src/plugins/platforms/eglfs/deviceintegration/eglfs_kms_egldevice/qeglfskmsegldeviceintegration.cpp: In member function ‘bool QEglFSKmsEglDeviceIntegration::setup_kms()’:
src/plugins/platforms/eglfs/deviceintegration/eglfs_kms_egldevice/qeglfskmsegldeviceintegration.cpp:391:28: error: ‘encoder’ may be used uninitialized in this function [-Werror=maybe-uninitialized]
m_drm_encoder = encoder;
^
src/plugins/platforms/eglfs/deviceintegration/eglfs_kms_egldevice/qeglfskmsegldeviceintegration.cpp:392:29: error: ‘connector’ may be used uninitialized in this function [-Werror=maybe-uninitialized]
m_drm_mode = connector->modes[0];
^
cc1plus: all warnings being treated as errors
Change-Id: Ic444450d06a6dd8047bd6871febb08a4a3fde6f9
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
When using threaded rendering the render-thread might be half-way into
rendering a frame when the application is backgrounded, resulting in
the following swap happening on a non-exposed window. This may result
in the system killing the application, as rendering is not supposed to
happen when an application is backgrounded, so we skip the flush.
Change-Id: I9ab8f2c4617391fd827558af9fb473f1734b3688
Reviewed-by: Richard Moe Gustavsen <richard.gustavsen@qt.io>
Using dispatch_async to deliver the initial application state at startup
was broken, as that would leave the application in the default application
state, inactive, until the next runloop pass. This became a problem when
an application was started backgrounded, eg. in response to location
updates or a Bluetooth accessory waking it up, as it would have a small
window of time at startup where it would think it was able to render
content (since the window was exposed), while in fact the application
was running in the background. iOS will in these situations kill the app
for doing background rendering.
Change-Id: I1ab4a6af08a154d8625c6451b4b5c8f4453e6b43
Reviewed-by: Richard Moe Gustavsen <richard.gustavsen@qt.io>
During device rotation, the backing CEAGLLayer of our custom UIView is
resized by the system. Normally this is the time where we would then
reconfigure the corresponding renderbuffer that we render to, which
shares memory with the CEAGLLayer, but we chose a lazy approach where
we'd defer the reconfigure until client code actually called makeCurrent.
This caused problems because not only did we implement the lazy reconfig
in makeCurrent, but in every QIOSContext function that operated on the
default FBO, including swapBuffers(). When using threaded rendering,
such as in Qt Quick, the render thread may be half way in rendering a
new frame when the system resizes the CEAGLLayer, and we pick up that
resize on the swapBuffer call and allocate a new renderbuffer, before
flushing the queued up GL commands that were operating on another
renderbuffer of a different size. This resulted in the following crash:
0 - gpus_ReturnObjectErrorKillClient()
1 - gpusSubmitDataBuffers()
2 - glrFlushContextToken()
3 - flush(__GLIContextRec*)()
4 - QIOSContext::swapBuffers(QPlatformSurface*)
...
We solve this by still being lazy in how we reconfigure, but limit the
reconfigure to makeCurrent(). If the CEAGLLayer is resized in between
two frames, we skip the half-drawn frame. The old frame will then be
scaled to match the new size by the system, but this is preferable to
flushing a new frame that may have been drawn with two conflicting
window geometries.
Task-number: QTBUG-50017
Change-Id: Ie229f26d156dfbfc7ed8d9efd0eb5e992eee73f1
Reviewed-by: Richard Moe Gustavsen <richard.gustavsen@qt.io>
Improves performance over the default timer-implementation, and allows
us to control the rate and paused state of the display link.
Change-Id: I05761b6eb48f5e91af35735e2faa477427cd8440
Reviewed-by: Richard Moe Gustavsen <richard.gustavsen@qt.io>
this apparently makes no difference except for not creating bogus .prl
files - presumably, the correct path is coming from somewhere else
already.
Change-Id: Ia3f3c44e506ba14b533ff097f05acecf1e86cfb3
Reviewed-by: Laszlo Agocs <laszlo.agocs@qt.io>
Currently TuioTouch plugin reports touch areas of size 1x1.
It's inconsistent with the platform plugins which report
empty touch areas when their size is unknown.
Change-Id: Iea8ad38a5712b666126780411380d963877b0229
Reviewed-by: Shawn Rutledge <shawn.rutledge@qt.io>
Reviewed-by: Ariel Molina R <ariel@edis.mx>
Reviewed-by: Robin Burchell <robin.burchell@viroteck.net>
A bug in the Windows C Runtime causes text mode pipes to drop newlines
sometimes. This bug was hidden because of another bug in rcc which
caused newlines to be redundantly duplicated. When the latter bug was
fixed (commit 53d5811b) the former bug was exposed, causing invalid
vcxproj files to be generated. The Windows bug is described here:
https://connect.microsoft.com/VisualStudio/feedback/details/1902345
The workaround is to avoid text mode, and do the conversion of "\r\n"
to "\n" ourselves (which we were already doing anyway).
