Preparation for also using it from QCoreTextFontDatabase.
Change-Id: I9cbef59c21f343a88a5cb3fdbee1ed4791d7a36e
Reviewed-by: Morten Johan Sørvig <morten.sorvig@theqtcompany.com>
Calling it directly is less confusing than calling
it indirectly via the "else" branch in
flushWindowSystemEvents.
Change-Id: I085deff09162137606922a5af7ead23e21497b11
Reviewed-by: Jørgen Lind <jorgen.lind@theqtcompany.com>
Set the accepted bit on the QWindowsystemInterface
keyboard event after processing the key event.
Change-Id: I2d8c9382f14840e464153870dff909000e64ddcd
Reviewed-by: Jørgen Lind <jorgen.lind@theqtcompany.com>
Add en "eventAccepted" field to WindowSystemEvent,
where the event subclasses can record the event
acceptance status. Make handleWindowSystemEvent()
return the accepted status.
This works for synchronous event processing only. If
the event is placed on the QPA event queue then there
is no way to return the accepted state immediately.
In the latter case handleWindowSystemEvent() always
returns "true".
Change-Id: I081aecc54f43588d42d3aaeec7f8458f06937601
Reviewed-by: Jørgen Lind <jorgen.lind@theqtcompany.com>
Convenience accessor for use by the Wayland
platform plugin.
Change-Id: I420209138cfc285f8396913548b9e158a35ee9c1
Reviewed-by: Paul Olav Tvete <paul.tvete@theqtcompany.com>
Propagate the line stride when creating the high-dpi
backing store image.
Change-Id: I15f41965d8eaf1d01ddac0a1a012b71148f757e3
Reviewed-by: Paul Olav Tvete <paul.tvete@theqtcompany.com>
qwidget.cpp:12864:70: error: cast from type ‘const QWidget*’ to type ‘void*’ casts away qualifiers [-Werror=cast-qual]
Change-Id: I7de033f80b0e4431b7f1ffff13f956cb26108af7
Reviewed-by: Martin Smith <martin.smith@digia.com>
qopenglengineshadermanager.cpp(430): warning #177: variable "none" was declared but never referenced
qopenglengineshadermanager.cpp(431): warning #177: variable "br" was declared but never referenced
Change-Id: I7de033f80b0e4431b7f1ffff13f958e4a4cca16e
Reviewed-by: David Faure <david.faure@kdab.com>
Reviewed-by: Laszlo Agocs <laszlo.agocs@theqtcompany.com>
In C++, type punning with a union is not allowed. It will result in
compiler defined behavior at best, and undefined behavior at worst.
Specifically, if QRgba64 is passed to a function by value, the different
members of a struct might be passed in separate registers. This means
that any write to the quint64 might not blank out the values of the
struct whenever the compiler looses track with TBAA.
Change-Id: I991b5492fe4bb13a14bb670fef5bf13dacbe6c0a
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@theqtcompany.com>
This permits text rendering consistent with other FreeType enabled platforms,
like Windows and Linux.
Task-number: QTBUG-42839
Change-Id: I8c99bcaa3fb07c16e935a0c3705af467bc3da584
Reviewed-by: Paul Lemire <paul.lemire@kdab.com>
Just like from RGB32 to RGB30 we must also repremultiply when converting
from RGB64 because the alpha channel loses more precision than the other
color channels.
Since this is not approximated accurately in the simple blending
functions and the functions are no longer needed now the main render
engine supports higher accuracy, the simple blending routines for RGB30
have been removed.
Change-Id: I2b7b8eb015e330a487848fc4370ad3a1e966be91
Reviewed-by: Gunnar Sletta <gunnar@sletta.org>
Avoid reading from the next pixel when the sample is fully based on the
current pixel. This is particular important for the last pixel in an
image when the next pixel might be outside the image buffer.
Change-Id: I3607f9c6c332d11ff944ca35d216d417368f9fd5
Task-number: QTBUG-47228
Reviewed-by: Konstantin Ritt <ritt.ks@gmail.com>
Reviewed-by: Gunnar Sletta <gunnar@sletta.org>
We shouldn't short-cut alpha8 formats in alphaChannel, the method is
used under the assumption that the returned alpha map has encoded the
alphas as an indexed grayscale image. The method is also documented
as not returning alpha8.
