347ddb6f
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>
Reviewed-by: Giuseppe D'Angelo <giuseppe.dangelo@kdab.com>
UCS-4 iterator over a QString.
Kept private for now so we can still work on the API.
Done-with: Thiago
Change-Id: I377f8bb1921e591ee3292c08c3e097fb6bc7f0c4
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
QWidgetWindow stores the normal geometry obtained from the widget when
transiting to other states. This does not work reliably on Windows,
where this geometry is already that of the new state. Instead,
introduce QPlatformWindow::normalGeometry(), add implementation
for Windows and use that in QWidgetWindow.
Task-number: QTBUG-21371
Change-Id: I3819ebaf55b4e7d2f7eef1affe6c20712ba45d7c
Reviewed-by: Joerg Bornemann <joerg.bornemann@digia.com>
Because they make sense. I'm even thinking that the char16_t version
should get a QString implicit constructor. Maybe both encodings.
Change-Id: Ifffc61dd890795fbbbd5f7cb5efb3e6287d1270e
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
Every time when a surface is (re)created we must paint something in it
otherwise it will appear black.
Task-number: QTBUG-36594
Change-Id: Ib4baaef189f59a83a251cf89db30b0a3aec16d92
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@digia.com>
UITextView has a property for specifying which UIView the CGRects
you return should be aligned with. This makes a difference for
widgets when not using alien, since then the view that draws the
text will usually not be the same as the view that backs
the top level QWindow.
Change-Id: I240d63c98544c39308cd91465ee84351e7d7d1f4
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@digia.com>
No need to call textWillChange all the time if the text is
really not changing. And report that selectionWillChange
when Qt reports that it has changed.
Change-Id: I7bd9f540cd9302c37888926a6152b803cc871ccb
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@digia.com>
By returning the font used at the cursor position, the
correction pop-up will be resized to match the
point size, and the text marking will get correct
height.
Change-Id: I362579b793794835323bb9ceb5ddb4655526f392
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@digia.com>
For legacy reasons, we send IM events to the focus
object directly instead of through QPA. To be consistent,
and to ensure that IM and key events end up at the same
object in the same order, we need to send key events
directly to the focus object as well.
We should consider fixing up QPA to support IM events
better, but this will do for now.
Change-Id: I8a18a1f7b7295e5c64a109fb98eee928fae06a0f
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@digia.com>
Sending faked key events is not such a good idea, since:
1. We don't get key events on iOS, but text events
2. We cannot determine correct key code or modifiers, nor
do we want to fake modifer press/release etc.
3. Android uses IM for all text input
So it seems that the correct solution is to avoid sending
key events in the first place. This will also bring the iOS
port on par with the Android port.
Change-Id: Ibac1d335184e62eb4185cfd4218a0ec73dffb2c4
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@digia.com>
We don't have a separate enum just for spell checking in Qt, but
Qt::ImhNoPredicitiveText should cover it. So use it
to enable/disable both spell checking and auto completion.
Change-Id: I7ad661cb7d720988f13bc1ed940573006c0ce229
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@digia.com>
This change will add support for input methods, word
completion, spell checking and related functionality.
Change-Id: I41d4de1cab521c679d414cfc7c1a2d0f9c1fcaaf
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@digia.com>