Change-Id: I792599a4cd7822f109fa921f02207fb1b144b1d1
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@theqtcompany.com>
If macx is present, osx AND macos should be required.
Change-Id: I5cd9d41270c741dc314720a1119b163dd17fdfd7
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@theqtcompany.com>
The visual column index needs to be replaced by the logical column
index in QModelIndex::sibling() call in QTreeView::moveCursor() method.
[ChangeLog][QtWidgets][QTreeView] Fixed a key navigation bug
when the columns were reordered.
Task-number: QTBUG-52793
Change-Id: I4dc89a2705966a94b20edd7f9a4422f089bb670f
Reviewed-by: Thorbjørn Lund Martsum <tmartsum@gmail.com>
If QT_NO_TABLETEVENT defined, then there is no sense to create tablet
events.
Its better to create general touch events in this case.
Task-number: QTBUG-53887
Change-Id: I2fabc2241158d54d6c39a2f6071ab874f7debd39
Reviewed-by: Shawn Rutledge <shawn.rutledge@qt.io>
This resolves an issue where qmake would generate a Makefile with an
install command immediately followed by a test command, with no
intermediary newline and tab to separate them.
Task-number: QTBUG-54035
Change-Id: I7f9226f25e92b49ce689d252e9c4a58b877f2972
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@theqtcompany.com>
As common bsd's are using the -ffunction-sections flags from
mkspecs/common/gcc-base.conf, also use the --gc-sections ld flags
like Linux as all are using GNU ld as their linker and all are
capable of using it. The last remaining problems with --gc-sections
removing the .name sections providing the OS tags for distinction of
the execution layer were solved with NetBSD 6.0 and OpenBSD 5.4, see
http://gnats.netbsd.org/40401 and http://www.openbsd.org/plus54.html
so this option can safely be used now to reduce the size of moc and rcc
tools.
Change-Id: I74ccd4f6bc607f6d82d32fc864875f26b26bf167
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
It's a problem when building for 64-bit where the two types no longer
match.
Change-Id: I8c31915caf81a60d635c79816a3a2d5d36742ff9
Reviewed-by: Dan Cape <dcape@qnx.com>
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@theqtcompany.com>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Avoid all inplace modification of images using external data
buffers. Since the QImage methods are documented to create a
(modified) copy, there is afterwards no API requirement on the
lifetime of the data buffer.
This patch supersedes 509bc7e59c
Task-number: QTBUG-53721
Change-Id: I3ccc01619eb61d8630104449394e0b76df0af695
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@qt.io>
Usage of this macro has been removed, and so can the macro.
Task-number: QTBUG-52389
Change-Id: I28a5459e577b78f0f9907612893d6850848f405d
Reviewed-by: Yoann Lopes <yoann.lopes@qt.io>
For QIconDirInfo::Scalable directories condition for
directoryMatchesSize was mistyped. In particular only the minSize was
considered which could lead to false positive checks.
[ChangeLog][QtGui][QIconLoaderEngine] Fixed theme lookup for scalable
entries
Change-Id: Ic7e06cc0a2e4be69e6633847cef8c2c5686378ea
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
The two PDB files that the MSVC compiler and linker create are supposed
to be handled differently and should not share the same file path.
Using the same file path for both can result in corrupted PDB files and
longer build times.
Use $${TARGET}.vc.pdb in the OBJECTS_DIR for the compiler and
$${TARGET}.pdb (the default) for the linker.
Task-number: QTBUG-53895
Change-Id: I31f06d4a674a3aa2afe5b30499bae820e5caf2c4
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@theqtcompany.com>
0 must not be used as a null pointer constant
Change-Id: I082d0e99c105fb02980b9cf390e7f6e4c9ad0869
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
- Adapt to the OS X => macOS rename in Q_OS_ macros/docs, qmake scopes,
file selectors, etc.
- Add new QSysInfo values and MAC_OS_X / __MAC_ / __IPHONE_ values for
macOS 10.12 and iOS 9.1 through 10.0.
- Update prettyProductName with new macOS "Sierra" codename.
Change-Id: Id976530beeafa01b648ebaa16f4a8f0613fcaf75
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Add the strip commands for installation commonly used also on BSD
systems.
Change-Id: I4113ffa559a737ef92afb7c90ab5e1bff902b1bb
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@theqtcompany.com>
This is the second attempt to remove 'insignificant' from UDP socket test to:
- make changes ending up in regressions more difficult (hopefully impossible) to merge
- switch to BLACKLIST if needed
- make flaky tests more visible.
For now this 'back to significant' will be accompanied by extended BLACKLIST.
