The QPixmap in the pixmap cache should not outlive the QPlatformIntegration.
Clear the cache shortly before destroying the platform integration. This is
fixing a crash with the DirectFB plugin in the tst_qpixmapcache test.
Change-Id: I0b1a304dc37d874b412aadb77f8dbaa7d292711f
Reviewed-by: Girish Ramakrishnan <girish.1.ramakrishnan@nokia.com>
Reviewed-by: Samuel Rødal <samuel.rodal@nokia.com>
There's no need to change behaviour now. The current marker (@Invalid())
returns an empty string list, as opposed to a list with one item.
Changing it would break compatibility unpredictably with Qt 4.
So I choose not to change QSettings. Let it be "Done".
Task-number: QTBUG-25110
Change-Id: Id1f353dfed800005d927183da237f3f8357c496d
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@nokia.com>
Reviewed-by: Jason McDonald <jason.mcdonald@nokia.com>
Useful when we get relative events, as otherwise the mouse might end up
far outside the screen boundaries.
Change-Id: I8e3884ab2acb03eaa6afce8926f503dbd03b0c5d
Reviewed-by: Girish Ramakrishnan <girish.1.ramakrishnan@nokia.com>
Growth computations are deferred to QArrayData::allocate, except in the
case of realloc as that functionality is currently lacking in
QArrayData. Since that sits in library code, anyway, it can be changed
later to use a future QArrayData::reallocate.
As it is, reallocData is becoming a model for QArrayData::reallocate
and what it can offer to containers of POD/movable types.
Change-Id: I045483f729114be43e4818149d1be1b333bcbe13
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
In this branch, !IS_RAW_DATA has already established that offset is
sizeof(QByteArrayData) and realloc maintains the assumption.
Change-Id: Ic160e36d7781d4c4f64a3b2ebec98c9cb605b3eb
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Jędrzej Nowacki <jedrzej.nowacki@nokia.com>
For the time being QByteArrayData keeps its independent existence, for
the sake of other modules. Once they have been ported to use the new
initializer macros it can be changed to:
struct QByteArrayData { QArrayData array; };
Extra braces can then be added to the macros.
With respect to source compatibility, the only concern is with other
modules, as QByteArrayData has already changed in incompatible ways with
Qt 4.x
Done-with: João Abecasis
Change-Id: I044e2a680317431777a098feec8839a90a3d3da3
Reviewed-by: João Abecasis <joao.abecasis@nokia.com>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
This eases porting to QArrayData and retains semantics. In append() and
prepend() a conditional was generalized so that no work is done if there
is nothing to add.
Done-with: Jędrzej Nowacki
Change-Id: Ib9e7bb572801b2434fa040cde2bf66dacf595f22
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Jędrzej Nowacki <jedrzej.nowacki@nokia.com>
In conceptual terms, this change increments the Data::alloc member by
one for all strings allocated and maintained by these classes. Instances
initialized with fromRawData retain 0 as the member value, so they are
treated as immutable.
This brings QByteArray and QString closer to QVector, making it possible
for them to reference and share the same data in memory, in the future.
It also brings them closer to QArrayData.
In practical terms all comparisons to the alloc member were changed to
take into account that it also tracks the terminating null character.
Aside from the increment in the alloc member, there should be no user
visible changes.
Change-Id: I618f49022a9b1845754500c8f8706c72a68b9c7d
Reviewed-by: João Abecasis <joao.abecasis@nokia.com>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Change-Id: Ic0bd8f68080bf139d81266d033cfd3c96a8d01f4
Reviewed-by: Donald Carr <donald.carr@nokia.com>
Reviewed-by: Samuel Rødal <samuel.rodal@nokia.com>
Change-Id: Ib40807b97bf8d9ec26e0dfe2797786d9352ce579
Reviewed-by: Kevin Krammer <kevin.krammer@kdab.com>
Reviewed-by: Robin Burchell <robin+qt@viroteck.net>
A recent change in platformsupport required "egl" to be defined,
otherwise the QNX plugin would fail at runtime with unresolved symbols.