Task-number: QTBUG-47138
Change-Id: I1cf16957d12e65d44f2b586d9f127fcb33c549b6
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@theqtcompany.com>
Reviewed-by: Gunnar Sletta <gunnar@sletta.org>
1. glTexImage*D does not accept compressed formats.
2. Replacing it with glCompressedTexImage*D is not an option as per GLES 2.0 spec
since passing null data is not allowed.
3. glCompressedTexSubImage*D must always be preceded by a glCompressedTexImage*d
(or glTexStorage*D) call.
4. Therefore the only way is to do nothing in allocateStorage() and switch to
glCompressedTexImage*D in setCompressedData() whenever mutable storage is in use.
This makes ETC1 textures working on the Beaglebone (and presumably others).
Change-Id: I21a040f6ed4aecaa494b6e5a6c6cd75b7389c15c
Reviewed-by: Sean Harmer <sean.harmer@kdab.com>
- Helper functions in LineBreakHelper have been renamed from
adjustFoo to calculateFoo, to signal that they do work, and
are not adjusting something relative to something else.
- The use of QFixed(1) and >= 0 has been replaced with an explicit
constant for the case of the bearing not being calculated yet.
- A helper function negativeRightBearing() has been added that
returns the absolute value of any negative right bearing, making
the width computations simpler.
- Comments have been added and rewritten to make the logic clearer.
Change-Id: I1d3a0214cfa8b4fed4551f3444b43a37d55bd69b
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@theqtcompany.com>
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@theqtcompany.com>
Even if QHighDpiScaling::isActive() returns false, there
is a small overhead from calling functions and multiplying
with 1.0. This change adds a QT_NO_HIGHDPISCALING #ifdef
that replaces all conversion functions with empty inline
functions.
Task-number: QTBUG-46615
Change-Id: Idf85e1700e3adf7f7174c6fa3f5e1d006d74c826
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@theqtcompany.com>
Adds the drawHyperlink method so that clients can draw a hyperlink
pointing to the specified URL at the specified rectangle. That new
method is to be used by GraphicContext::setURLForRect implementations
that want to render anchors as clickable links in PDF documents.
Task-number: QTBUG-44563
Change-Id: I7b0c602da37ee157d18115c531ab1b11fe304c13
Reviewed-by: Lars Knoll <lars.knoll@theqtcompany.com>
On mobile and embedded ETC1 (via GL_OES_compressed_ETC1_RGB8_texture) is
still the most common (and sometimes the only) option for compressed textures,
at least until GLES3 (with ETC2) becomes widely available.
Note that we treat this format as unsized to prevent going on the immutable storage
path as there is no word on glTexStorage and friends in the extension so that may
just fail as well.
Change-Id: Icbc3c6f62f314d63c4df289ef456f047c5e84cf3
Reviewed-by: Sean Harmer <sean.harmer@kdab.com>
Since d7ca800a87 the shortcut events were
only sent once. The one sent by QGuiApplication did not create a
QKeyEvent with the full native modifier state - basically the extended
key event expected everywhere.
That means that shortcuts on some keyboard layouts - like ctrl-shift-7 on
the German keyboard (resulting in ctrl+/) - would not work when the
shortcut override was tested from QGuiApplication, but then the same shortcut was
sent from QApplication with the full information, so it worked the
second time. Shortcuts of this type in Qt Quick were broken before.
Task-number: QTBUG-47062
Change-Id: I8390b9a96d0d998a2a538ac65503702e0d299cc7
Reviewed-by: Andy Shaw <andy.shaw@theqtcompany.com>
Reviewed-by: Gatis Paeglis <gatis.paeglis@digia.com>
Resolve ambiguity with function calls (due to double->float conversion),
use qreal where needed intead of explicit 'double'.
Change-Id: I28f1fb4a2b424b1399fb082a36ba9cff666f7c62
Task-number: QTBUG-47433
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>
Checking for the exact same context is too strict. The texture is valid
in sharing contexts too.