New in BLACKLIST:
* OS X - datagram size-related problems (fixed in 5.7)
* OS X - multicastLeaveAfterClose - will probably stay BLACKLISTED,
seems to be a Darwin's quirk, can be fixed in OS X >= 10.10.
* windows ...
tst_QUdpSocket::echo seems to fail randomly on OS X/linux and
it looks like it fails at the same time on different machines,
should be something server-related (a guess only).
Change-Id: Ib344348ffab03fab1b9309b80449a04d8ce247c6
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
The unsigned flag in columns was ignored when creating the list of
bound values in a mysql table. So the result iteration with
QSqlQuery::next stops after the first wrong truncated value.
[ChangeLog][QtSql] Fixed QSqlQuery::prepare value truncation error when
using UNSIGNED values in a MySQL database.
Task-number: QTBUG-53969
Task-number: QTBUG-53237
Change-Id: I10d977993445f2794f1dd8c88b2e83517ef524f3
Reviewed-by: Milian Wolff <milian.wolff@kdab.com>
[UIFont familyNames] will return a list of fonts that include the fonts
Telugu Sangam MN, Heiti SC, Heiti TC, and Bangla Sangam MN, but when
calling [UIFont fontNamesForFamilyName:] for these fonts we get an
empty list.
The problem appeared when we tried to then populate these fonts, as
CTFontDescriptorCreateMatchingFontDescriptors() would return a list
of font descriptors from the PingFang SC font when called with
NSFontFamilyAttribute = "Heiti SC". This is due to PingFang being a
replacement for Heiti in later iOS versions.
Task-number: QTBUG-50624
Change-Id: I22684e247d472c30775321b6976b3aeb6ea579f5
Reviewed-by: Jake Petroules <jake.petroules@qt.io>
So that it actually compiles with gcc-linaro-arm-linux-gnueabihf-raspbian.
Remove also flags a device spec should not set (like -std and -O).
Change-Id: Ib7a3bc298e60715410d5c00fbc22199ab4711fa7
Reviewed-by: Andy Nichols <andy.nichols@qt.io>
The rpi3 spec added previously is only half of the story now that we
have the option to use another GL driver.
Change-Id: I1b1edde77bcc6d2f382f1021de9c594c27c34d6f
Reviewed-by: Andy Nichols <andy.nichols@qt.io>
Do the multiplication of the normal components in floating point to
avoid integer overflows. Also add an assert, since a scale of 0 here
will cause a normal of (0, 0) which will assert further into the
drawRectangle() function, and the cause is not immediately clear.
Task-number: QTBUG-51956
Change-Id: If7187d56af28eaa149f8f362050a587da5adb262
Reviewed-by: Yoann Lopes <yoann.lopes@qt.io>
Assuming a certain max height for glyphs would make it impossible
to render certain fonts. The follow up to this change is in
Qt Quick, where the code must also be adapted to make it work.
Task-number: QTBUG-52389
Change-Id: Iabebb2de21a92d1537b2965aa6603529c1d5d587
Reviewed-by: Yoann Lopes <yoann.lopes@qt.io>
If a user's code, attached to sslErrors signal, calls abort/close
or disconnectFromHost but our SSL socket was configured not to verify a peer,
no need to continue handshake after calling checkSslErrors
(and finally crashing on invalid 'ssl' pointer).
Task-number: QTBUG-53906
Change-Id: I7f185511d278f9d6f16e7d6c5ba424707141459c
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@theqtcompany.com>
Push conversions from pthread_t to Qt::HANDLE and back into functions.
The casts that were being used didn't work for the unusual 64-bit
pointer/32-bit int combination that QNX is using for 7.0. HANDLE ends
up as a 64-bit pointer and pthread_t ends up as a 32-bit integer. g++
considers the precision loss when converting from the 64-bit pointer
to the 32-bit integer an error. Better to have the casts hidden in
functions so it's easier to adjust them for unusual combinations such
as this.
Change-Id: Ia156b26224a0f7edc1c31e3d1ee8b21191381698
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
GetModuleFileName exists for Windows 10 and upwards, hence use the
generic version from the win32 mkspec. This allows to create a
QCoreApplication object with nullptr argv, as the application filename
is identified via the binary itself and not via arguments. A couple of
auto-tests use this method to create multiple application objects during
runtime.
Unfortunately we cannot apply this for msvc2013, even though MSDN states
the GetModuleFileName exists, it fails to compile for Windows Phone 8.1.
Change-Id: I2b8b988107487ef3785462f8ca40b0a6e0623e32
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Oliver Wolff <oliver.wolff@qt.io>
In case QEventDispatcherWinRT::runOnXamlThread returns an error the
runtime sets the status of IAsyncInfo to Error. At the point when the
IAsyncInfo destructor is invoked, an unhandled exception is thrown
indicating the error has not been handled, causing any application to
just crash deep inside the Windows platform libraries.