Now the EGL config test passes.
Change-Id: I111eb939abac86885bcdb35d5f5899f515bd8c4e
Reviewed-by: Sean Harmer <sean.harmer@kdab.com>
Reviewed-by: Kevin Krammer <kevin.krammer@kdab.com>
Reviewed-by: Rafael Roquetto <rafael.roquetto@kdab.com>
Reviewed-by: Robin Burchell <robin+qt@viroteck.net>
Instead of having to pass -no-xcb -no-eglfs -no-directfb, it's
better to pass -no-qpa-platform-guard which is also resilient
against future platform additions.
Change-Id: Id68bfe3688980fa273665b01b9332f5d6f359491
Reviewed-by: Jørgen Lind <jorgen.lind@nokia.com>
Reviewed-by: Donald Carr <donald.carr@nokia.com>
qpa includes are considered private and should not be included
in public header files.
Change-Id: I26c744ec1d8ddef7b0c11c3d26b593be05f5aa54
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Should be const just like Qt4's QGLContext::getProcAddress.
Change-Id: I273467d5cf852cd49f48cec3f335c4ddac795363
Reviewed-by: Gunnar Sletta <gunnar.sletta@nokia.com>
Change-Id: I08048ecee9b243b122ee93fce316e498aa7e2d51
Reviewed-by: James Turner <james.turner@kdab.com>
Reviewed-by: Samuel Rødal <samuel.rodal@nokia.com>
Reviewed-by: Morten Johan Sørvig <morten.sorvig@nokia.com>
The "Hide <app>", "Hide Others" and "Show All" menu items in Mac
application menu are always enabled, and do not get disabled correctly.
Fix by turning on autoenable for the menu in qt_menu.nib, and by
implementing menu item validation in QCocoaMenuLoader.
Task-number: QTBUG-10705
Change-Id: Ic181dfa26a71acad0067f5269c72517b50b17362
Reviewed-by: Morten Johan Sørvig <morten.sorvig@nokia.com>
(cherry picked from commit 8f23a6be1069455e609e8bea7527726c24bebb36)
This makes it easier to find insignificant tests that have no associated
bug report.
Change-Id: Ia71d59da062818d3860b0365d063e044705267fd
Reviewed-by: Rohan McGovern <rohan.mcgovern@nokia.com>
When encountering a null preprocessing directive (which is supposed to
be ignored), the moc preprocessor will leave a PP_NEWLINE token in the
token stream. That will confuse the parser.
The PP_NEWLINE token need to be ignored in the preprocessing phase.
Task-number: QTBUG-22717
Change-Id: I1e502a7e5bc6fa8ce2f82109ba7199b95747ff0a
Reviewed-by: João Abecasis <joao.abecasis@nokia.com>
Reviewed-by: Bradley T. Hughes <bradley.hughes@nokia.com>
The tools will now generate the new org.qtproject annotations only,
matching the XML generator in the library. They accept both types of
annotations as input though -- and will generate a warning about the
older one.
This commit should be backported to Qt 4, so XML files can start to be
ported.
Task-number: QTBUG-23274
Change-Id: If298c342ab4774cbca1be1898a01af8b46e80446
Reviewed-by: Jason McDonald <jason.mcdonald@nokia.com>
Reviewed-by: Lorn Potter <lorn.potter@nokia.com>
Those error codes have been standardised for years but we haven't used
them until now.
Task-number: QTBUG-23274
Change-Id: Iebc9ded949f363281a4d43fd9d29a284f2e2df08
Reviewed-by: Jason McDonald <jason.mcdonald@nokia.com>
Reviewed-by: Lorn Potter <lorn.potter@nokia.com>
Lots of uses of the annotations and error names, plus a bunch of local
unit test names (including one file that had to be renamed).
The meta object generator is updated to support both the old and new
names. That means some references to com.trolltech *must* remain in the
source code.
