Change 939f21be53 introduced a crash when dragging data from
a 3rd party app into the app. Basically the current code assumed
that we always have QCocoaIntegration::instance()->drag()->currentDrag()
but this seems to only be the case if the drag was started by the
app itself.
The crash was found by testing the fridgemagnets example. Just
drag text from other app into the example to see it crash.
Also, refactored the cursor code into a separate method
to simplify code reading.
Change-Id: Ica611a4452a0dd02e01451111aeda14c879f8f1b
Reviewed-by: Filipe Azevedo <filipe.azevedo@kdab.com>
Reviewed-by: Morten Johan Sørvig <morten.sorvig@theqtcompany.com>
The GLX and EGL integrations are skipped correctly but the base class still
contained createPlatformOpenGLContext() with OpenGL-specific types.
Task-number: QTBUG-44998
Change-Id: I727ded7ca8589b163fc1271709dd718572b51c3e
Reviewed-by: Gunnar Sletta <gunnar@sletta.org>
Necessary for iOS. In addition to swizzle we also need to communicate the need
for flipping, so switch to flags instead of bools.
Task-number: QTBUG-40034
Change-Id: I055e591afd838878503be6f5f69aa7347965d9cf
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@theqtcompany.com>
If the asset url indicates that this file engine points
to the asset folder, report the file as a directory.
Also, if the app is authorized to access assets, report
that the directory has read access.
Change-Id: Ic8f656fa30a1b2a0ec6402e8b19256bdf5f7345e
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@theqtcompany.com>
Don't report read access to assets if we know that we are unauthorized
to load them. Note that if authorization is ALAuthorizationStatusNotDetermined
we continue to report read access, since we don't really know the permissions
until we try to load.
Change-Id: If51cfe9f5c57f8f33f463bddf81a77fade5fb89d
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@theqtcompany.com>
Resolve m_assetUrl already when setting file name. The variable will be
used several places in patches that follows.
At the same time, change the logic to be more robust to work around QDir
removing slashes (both single a double) after the scheme. In the end, what
matters is that we still recognize the file name as an asset url, and that we can
restore the original url based on the hash-tag contained inside the file name.
Change-Id: I988c6a73b2484e46d63917b442c13aa5a3666787
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@theqtcompany.com>
On touch platforms, QCursor::pos() will only be 'valid' when a
touch event has (at least once) been translated to a mouse event.
Currently this never happens in QtQuick since QtQuick always
accepts all touch events and performs its own translations.
So rather than setting the geometry of QShapedPixmapWindow from
QCursor directly, we instead base it on mouse events. This will
ensure that we never try to set the geometry of the window to
an 'invalid' value, which can cause a crash on platforms like iOS.
Note that we currenly miss an API in Qt to get the current
touch points. When that is in place, we can also set
a correct start position for the window before the
first mouse move event arrives.
Task-number: QTBUG-45877
Change-Id: I320598e87d43f6e9e087c204a69b95465128f468
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@theqtcompany.com>
Inform the platform that it should not activate the window
when shown by setting Qt::WindowDoesNotAcceptFocus. This compliments
the already set Qt::WindowTransparentForInput, which specifies
that mouse/touch events should also pass through the window. In
other words, the window is just for showing output and should not
respond to input.
Change-Id: I3e90a28be2f2e27e2044effedf64f47c94a857a5
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@theqtcompany.com>
The class inherited QSharedData, had a non-virtual clone() function and
a non-virtual operator==() which compared QPlatformPrintDevice::id().
Derived classes implemented clone() and operator==() comparing ids
to no effect. The class does not have any setters modifying its values,
so detaching, copying and assigning does not make sense.
Remove the inheritance, clone(), and operator==() and make the class
a non-copyable base class. Use a QSharedPointer instead of
a QSharedDataPointer to store it in QPrintDevice.
Remove copy constructors and clone() reimplementations that were never
called in implementations.
Found while investigating QTBUG-44991.
Task-number: QTBUG-44991
Change-Id: Ib79354b37048d04d50d936f1d0ae06c36efaac00
Reviewed-by: Morten Johan Sørvig <morten.sorvig@theqtcompany.com>
Reviewed-by: Paul Olav Tvete <paul.tvete@theqtcompany.com>
In single-exec mode, QDoc must never skip documentation warnings
as there is no subsequent QDoc run to re-execute the same code
paths (in generate phase) that generate warnings.
