Instead we deal with any differenced by setting variables in the top
level makespecs, that are used by the common makespec configs.
Change-Id: Iae1fb5fef8c95778511ed400008731989b446f3c
Reviewed-by: Morten Johan Sørvig <morten.sorvig@digia.com>
The LLVM version in Xcode 4.5 and below does not handle the GNU
assembler syntax in the pixman assembly file.
A possible alternative workaround is to include a preprocessed assembly
file using https://github.com/hollylee/gas-preprocessor.
Change-Id: Id95add669c60d3a7da823e5975afdd1f88f71977
Reviewed-by: Morten Johan Sørvig <morten.sorvig@digia.com>
NEON detection is handled by configure's arch test, and THUMB2 is the
default for ARMv7.
Change-Id: I8ec3ce0ec6af9ad8d9509890aa1f8c87e18364d5
Reviewed-by: Morten Johan Sørvig <morten.sorvig@digia.com>
We now require SDK version 4.3 or above, and armv7.
Change-Id: I4766e277a3a4a32712bf2ec27fede694e8316c95
Reviewed-by: Morten Johan Sørvig <morten.sorvig@digia.com>
There's no need to duplicate the logic for device vs simulator. The only
difference is the iPhoneOS/iPhoneSimulator name.
Change-Id: I87c57fa785279a3ee258b76fdac8317e52e7daa2
Reviewed-by: Morten Johan Sørvig <morten.sorvig@digia.com>
We treat iOS as a variant of Mac OS, so for iOS both Q_OS_MAC and
Q_OS_IOS will be defined. This matches what Apple assumes in the
header file TargetConditionals.h
Change-Id: I55cc851401b748297478e4c32e84e0f6e1fdfc28
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
QMacStyle is not buildt as a part of iOS. So make sure we dont
reference it from QStyleOption
Change-Id: I98e779c576d0607402e45a19b457144a6bdfc73b
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@digia.com>
Make sure the libraries dont depend on Cocoa. This will be
picked up by libtool, and make all apps and examples link
against cocoa too (which will ofcourse fail)
Change-Id: I5654bb08c4ed376fc7ee74da422d903270a8af38
Reviewed-by: Morten Johan Sørvig <morten.sorvig@digia.com>
The plugin has been renamed from uikit to ios.
Other than that, the plugin will now build, but do nothing. Most of
the Qt4 code is preserved, with a rough translation
into the Qt5 qpa API. A lot of code has simply been commented
out so far, and most lacking at the moment is the event dispatcher
which will need to be rewritten, and the opengl paint device
implementation. But it should suffice as a starting ground.
Also: The plugin will currently not automatically build when
building Qt, this needs to be enabled from configure first.
Change-Id: I0d229a453a8477618e06554655bffc5505203b44
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@digia.com>
We can use xcode-select -print-path to get the /Developer directory.
This also removes the need for the "legacy" makespecs, which only
differ from their non-legacy counterparts in the location of the
Xcode developer directory.
Change-Id: Ia9245033a4b82cc3933226bf998f07177b60871f
Reviewed-by: Morten Johan Sørvig <morten.sorvig@digia.com>
Reviewed-by: Richard Moe Gustavsen <richard.gustavsen@digia.com>
We're not in the output directory yet, so resolving using vanilla
fileFixify will end up failing when we check if the file exists, since
QFileInfo resolves relative paths against the current directory.
Change-Id: I414c6a2e83b49e3fb30e6153a49f7a90a8e528a0
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@digia.com>
The Xcode generator seems to have been written with the assumption that
writeMakeParts() would be called with the output directory as the current
directory, but that's not the case when shadow-building. Perhaps this
was changed in qmake at some point, and the Xcode generator was not
updated to reflect that.
Instead of replacing every occurance of fileFixify and other logic to
deal with paths, we just chdir into the output_dir for the duration
of the function (except when writing the 'make qmake' makefile, as
the regular makefile generator works as expected with the current
directory set to the input directory).
Change-Id: I6ba492036d73f29f4adbd7cd554db9504050629e
Reviewed-by: Richard Moe Gustavsen <richard.gustavsen@digia.com>
We already assume that if a source is buildable and should end up in OBJECTS we
can let Xcode build it, so we skip this input for the extra compiler.
Change-Id: I17b2408925b8e6513f0fa0d2459ec539bf7381d3
Reviewed-by: Richard Moe Gustavsen <richard.gustavsen@digia.com>
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@digia.com>
The PBXResourcesBuildPhase will optimize resources, such as turning XIB
files into NIB files, running pngcrush on images, turning string files
into binary plists, etc, so we prefer that if possible.
Unfortunatly this phase does not support custom paths, so whenever we
encounter bundle data with a custom path we fall back to the regular
PBXCopyFilesBuildPhase.
Change-Id: I539db03dd7982fd37293123b6428cdb695f64d2b
Reviewed-by: Richard Moe Gustavsen <richard.gustavsen@digia.com>
The Xcode generator creates a makefile for running 'make qmake', and the
makefile passes -o to ensure it writes to the same Xcode project. This
fails when qmake then treats -o foo.xcodeproj/project.xcproj as not only
setting the output filename, but also the output directory to foo.xcodeproj,
which results in the Xcode project trying to reference files relative
to this directory, such as '../main.cpp'.