Task-number: QTBUG-23274
Change-Id: Icc38ae040232f07c437e7546ee744a4703f41726
Reviewed-by: Jason McDonald <jason.mcdonald@nokia.com>
Reviewed-by: Lorn Potter <lorn.potter@nokia.com>
Change the old com.trolltech ones to org.qtproject and introduce Use
the alternate domain name for the Qt Project because the dash
character is not valid in interface and error names.
Task-number: QTBUG-23274
Change-Id: Iac1699e70525d67f983c10560932acff6b2ecde6
Reviewed-by: Lars Knoll <lars.knoll@nokia.com>
Reviewed-by: Jason McDonald <jason.mcdonald@nokia.com>
The testChildrenLayoutsChanged fails randomly. This happens rarely,
f.i. wasn't spotted by CI when QHash randomization itself was merged;
but is indeed reproducible by running the test a few times in a row.
This is now blocking api_merges integration, and I have no idea
how to fix it.
This patch marks the test as insignificant for now (the bug
tracking this test failure is QTBUG-25325), and switches the failing
tests from QVERIFY(a == b) to a proper QCOMPARE (so that the
expected values do show up in the build logs).
Change-Id: I16f0e28bcbb06dbac2e7169f4676a19ccf626a92
Reviewed-by: Stephen Kelly <stephen.kelly@kdab.com>
QWS used to have a line to change the composite mode from SourceOver
to Source for the top level widget. This wasn't used with QPA and I
removed the internal DontSetCompositionMode in qtbase. It turns out that
the QWS way is the most efficient one to initialize the background of
the widget.
The alternative is to have the QPlatformBackingStore::beginPaint
always clear the entire to be painted area and then paint the background
of the widget. The difference of painting each pixel once or twice is
noticable on embedded platforms and in the range of one to two fps.
Reproduce the issue with:
echo "QWidget {background: transparent}" > style.css
./examples/widgets/wiggly/wiggly -stylesheet style.css
Task-number: QTBUG-24526
Change-Id: Ica4c980bb3bf6eb87ddb5b510ac7493292d01543
Reviewed-by: Samuel Rødal <samuel.rodal@nokia.com>
Reviewed-by: Girish Ramakrishnan <girish.1.ramakrishnan@nokia.com>
http://unicode.org/versions/corrigendum6.html:
> in Unicode 5.0, the list of characters with the Bidi_Mirrored property
> was made consistent for brackets and quotation marks, in preparation for
> new constraints on bidi mirroring. However, after publication of
> Unicode 5.0.0 it was discovered that this change adversely affected
> several quotation mark characters in deployed data.
Task-number: QTBUG-25169
Change-Id: Id49caf401af2d5a1e6dbcc32b2f350aa20b7f901
Reviewed-by: Lars Knoll <lars.knoll@nokia.com>
If we do multiple reads we need to accumulate the total amount of bytes
read, instead of just taking the last read amount into account.
Change-Id: Iaa9b90c269f3ed9d09dae67452ca816d9db6217f
Reviewed-by: Johannes Zellner <johannes.zellner@nokia.com>
Reviewed-by: Girish Ramakrishnan <girish.1.ramakrishnan@nokia.com>
The test fails consistently on the CI (but never fails on any
of the developer machines). This is possibly a timing issue.
Change-Id: Ie40d9c38c3128a93898b0e50bfde5a754bd2b7fb
Reviewed-by: Giuseppe D'Angelo <dangelog@gmail.com>
Reviewed-by: Robin Burchell <robin+qt@viroteck.net>
It goes by the name of QBackingStore these days.
Change-Id: Id46254e47ee0abf9e669862eec6c2205d9634368
Reviewed-by: Holger Freyther <holger+qt@freyther.de>
Reviewed-by: Samuel Rødal <samuel.rodal@nokia.com>
Fixed a bug in qdoc that caused too many end elements to be generated.
Also fixed some doc errors that caused invalid DITA to be generated.