Change-Id: I8da2f16cfb12b3b3509249d1c9941d63733176a9
Reviewed-by: Martin Smith <martin.smith@digia.com>
The failing test has been blacklisted in 3ed6f74fb2.
Change-Id: I5a2defd839e2f98690fc8cea9ff18e7503caf0de
Task-number: QTBUG-25294
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@theqtcompany.com>
this allows LD_LIBRARY_PATH to take precedence over the hard-coded
rpath, which is the only sane thing to do (which is also why i'm not
adding an option to disable it).
this behavior is consistent with non-linux systems.
the windows version has no auto-detection, just like for gold linker
usage.
Task-number: QTBUG-3069
Change-Id: Ief9ba032291c898d75d76ecc740390954382a804
Reviewed-by: Simon Hausmann <simon.hausmann@theqtcompany.com>
This is for example useful when looking for a possible BLACKLIST
file while doing a shadow build.
Add autotests for blacklist.
Change-Id: I41d3939d31d21d10187fefcb82604736d911b6ad
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@theqtcompany.com>
Reviewed-by: Simon Hausmann <simon.hausmann@theqtcompany.com>
The change would remove the noexception hint on MSVC versions that
doesn't support noexcept but supports their older 'throw()' hint.
Change-Id: Ie5163f2413522f427279f59c8562c0ce4769bc82
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>
There is no ApplicationsLocation on iOS (at least not one that
is public API). NSApplicationDirectory just points to a non-existing
write-protected path inside the app sandbox. Rather than returning something
we know is wrong, it's better to return an empty string.
Change-Id: I2ebc151f15509ed5699af05def5c708a56eeaf31
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@theqtcompany.com>
NSDesktopDirectory points to a non-existing write-protected path inside
the app sandbox. According to QSP documentation, we should fall back to
use the home directory instead.
Change-Id: I2c370af7758ac043eddcff84aa287eacc754ae38
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@theqtcompany.com>
This was completely broken until now because the ranges were assumed to be
from the OS/2 header map. These are actual unicode ranges from the cmap,
so they need to be matched by looping over the values and checking if they
fall within a given writing system range.
Task-number: QTBUG-44155
Done-with: Peng Wu <peng.wu@intopalo.com>
Change-Id: I933429627c4dbf3377f41c9281df5a801057698f
Reviewed-by: Oliver Wolff <oliver.wolff@theqtcompany.com>
Moves some of them to the .rodata section, the rest at least to
.data.rel.ro[.local].
Change-Id: I85676ddf22b0c0097f3f0dce4c3dc018dc29d045
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Giuseppe D'Angelo <giuseppe.dangelo@kdab.com>
QGraphicsWidgetPrivate::init had a special code path for setting
the item's parent. For some reason that code path caused the
ItemChildAddedChange notification not to be sent to the parent
element, which is wrong. Instead use the "normal" path, which is
what the QGraphicsItem constructor does anyhow.
Change-Id: Iad84cae05d797022a45977d35ca00c80c17c306a
Task-number: QTBUG-45867
Reviewed-by: Andreas Aardal Hanssen <andreas@hanssen.name>
...by dropping the use of qIsNull(), which, in Qt 4, distinguished
between -0.0f and +0.0f.
But mathematically, whether x, y, z are ±0 doesn't change the fact
that the result is the identity element (x, y, z should contain the
identity element for addition and w the one for multiplication), or
the null element (additive identity).
So using qIsNull() was wrong even in Qt 4. In Qt 5, qIsNull() returns
true for both ±0, so we can just as well compare to 0.0f instead,
which allows to mark these functions constexpr (qIsNull() can't be).
Do so.
Change-Id: I78b1fa7890036dd3cb4de7f90b75d439f9835e73
Reviewed-by: Konstantin Ritt <ritt.ks@gmail.com>
currently when adding an event it is not possible to know when processing
it has finished.
In particular if the event is ignored no method is called.
Adding virtual methods to the private implementation (binary compatibility).
These methods allow for extended automatic testing of the state machines.