Task-number: QTBUG-47521
Change-Id: Ifdf329ef5361b48abcb7c69e08acd7e35d624a08
Reviewed-by: Giuseppe D'Angelo <giuseppe.dangelo@kdab.com>
opengl\qtriangulatingstroker.cpp(130): warning C4305: '=': truncation from 'double' to 'float'
opengl\qtriangulatingstroker.cpp(546): warning C4305: '=': truncation from 'double' to 'float'
opengl\qopenglshaderprogram.cpp(1779): warning C4312: 'reinterpret_cast': conversion from 'int' to 'const void *' of greater size
qglshaderprogram.cpp(1709): warning C4312: 'reinterpret_cast': conversion from 'int' to 'const void *' of greater size
Change-Id: If1de629aada308d18c3641b504d74d3294b1e8bf
Reviewed-by: Laszlo Agocs <laszlo.agocs@theqtcompany.com>
the handlers' .pri files added $$PWD to INCLUDEPATH to make the files
self-contained when used externally, but this polluted the include path
of the gui module itself, thus hiding incorrect use of QPA includes.
Task-number: QTBUG-47400
Change-Id: I576469a71e8ded0b409d62687999c0fa884613f9
Reviewed-by: Joerg Bornemann <joerg.bornemann@theqtcompany.com>
Build it only in -developer-build mode for tests that might depend
on exact-matching font behavior.
Return earlier to avoid doing any useless job.
Change-Id: I966ee5689f03403e45f4c957b63e3113f0467803
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@theqtcompany.com>
...mainly for platforms that do not provide a native/unified way
to obtain system-defined font fallbacks list
(ie !CoreText && !FontConfig).
Change-Id: I23c5589d79ddecb6311ccc52ec8b29977f06d408
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@theqtcompany.com>
Use the same trick the WebKit related path has used for years:
if there is no explicit list that takes a precedence over the
default one (i.e. QFont("Arial,SimHei")), then defer the fallback
families list initialization until it gets requested.
Change-Id: If5a74294bdebb24865c619bc0d7328b3e706de76
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@theqtcompany.com>
This allows us to link to it from the Qt licensing information page.
Also, replace the \legalese command with \badcode in order to
keep the license text formatting intact.
Task-number: QTBUG-46478
Change-Id: I1b5100ca0373a533bd56852ab2a8f18e72404334
Reviewed-by: Sami Makkonen <sami.makkonen@digia.com>
Reviewed-by: Pierre Rossi <pierre.rossi@theqtcompany.com>
Add a coordinate scaling layer to QtGui, which supports 'devicePixelRatio'
type high-dpi on all platforms, in production and also for
development and testing purposes.
High-DPI scaling is opt-in, by setting environment variables:
QT_SCALE_FACTOR - sets a global scale factor
QT_AUTO_SCREEN_SCALE_FACTOR - sets per-screen scale factors,
where the scale factors are provided by the platform plugin.
This QtGui scaling can be used instead of or in addition to scaling
done by the window system. This distinction is not visible to
applications [when they use Qt API], which will see a change in
the value returned by the devicePixelRatio() accessors as usual.
Introduce a new (private to Qt) coordinate system: native pixels.
The coordinate system stack now looks like:
device-independent pixels (app, upper parts of Qt)
native pixels (lower parts of Qt Gui, platform plugins)
device pixels (backing stores and OpenGL)
Add private QHighDpi namespace with scaling functions that convert
between device-independent pixels and native pixels:
T toNativePixels(T, QWindow *);
T fromNativePixels(T, QWindow *);
Add scaling calls the QWindow (and friends) cross-platform implementation,
around the calls to QPlatformWindow functions. QPlatformWindow now uses
native coordinates - platform code remains largely unchanged since native
coordinates are window system coordinates.
QWindow now uses (possibly) scaled coordinates. This means that
platform plugins no longer can rely on QWindow::geometry() and
related functions. QPlatformWindow::windowGeometry() and other
convenience functions have been added for use when the platform
plugin needs to convert scaled geometry to native geometry.
Add Qt::AA_NoHighDpiScaling, which can be use to disable any
scaling in QtGui, effectively ignoring the environment variables.
(Note that this does not disable any scaling done by the window
system.)
Contributions from Friedemann and Paul.
Task-number: QTBUG-46615
Change-Id: I673bbd69c130e73b13cce83be11bfb28f580bf60
Reviewed-by: Lars Knoll <lars.knoll@theqtcompany.com>
Work around QPaintDevice::metric's int return type
by adding a new metric that returns a scaled devicePixelRatio.