Task nr: QTBUG-25302
Change-Id: Ifbbf457d28c51c2691a252888447739da7713bc9
Reviewed-by: Martin Smith <martin.smith@nokia.com>
Reviewed-by: Casper van Donderen <casper.vandonderen@nokia.com>
QHash employs an optimization for int/uints, squashing the hash
value and the key value inside an union. This obviously works
iff qHash(int k) = k. If the hash value gets salted, the hash
table is corrupted.
This patch removes that optimization by means of a #if 0,
so if further research finds out that we want those 4 bytes back
it's pretty simple to revert.
Change-Id: If273f0bf2ff007f4f2f7c46d2aab364a3b455cf1
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
The key returned by QHash::key is an arbitrary one that maps to the
given value. The test instead relied on it being a specific one.
Change-Id: I090351797e8b52036d78160fd810518a11e8107d
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@nokia.com>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
_QPA_ should only be used for qpa (plugin specific api) files.
Change-Id: I1fb6cd5973fcabec4c7e87eb6ccb048f825e2aad
Reviewed-by: Holger Freyther <holger+qt@freyther.de>
Reviewed-by: Samuel Rødal <samuel.rodal@nokia.com>
The version without argument was kept for binary compatibility when
the configurable ReadElementTextBehaviour was introduced. It is now
dropped in favour of using a default argument value.
Change-Id: Ic08c41d5a5aad9f22df7fc37a2d53ffbc6df1fe9
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Applied Bill King's suggestion in QTBUG-1363. Columns of
hStmt must be accessed in order.
Verified using ODBC driver on SQL Server 2005 on Windows 7.
Added test for length of text field for MS SQL Server over ODBC.
Task-Id: QTBUG-1363
Change-Id: I6673dafe75e3ef394d41e439adb45096c1421068
Reviewed-by: Bill King <bill.king@nokia.com>
Reviewed-by: Mark Brand <mabrand@mabrand.nl>
The Freetype engine currently works better for QML2, whereas
the native engine is better suited to the widgets.
Task-number: QTBUG-24205
Change-Id: I76de98c3e7c84a2d72542ea9860b8be1e67f7e04
Reviewed-by: Jan-Arve Sæther <jan-arve.saether@nokia.com>
Itemview were not following the established pattern
of keeping the referenced object in a QPointer.
This made them a lot more crash prone.
Port of Qt 4 change.
Conflicts:
src/plugins/accessible/widgets/itemviews.cpp
src/plugins/accessible/widgets/itemviews.h
Change-Id: I210b112b9c3647b246fde2d6c69aba9ce8d25bd3
Reviewed-by: Jan-Arve Sæther <jan-arve.saether@nokia.com>
(cherry picked from commit f420ab2870d3c96e9678c24c64b4e26ec8a771bd)
Unfortunately, qdoc is too simple to understand any pointer to member
function syntax.
(the ::* token is not even tokenized. And even if it was, it would be a
difficult way to hack that into the parser. (there is already an ugly
workaround for non-member pointer to function hat works by having '(*'
as a token,but the same hack is not possible for pointer to member function))
So I just put verbatim 'PointerToMemberFunction'
Also remove the obsolete mention that Qt::UniqueConnection is not supported
in that overload. It is now. (And it even contradicts the previous
paragraph)
Change-Id: I8fc9544808c9a462b0f11ccea406e2e33dee15b1
Reviewed-by: Geir Vattekar <geir.vattekar@nokia.com>
Reviewed-by: Casper van Donderen <casper.vandonderen@nokia.com>
This was made inline for the sake of performance, but has in hindsight
proven to be not really worth it. It also prevents us from aligning
the internal datastructure of QPainterPath with that of QPolygonF
and the internal QVectorPath class which would make mapping
between the two inside QPainter a lot faster.
Making all this out-of-line now as discussed in the task
https://bugreports.qt-project.org/browse/QTBUG-19998
so we can adress this in a binary compatible fashion during
Qt 5.x
Change-Id: I61058171ed31f8a845fa45517248367c85ce52cd
Reviewed-by: Lars Knoll <lars.knoll@nokia.com>
Reviewed-by: Samuel Rødal <samuel.rodal@nokia.com>