(cherry picked from commit e7feb95628)
Change-Id: Iaa48fb9d7f6a6cde1a8a7a2bece7b4df55c147e8
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@theqtcompany.com>
The behavior of "external" and "internal" transitions is identical,
except in the case of a transition whose source state is a compound
state and whose target(s) is a descendant of the source. In such a case,
an internal transition will not exit and re-enter its source state,
while an external one will.
[ChangeLog][State machine] Added support for internal transitions.
Change-Id: I9efb1e7368ee52aa2544eb84709a00ae3d5350d3
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@theqtcompany.com>
As nothing changes in the state machine when selecting transitions for
events and then calculating the exit- and entry-sets, some calculations
can be cached.
The exit set for a transition was calculated multiple times. First in
removeConflictingTransitions, where the two loops would each calculate
them multiple times. Then secondly in microstep(), which would calculate
the exit set for all transitions.
Transition selection, exit set calculation, and entry set calculation
all calculate the transition domain and effective target states for
transitions.
Change-Id: I217328a73db2f71e371eb5f60a0c7b222303f0ca
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@theqtcompany.com>
Add the noexcept attribute to all move constructors and assignment
operators in QtGui that didn't already have it.
Change-Id: Idcdf79ae8349b8793e7394b5ae7c08e6111fbc9a
Reviewed-by: Konstantin Ritt <ritt.ks@gmail.com>
Reviewed-by: Gunnar Sletta <gunnar@sletta.org>
It's undefined behavior to left or right-shift a signed integer such
that the sign changes. Since SymbolCsbBit is 31, make sure we use
unsigned numbers.
Found by ICC
qplatformfontdatabase.cpp(614): error #68: integer conversion resulted in a change of sign
Change-Id: Idf715b895bac4d56b4afffff13da78d294b1248e
Reviewed-by: Konstantin Ritt <ritt.ks@gmail.com>
Don't remove ampersands when setting the text: they will be removed when
the text is displayed. This fixes double removal of ampersands.
[ChangeLog][QtWidgets][QToolButton] Fix double removal of ampersands
Task-number: QTBUG-23396
Change-Id: I56bf50eb24aae32a81d614824aca0b63363587c8
Reviewed-by: Timur Pocheptsov <Timur.Pocheptsov@digia.com>
Always accept drag enter events if the mime type and the drop actions
matches. Whether the drop can actually happen on the currently hovered
index will be decided in the following drag move event.
Change-Id: I27e865cb65513dfe3f57ad3f1bc8cebf4c29a692
Task-number: QTBUG-45037
Reviewed-by: David Faure <david.faure@kdab.com>
Make it clear (just like the other methods) that the indexes refer
to the source model, not the proxy model. There was already a note
in the documentation, but it was at the end of it; instead, change
the formal arguments names.
Change-Id: Ia9592f2b080ff276a62de1713a9623e0f3a50cf6
Reviewed-by: Tobias Koenig
Reviewed-by: Sérgio Martins <sergio.martins@kdab.com>
Reviewed-by: David Faure <david.faure@kdab.com>
There is no any sense in detecting the encoding of an empty string ;)
Change-Id: I1c7af07bd7c3e7e7cf67421a2cb3a1123ca57650
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
The init()/cleanup() code in tst_qfiledialog and tst_qfiledialog2
currently differs and fails to clean up the settings file
since it only removes the legacy settings under the Qt group
and instantiates a new QFileDialog while the QSettings class
is still in scope. Also, it has no means of clearing the
setLastVisitedDirectory(), which causes the
tst_QFiledialog::completer() and tst_QFiledialog::history() tests
to interfere, leaving the settings in an invalid state.
tst_qfiledialog2 does not use QStandardPaths::setTestModeEnabled(().
- Ensure the last visited URL is always clean by
making QFileDialogPrivate::setLastVisitedDirectory()
static and calling it from init().
- Introduce a cleanupSettingsFile() function to the tests that
cleans both groups and call it from initTestCase() and cleanup()
to ensure a clean state.
- Add QStandardPaths::setTestModeEnabled() to tst_qfiledialog2.
Fixes sporadic test fails when executing
tst_QFiledialog::completer() and tst_QFiledialog::history()
in a sequence.
Task-number: QTBUG-45764
Change-Id: I24de3caabf77be067b385d64ff11b7a07fe12b72
Reviewed-by: Simon Hausmann <simon.hausmann@theqtcompany.com>
This one has not been unsed for ages (since Qt3?).