Choose a scale factor that gives us more than enough
range.
The QPaintDevice::devicePixelRatio() convenience accessor
is public API and can unfortunately not be changed
to return a qreal. Add devicePixelRatioF() which
returns the (unscaled) devicePixelRatio.
Change all call sites of QPaintDevice::devicePixelRatio()
to use QPainDevice::devicePixelRatioF().
Task-number: QTBUG-46615
Change-Id: I97ec4000fe379b7ff5e1624a871ae2512790aad9
Reviewed-by: Paul Olav Tvete <paul.tvete@theqtcompany.com>
Reviewed-by: Lars Knoll <lars.knoll@theqtcompany.com>
Allows platform plugins to set the power state for each screen.
Change-Id: I4164ca63cbae9fa2ab186dc84c8f704c33fd7c69
Reviewed-by: Laszlo Agocs <laszlo.agocs@theqtcompany.com>
Introduce overloads in the API to allow specifying multiple color
attachment sizes and formats. When these are in use and MRT is supported,
a texture or renderbuffer is created for each of GL_COLOR_ATTACHMENT0, 1, 2, ...
[ChangeLog] Added support for multiple render targets in QOpenGLFramebufferObject
Task-number: QTBUG-39235
Change-Id: Ie7cfd81d1b796a9166b80dff7513aafe0120d53d
Reviewed-by: Jørgen Lind <jorgen.lind@theqtcompany.com>
There were two problems:
1. qHash() hashed FaceId::uuid, but op== didn't use it for
determining equality. Thus, x == y => qHash(x) == qHash(y)
might be broken if fileName and uuid have an orthogonal
component, if you excuse my vector analogy.
2. To hash fileName and uuid, it concatenated the two,
causing a temporary QByteArray to be created just for
hashing its contents. That prevented marking qHash()
as nothrow.
Fix by using new QtPrivate::QHashCombine and adding
uuid to the list of fields compared in op==.
Change-Id: I49f2379d514d6d3669929e737562cf6de823127e
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@theqtcompany.com>
The sources for QSessionManager live under Qt Gui so it doesn't make
sense to include the class under the Qt Widgets doc module.
Task-number: QTBUG-47274
Change-Id: I18b58c2a274bff323c7364cebbb0429a5a3a97d6
Reviewed-by: Topi Reiniö <topi.reinio@digia.com>
Using the approach we already do for some GLES 3.0 functions we can provide a cross-platform,
cross-GL-GLES wrapper for ES 3.0 and 3.1 functions too.
Applications only have to take extra care about the version requests (context version
and version directives in shader code), the rest of their code can stay the same across
desktop/mobile/embedded, even when ES 3 functions are used.
The new functions are placed to a new subclass which is placed between QOpenGLFunctions
and the internal QOpenGLExtensions. This is necessary because, unlike with QOpenGLFunctions,
there is no guarantee that these functions are always available in all configurations.
When running on desktop OpenGL, we resolve as usual. If the OpenGL version contains the
function in question, either in core or as an extension, it will all just work. This is
handy because it does not rely on 4.x extensions like GL_ARB_ESx_compatibility, and so
ES 3.0 functions will be functional on OpenGL 3.x systems too by just setting a 3.x version
number in the QSurfaceFormat.
We will no longer qFatal on broken systems where the driver returns a 3.0 or 3.1 context without
the corresponding functions present. Instead, we show a warning and gracefully fall back to
resolving as usual, via eglGetProcAddress or similar. For functions that are available in ES2 as
an extension this may just work fine.
Added also an example that runs identically both with OpenGL and OpenGL ES 3 and utilizes some
ES 3.0 features like instanced drawing.
[ChangeLog] Added QOpenGLExtraFunctions providing OpenGL ES 3.0 and 3.1 function wrappers in
a cross-platform manner.
Task-number: QTBUG-46161
Change-Id: I9f929eb61946c35c415b178c4d6ab2c1c958684e
Reviewed-by: Paul Olav Tvete <paul.tvete@theqtcompany.com>
The move special members were inhibited by the presence of
user-defined copy ctor and assignment operator, which we
cannot remove since the class is exported.