Hence, in case runOnXamlThread returns non-S_OK we have to manually
invoke Close() of the IAsyncInfo to tell the system we have taken care
of everything.
Change-Id: I3ac1e2ec2726f42e44f4f9a92191e454711120dd
Reviewed-by: Oliver Wolff <oliver.wolff@qt.io>
Contrary to our initial assumption, this should be added by default to
allow full functionality during development by default. Developers need
to decide which capabilities to use during the publishing step already,
hence improve DX by adding this.
Task-number: QTBUG-50847
Change-Id: I36e0214f7bcf8610d31851eea172aba3944cfd99
Reviewed-by: Oliver Wolff <oliver.wolff@qt.io>
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
There is no need for the functions to assert. By returning false they
show that something went wrong and the error will be handled
gracefully.
Change-Id: Ib026adf5c6fb23b5e6b5598533caec3b3669220c
Reviewed-by: Maurice Kalinowski <maurice.kalinowski@qt.io>
If runOnXamlThread returns anything but S_OK an exception is thrown
which might cause the application to terminate. So we give the lambda
a reference to hr and check that reference instead of runOnXamlThread's
return value for errors.
Change-Id: I1188ea720c63f6fdf43400f2f3ff928b72afc58e
Reviewed-by: Maurice Kalinowski <maurice.kalinowski@qt.io>
runOnXamlThread should always return S_OK (causes exceptions otherwise)
so we need another way of error reporting to the outside (hr reference).
A failed hr is reported to the outside (of the lambda) and interpreted
as an unknown error.
Specific error cases like the given address not being a local address
or the given address being in use already are handled inside the lambda.
The specificErrorSet variable is necessary in these cases so that the
error is not overwritten by unknownError.
Change-Id: I198d66fe97726d5127bf31e50c7eff3363d5259c
Reviewed-by: Maurice Kalinowski <maurice.kalinowski@qt.io>
Q_COMPILER_CONSTEXPR can be undefined (or not defined at all) to
indicate that constexpr should not be used regardless of the compiler's
ability to support it. This is done for QNX because some C library
floating point functions used in the Dinkumware C++ library aren't
constexpr functions; i.e., the library doesn't have proper constexpr
support even though the compiler does.
(cherry picked from commit d87242968f)
Change-Id: If0bdeb2180710dd9ccd97d79fa91cf9ff42f7990
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
BSD OS mkspecs share mainly the same configurations except very few
differences. Merge into a common/bsd/bsd.conf file to be used across all BSD
OSes in their respective qmake.conf and add a qplatformdefs.h that contain
the common defines to be re-used in the BSD mkspecs.
The change includes the usage of <sys/param.h> also on NetBSD through the
common qplatformdefs.h, which is intended for using NetBSD's version defines.
Change-Id: Ibb0ac9e4c8bb5aff7d0febdcab1a4b9600a61117
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Use larger system metric SM_CXICON instead of SM_CXSMICON and set NIIF_LARGE_ICON
in the NOTIFYICONDATA. De-inline function iconFlag() and set NIIF_USER for user
icons for NIIF_LARGE_ICON to take effect.
Task-number: QTBUG-53591
Change-Id: I47c8e0a020ef94241403b1fbae76d5ef2e074301
Reviewed-by: Oliver Wolff <oliver.wolff@qt.io>
The animated glow always has a length of 120 but was previously drawn
with the clipping rect of the bar's actual size. For sizes smaller than
120 the native theme part would be clipped and the black gradient would
show.
Change-Id: Id81e39c405ef81ae08af0f64249568016944bdf1
Task-number: QTBUG-51266
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
Reviewed-by: Alessandro Portale <alessandro.portale@qt.io>
This makes it easier for users to find out how to get a file's
extension when browsing docs.
Change-Id: I08a1b620dea5432462133324824fae85754b9a09
Task-number: QTBUG-53380
Reviewed-by: Topi Reiniö <topi.reinio@theqtcompany.com>
The backing store would not report itself dirty if only render-to-
texture widgets were dirty.
This caused QOpenGLWidgets not be repainted after being remapped if they
marked themselves dirty while unmapped.
Task-number: QTBUG-50818
Task-number: QTBUG-51815
Change-Id: If43f7cbb6f4d44831b4c21b89f884d6bcaebf87c
Reviewed-by: Laszlo Agocs <laszlo.agocs@theqtcompany.com>
the assumption stated in b67a0836d is actually invalid - configure sets
build_all without debug_and_release there. debug_and_release does
actually imply build_all, though.
to make things less confusing, don't let configure inject
iphonesimulator_and_iphoneos into all projects, but handle it like
debug_and_release instead.
Change-Id: Ib7acdc63308a538862fc603428f81aba60bca08e
Reviewed-by: Jake Petroules <jake.petroules@qt.io>