Change-Id: Iaf514db1b698b34a303f34c150b72db989eb176c
Reviewed-by: Simon Hausmann <simon.hausmann@theqtcompany.com>
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@theqtcompany.com>
Many times set but never used.
Change-Id: I297f21d4b9878e5f9559a2dfdb2d3de34107bb7d
Reviewed-by: Simon Hausmann <simon.hausmann@theqtcompany.com>
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@theqtcompany.com>
1. Checking if `db->count == 0` inside the
`for (int x = 0; x < db->count; ++x)` loop makes absolutely no sense;
2. The family gets loaded just a line above (ensurePopulated()).
Change-Id: I72dbd42565c4f5a5d9ff8879f10ee0ece7298fa7
Reviewed-by: Simon Hausmann <simon.hausmann@theqtcompany.com>
There is no sense in comparing fonts with different resolve_mask-s.
Change-Id: Icfdaf494fce8a59b7138d96fdf7354cc0514ca6a
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@theqtcompany.com>
When resolve_mask is 0, the font inherits just everything, and there is
no need in checking for any resolved attributes;
the resolved font always inherits dpi, so do not check it either.
Change-Id: Ie5d7ced0859d46a9237447e29051a22569480a51
Reviewed-by: Simon Hausmann <simon.hausmann@theqtcompany.com>
These values were added as part of upgrading to Unicode 7.0
Change-Id: Ib208828a7685ba4f89f0e2f06033f74c8ff13532
Reviewed-by: Andy Shaw <andy.shaw@theqtcompany.com>
The flags are the same across all GCC flavors so put the definition
in the common gcc-base.conf instead of duplicating it in g++-base.conf
and clang.conf
Change-Id: I1ba2c3c314d3a02b559c384aecef74240f69f659
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@theqtcompany.com>
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@theqtcompany.com>
In Android 5.1 the field name for the inset state member in
InsetDrawable changed from mInsetState to mState.
Task-number: QTBUG-45714
Change-Id: I0ebada1ef90954013e5357cbd10df925f8f05295
Reviewed-by: BogDan Vatra <bogdan@kde.org>
Completes the inplace converters so that we can rely on inplace
conversions to succede as long as the image depth is the same.
Change-Id: Ia1ae34b5de1bc16e87ff5403bdacfcae44a22791
Reviewed-by: Gunnar Sletta <gunnar@sletta.org>
Some inplace conversions would not set the requested image format in
the returned image, due to the same conversion being used for several
destination formats.
This patch ensures all inplace conversions return the right format,
and adds testing infrastructure to easily test iterate over all formats
so that they all can be tested.
Change-Id: I68ab17a713ddb2aa8ed6aaf0a0a1ec8bbcb56090
Reviewed-by: Gunnar Sletta <gunnar@sletta.org>
- Introduce QDebugStateSaver for all debug operators.
- Remove the "Flags=" from enumerations since their type is now output
by default.
- Added some spaces since the previous formatting relied on space=true
as a result of some debug operators erroneously returning debug.space(),
which is now fixed in qtbase.
- Fixed formatting, added noquote() where necessary, added some newlines,
used stream modifiers instead of QString::number(n, 16) to output
hex numbers.
- Fix indentation.
Change-Id: I64123a4262916e21448cda2aa61ae1100f07291a
Reviewed-by: Joerg Bornemann <joerg.bornemann@theqtcompany.com>
The one converters from RGB30 was misplaced in the method table, and
the unpremultiplication from A2RGB30 to RGB30 had an underflow mistake
when alpha was 2.
Change-Id: I92c11ede28611a3dbdce72aca1898845c120c209
Reviewed-by: Gunnar Sletta <gunnar@sletta.org>
X server may return an empty physical screen size, for example
on VNC, Xephyr or some not very well supported hardware.
In this case it's possible to use the size of the virtual desktop,
but until now it was done only in the QXcbScreen constructor.
Move it to QXcbScreen::updateGeometry() and calculate physical screen
size using the DPI of the virtual desktop.
Task-number: QTBUG-45564
Change-Id: I6b757818a2fcefdd7b2c0aa31b840a88d625d6ae
Reviewed-by: Daniel Vrátil <dvratil@redhat.com>
Reviewed-by: Paul Olav Tvete <paul.tvete@theqtcompany.com>