Change-Id: I54fe6c28351fe69ca4b75066adb76ea07c959dfe
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
QTextDocument::find has a behavior that is different from the typical
QRegExp/QRegularExpression behavior where finding a given expression is
always constrained by the paragraph (block), regardless of newline
characters in the expression.
Task-number: QTBUG-21785
Change-Id: Ia6a061c6feb5a19e6b8caa3d8009da5ca31ec422
Reviewed-by: Konstantin Ritt <ritt.ks@gmail.com>
Reviewed-by: Simon Hausmann <simon.hausmann@theqtcompany.com>
Literals 1. or 0. have type double, but QGradientStop.first is qreal and
thus can be float depending on platform/configuration, making
qBound call invalid (qBound(0., qreal_value, 1.) for example).
Change-Id: Idab668d417a148bffe2495009ae700de980b73cc
Task-number: QTBUG-47319
Reviewed-by: Lars Knoll <lars.knoll@theqtcompany.com>
Every bit combination is a valid object, so the
correct type classification is primitive, not
just movable.
This is BC, because it just changes whether
the default ctor and dtor are run or not,
and these types don't care.
Change-Id: Ifbfa83128778744f15a086a65c8be1ce01a58820
Reviewed-by: Sean Harmer <sean.harmer@kdab.com>
It doesn't save any space, is not required for ABI compat
(because it's private API), generates more code to extract
the field, and triggers a bug in older GCCs when synthesizing
a move constructor for this type:
src/gui/image/qiconloader_p.h:64:8: error: invalid conversion from 'unsigned char:4' to 'QIconDirInfo::Type' [-fpermissive]
src/corelib/tools/qvector.h:641:13: note: synthesized method 'QIconDirInfo& QIconDirInfo::operator=(QIconDirInfo&&)' first required here
Change-Id: I61e886566b67c7a18a318a3d026dc762600f8ab4
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
... except if the function is a ctor, callers pass rvalues,
and the type has a fast move ctor. In that case, keep passing
by-value and qMove() into place.
Change-Id: I2c0be7d4436327834ddee0531c51c5af352ac74c
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>
Makes for much more readable callers, and, d/t NRVO, isn't less efficient
than the old code, either. Transparently enables moves when QVector should
get them, at some point, too.
Change-Id: I77327cf665272b80277deaaa6dd2ee9c00806464
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
FileInfo is larger than a void*, so holding them in a QList is needlessly
inefficient. Worse, the code could come to depend on the fragile property
of (inefficient) QLists that references to elements therein never are
invalidated.
Change-Id: I772177c5ac544a5fecce2368f628148308ef260f
Reviewed-by: Konstantin Ritt <ritt.ks@gmail.com>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
- don't define special member functions if the compiler-generated
ones are just fine (inhibits move semantics)
- implement all remaining methods inline
- unexport
- remove unused (and, to add insult to injury, never init'ed)
d-pointer. This is private API. No need for a d-pointer,
much less an unused one.
Change-Id: I6979cb5103a361c0313c252d3bf7073a3c47addd
Reviewed-by: Konstantin Ritt <ritt.ks@gmail.com>
These all invoke the QFlags<>(Zero *) ctor, which is designed to
accept a 0, but no other int. But in doing so, it requires passing
a nullptr literal, and 0 is not a nullptr literal accepted under
-Wzero-as-null-pointer-constant or similar warnings.
Fix by using the QFlags::QFlags() ctor instead.
Task-number: QTBUG-45291
Change-Id: I73f9c9f4de11eeb1ba04ace6c7121d17510ea29f
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
This is in preparation of adding -Wzero-as-null-pointer-constant (or similar)
to the headers check.
Task-number: QTBUG-45291
Change-Id: I72ab40b21a9499b53a639564fa45884de17b6c98
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
This also finishes the implementation of the static methods which
were being passed a supportedSchemes argument but weren't using it.
Now they can pass it along to the QFileDialog instance, which then
passes it to the helper used by the QPA implementation of the file
dialog. The default implementation only supports local files and
can therefore ignore this, but other implementations can use this
argument to restrict the protocols allowed to the user.
[ChangeLog][Widgets][QFileDialog] Add supportedSchemes property.
Change-Id: I5235f70e785da1c06866a8355ef98f571890c4a2
Reviewed-by: Kevin Ottens <kevin.ottens@kdab.com>
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@theqtcompany.com>
QBackingstoreTextureInfo is larger than a void*, so holding them in
a QList is needlessly inefficient. Worse, the code could come to
depend on the fragile property of (inefficient) QLists that
references to elements therein never are invalidated.
Change-Id: I5edf846ee8f01ae36c9314147261748270e1fdf6
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@theqtcompany.com>
Reviewed-by: Laszlo Agocs <laszlo.agocs@theqtcompany.com>
First, all values in the array fit into 8-bit, so change the type
from ushort to quint8.
Then, instead of appending a zero as end marker, statically
determine the size of the array and use that number.
Also fixes (benign) off-by-one error in the existing reserve()
call.
Change-Id: Id05b6a56b55d2e0769b00496a55808bb7351b084
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
Any function using the non-dsa helpers and glTextureParameterf() would
crash as this code path failed to initialize the function pointers.
Task-number: QTBUG-47133
Change-Id: I8970c9693ec471a96b135ae5728ab6407ee06a6f
Reviewed-by: Mike Krus <mike.krus@kdab.com>
Reviewed-by: Giuseppe D'Angelo <giuseppe.dangelo@kdab.com>
ItemDecoration is larger than a void*, so holding it in QLists
is needlessly inefficient. Worse, the code could come to depend
on the fragile property of (inefficient) QLists that references
to elements therein never are invalidated.
Fix by marking the type movable and holding in QVector instead.
Change-Id: I06d7a77529f7cff2ba503c5e8d6e5df0ad801a21
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
Reviewed-by: Konstantin Ritt <ritt.ks@gmail.com>
Q(Multi)Map::keys() returns duplicated keys, so keys().size() will
always be the same as the map's size().
So use that directly instead of creating temporary QLists to check
the size.
Change-Id: I0600d1845e25be3b825e4ae470c1ef41a3a5d2c9
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
QModifKeyName is larger than a pointer, so holding it
in a QList is horribly inefficient.
Fix by marking it movable and using QVector instead.
The logic for filling the lists-turned-vectors is a bit
fishy in that it is not thread-safe. Maybe it doesn't
have to, it's not marked as \reentrant.
Change-Id: I8421e6d8b980eff022badfe3c61b3537783b5cfa
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
All values in the table fit into 8 bits, so don't use a
16-bit type.
Change-Id: I685727e16a264a49ea501d36eb6815c6d5aaa9b2
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
Reviewed-by: Konstantin Ritt <ritt.ks@gmail.com>
The MaxKeyCount symbolic constant was introduced for use by QKeySequenceEdit,
but never applied to QKeySequence itself. To prevent the two from getting
out of sync, use it in the QKeySequence implementation itself, too.
This required re-writing most expressions involving QKeySequenceEditPrivate::keys
with algorithms, but we already determined in the discussion around 1cf9a139
that GCC generates the same code as the loops unrolled by hand, and if Clang
doesn't, it's its own fault.
Where using MaxKeysCount was not possible, e.g. in documentation,
static_asserts were used to indicate the need for manual changes.
Change-Id: I559fffdfd552d488448ba4a5a1ac10708c16a2ae
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
Specifically when reading files with broken cmap tables, we could
get some undeterministic results. We handle this more gracefully
by verifying that the offsets are sane and bailing out early if not.
This replaces the current pattern throughout the font engine for
consistency.
Change-Id: I507bba49c0be634afca77d1eb3d199a427451bee
Reviewed-by: Simon Hausmann <simon.hausmann@theqtcompany.com>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
There was an attempt to allow static instances of QIcon in
7727a4355876607a1a022ff54e2570dae883f79c (Qt 4). This patch does only
solve some of the corner cases and broke with
aa5f70c00a. Since the "breakage" has been
there for two years, let's officially declare it unsupported instead of
trying to work around the issue.
Change-Id: I61e12fd03953763ee2e70eae58bcaecabdcb85b8
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
This allows end users to create QSystemTrayIcons that are properly
displayed as monochrome images on OS X, which is especially important
on Yosemite and above when Dark Mode is turned on.
[ChangeLog][QtGui][QIcon] Add ability for QIcons to be marked
as template images. This allows end users to create QSystemTrayIcons
that are properly displayed as monochrome images on OS X, which is
especially important on Yosemite and above when Dark Mode is turned on.
Task-number: QTBUG-42109
Change-Id: I095fb71355ffa572d0abb4769341f68b89b877c4
Reviewed-by: Jake Petroules <jake.petroules@petroules.com>
DestFormat is not defined here but should be Format_RGBX8888
Change-Id: Ie33989a6fc31650434281a6b7b0f4806524268af
Reviewed-by: Konstantin Ritt <ritt.ks@gmail.com>
Reviewed-by: Timo Jyrinki <timo.jyrinki@canonical.com>
Reviewed-by: Simon Hausmann <simon.hausmann@theqtcompany.com>
Found by ICC:
qtextengine.cpp(1259): error: more than one instance of constructor "QFixed::QFixed" matches the argument list:
function "QFixed::QFixed(int)"
function "QFixed::QFixed(long)"
function "QFixed::QFixed(qreal={double})"
argument types are: (uint)
QFixed stretch = QFixed(actualFontEngine->fontDef.stretch) / QFixed(100);
^
Change-Id: I790fb4cae1abc692705253a7863291bc8743b832
Reviewed-by: Konstantin Ritt <ritt.ks@gmail.com>
Reviewed-by: Jake Petroules <jake.petroules@petroules.com>
When cleaning up the shortcut override code, in
d7ca800a87 passing on the autorepeat flags
from the native event was accidentally dropped in the call to
QWindowSystemInterface::tryHandleShortcutEvent.
Task-number: QTBUG-46569
Change-Id: Ib053e11e582de630a1fc1b754f8d10d995c88026
Reviewed-by: Gabriel de Dietrich <gabriel.dedietrich@theqtcompany.com>
Fixed a few qdoc error messages caused by errors in qmetatype.cpp and
qtextlayout.cpp and qtestcase.cpp and qpointer.cpp
Change-Id: I662aa25bedba057e125e289c34787e81793941ff
Task-number: QTBUG-46939
Reviewed-by: Martin Smith <martin.smith@digia.com>
The order of member variables is wp, xp, yp, zp. When performing
operations that touch all members, do so in declaration (=memory)
order.
Might enable autovectorization. Probably leads to slightly
better cache behavior.
Change-Id: Id9541e5a8011c349f5d24c5e9322bb232a7607b6
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Since the introduction of jpeg auto transform, it is possible to read jpeg
header after its state was changed from ReadHeaer to Ready which will lead
to creating some resources again without releasing them.
For example, if you call QImageReader::setAutoTransform(true) and then
call QImageReader::read(), QJpegHandlerPrivate::readJpegHeader() will be
called twice and it will allocate resource again without releasing the old
one.
This patch add a new state ReadingEnd to prevent the header from being
read twice.
Change-Id: If2497f6e3668958c0c792a66e1b77eb2773584a2
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@theqtcompany.com>
In particular the move constructor now performs only
two instead of four assignments.
Change-Id: I033d75635288fbd44060a1ba5f2da23b7e7729f6
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@theqtcompany.com>
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
Task-number: QTBUG-46826
Change-Id: Ie04e69cc5af05650ec6aa42f534f494049a54ee8
Reviewed-by: James Turner <james.turner@kdab.com>
Reviewed-by: Mike Krus <mike.krus@kdab.com>
Reviewed-by: Giuseppe D'Angelo <giuseppe.dangelo@kdab.com>
This is a separate commit to allow other modules to do the transition
before this may vanish or cause a warning.
Change-Id: I556837bf0d2c667f067f4e1fa7918bd1a3f55025
Reviewed-by: Konstantin Ritt <ritt.ks@gmail.com>
Reviewed-by: Simon Hausmann <simon.hausmann@theqtcompany.com>
Instead of reading a QList with the keys and then assigning the list
contents to the keysequence internals without any further checking
(not even for the size of the list, even though op<< creates lists
with just one element), do the processing by hand.
The greatest benefit is that there is are no memory allocations
anymore, except for the detach and whatever QDataStream does
internally.
The other benefit is that the output key sequence is not touched
until the necessary values have been successfully read. This
includes the detach. For this, some very basic error checking has
been added.
Also removed the magic number 4 in favor of the recently introduced
QKeySequencePrivate::MaxKeyCount.
Change-Id: If70f75cc043468d2774a7bb03eebdbf39422043a
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@theqtcompany.com>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
The addShader() function takes a pointer.
Change-Id: Ibd6549ff06cb3bd86096e90ab2637845f032fc68
Reviewed-by: Venugopal Shivashankar <venugopal.shivashankar@digia.com>
We can not rely on the alpha being preserved as part of the source when
the source is converted to a2rgb30 which has much less alpha accuracy
instead we need to apply the alpha on the result.
Change-Id: Iac3104666980825e44ed3490370cdbe21dc68008
Reviewed-by: Gunnar Sletta <gunnar@sletta.org>
Excepting QBitmap and the QTextFormat heirarchy, which don't
have 100% value semantics (still pondering them).
In QPolygon(F), adding the move assignment operator disables
the implicitly-defined copy assignment operator, which therefore
have to be made user-defined. That doesn't change the ABI of
the class, since the base class QVector already had a user-defined
copy assignment operator.
Change-Id: I0b111c1d21cf47f559ada6357c982e3dc26aca68
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
Windows Embedded Compact 7 seems to be affected
by the optimization bug as well.
Task-number: QTBUG-46666
Task-number: QTBUG-46667
Change-Id: I169588cfbc570378ee3066b760d8c084fd28e6be
Reviewed-by: Andrew Knight <andrew.knight@intopalo.com>
Reviewed-by: Maurice Kalinowski <maurice.kalinowski@theqtcompany.com>
Reduces internal memory fragmentation.
The search criteria was:
QVector::append(), QVector::push_back(), QVector::operator<<()
and QVector::operator+=() calls inside for, do and while loops.
Statements inside ifs and out of loops weren't considered.
Change-Id: Ie5aaf3cdfac938994e6e5dfa5f51de501ed79a0c
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>
unless GL_EXT_texture_norm16 (and so GL_RGBA16) is present.
Change-Id: Ic15b81b8134fda147e96301b7f78cabe07a05d9e
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@theqtcompany.com>
Caught being detached a few thousand times per second.
Change-Id: I9f3c80cc5178ec2016041a4f2f6e6e1f458171f1
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@theqtcompany.com>
Reviewed-by: Giuseppe D'Angelo <giuseppe.dangelo@kdab.com>
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>
Reviewed-by: Konstantin Ritt <ritt.ks@gmail.com>
We should accept JPEGs with broken EXIF headers since the header is
optional and can be ignored if broken.
Task-number: QTBUG-46870
Change-Id: I494e5497b8449ce6494285d4d77caadbbb0ccaf8
Reviewed-by: aavit <eirik.aavitsland@theqtcompany.com>
Conflicts:
src/plugins/platforms/cocoa/qcocoafiledialoghelper.h
Manually fixed src/testlib/qtestcase.cpp to return the right type.
Change-Id: Id1634dbe3d73fefe9431b9f5378846cb187624e4
Guarantee the values are exactly the same on all platforms
and avoid the macro redefinition(s) on a newer MinGW.
This also fixes a typo in #define S_IFLNK 020000 (<-- must be 0120000).
Change-Id: I4993d2a9bdb4b0871af110b89c137ce0ac0e953a
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@theqtcompany.com>
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@theqtcompany.com>
Reviewed-by: Konstantin Ritt <ritt.ks@gmail.com>
os.version is the kernel version which is just not very useful or easy to
use for us. Instead, introduce a string that allows easy differentiation
between Windows 7 and 8.
From the adapter identifier the driver description is often helpful too,
especially in virtual machines.
These allow writing rules like:
{ "description": "Use WARP in some Win 8 and 8.1 VMs",
"os": { "type": "win", "release": [ "8", "8.1" ] },
"driver_description": "VMware SVGA 3D",
"features": [ "disable_desktopgl", "disable_d3d9", "disable_d3d11" ]
}
Change-Id: I196f6b44d7bb9e42ef47f61f4c28d8aa79afb7c4
Reviewed-by: Simon Hausmann <simon.hausmann@theqtcompany.com>
This is used by the scene graph to automatically switch over from distance
field text to native text rendering for the given glyph node, which allows
mixing regular text with Emoji in e.g. a Text item without having to set
renderType to Text.NativeRendering.
Change-Id: I5d96d1dab329a975e3442284bf4c5a82174177c9
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@theqtcompany.com>