Unfortunatly the output filename parsing happens too early for us to know
whether or not the generator is Xcode, so we just have to assume that
a certain combination of output filename and directories means we are
generating an Xcode project.
Change-Id: I0901d4db995f287c35cbbbd015683d5abda6d0f5
Reviewed-by: Richard Moe Gustavsen <richard.gustavsen@digia.com>
Replace them with dashes, like Xcode itself does.
Change-Id: I302425363a2eef13394025cd4a9e414048ce55ce
Reviewed-by: Morten Johan Sørvig <morten.sorvig@digia.com>
There was quite a bit of cruft left over from older Xcode version. We
now produce Xcode 3.2 compatible files, similar to what Xcode would do
when asked to upgrade one of our generated files. In particular:
- Removed refType
- Set more lastKnownFileTypes
- Renamed defaultConfigurationIsName to defaultConfigurationName
- Add runOnlyForDeploymentPostprocessing = 0 to build phases
- Don't put buildSettings directly into PBXNativeTarget
- Don't write productSettingsXML
- Don't write startupPath
- Don't write name when path is the exact same
- Write empty buildSetting lists as empty string
- Don't write empty PBXBuildFile settings
- Don't write generated/neede filenames for PBXShellScriptBuildPhase
- Use PBXFileReference instad of PBXFrameworkReference
- Prune deprecated buildSetting variables
- Remove deprecated PBXBuildStyle sections
- Resolve correct CC/CPLUSPLUS/LDPLUSPLUS
- Write IPHONEOS_DEPLOYMENT_TARGET
Change-Id: Ia2365c2623fe898878bd10636c3b85145c1cff04
Reviewed-by: Andy Shaw <andy.shaw@digia.com>
Reviewed-by: Morten Johan Sørvig <morten.sorvig@digia.com>
Instead of letting each qmake variable have its own auto-generated name
we try to group common variables into similar groups as used by the Xcode
templates provided by Apple.
We also prevent the same files from ending up multiple times in a group.
Change-Id: I73b13d6071bb7b3cd1501c422a99c60743221485
Reviewed-by: Andy Shaw <andy.shaw@digia.com>
Move the Carbon dependency to the Cocoa platform plugin instead, where
it's actually used.
CoreFoundation was not used by any plugins and could be removed
completely.
Change-Id: I1c825cdf94e2cc348ea13519b894fd868be0d14a
Reviewed-by: Morten Johan Sørvig <morten.sorvig@digia.com>
Added the following functions to QStringRef: toShort, toUShort, toInt,
toUInt, toLong, toULong, toLongLong, toULongLong, toFloat, and toDouble.
These functions use the corresponding functions found in QLocale.
Updated tst_qstringref.cpp to exercise the new functionality.
Change-Id: I38668a0cc7da0c101a62613fd16cb5a98286617f
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@digia.com>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Allows us to dynamically generate the command line option for iOS later,
and allows the user to override QMAKE_MACOSX_DEPLOYMENT_TARGET with the
expected effect on the command line options.
We unset PERL5LIB to ensure we get the system Perl libraries, since the
Mac OS 10.6 CI machine seems to have a broken XML::Parser::Expat from
macports/CPAN.
Change-Id: I04430c7b1daf9452d72f9a04a6b7f8d0d6926884
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@digia.com>
Change-Id: I09286388e9af7ec472b394be87204746f8ae22b1
Reviewed-by: James Turner <james.turner@kdab.com>
Reviewed-by: Gunnar Sletta <gunnar.sletta@digia.com>
This now includes support for OpenGL 4.3
Change-Id: I964284843dffe806280e7f67cde67f17e84dc6df
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
Reviewed-by: Sean Harmer <sean.harmer@kdab.com>
Must use ; as the path delimiter instead of :
Change-Id: I549e1652ef5bbae09c8fddec3e83ac9f52cec3a4
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@digia.com>
Before a buffer swap the new QEglFSHooks::waitForVSync method is
called which looks at QT_QPA_EGLFS_FORCEVSYNC and - if that is set
and non-null - calls ioctl with the FBIO_WAITFORVSYNC request on
the framebuffer device.
This is required on some embedded platforms where the driver does not
support VSYNC yet the Kernel provides a generic implementation.
I tested this using QML_RENDER_TIMING=1 which proofs that the frame
rate for an example of mine drops from >125fps to a straight ~60fps
with a few frames that take ~33ms (i.e. 30fps) as expected for VSYNC.
To prevent excessive open/close calls on the frame buffer device
per frame, the file descriptor is now cached. To keep the QEglFSHooks
interface as clean as possible this is done via a global static in
qeglfshooks_stub.cpp and initialized and freed in platformInit and
platformDestroy.
Change-Id: I4d31b227c65ff22aa089db0fbc62c89a59cbb6c7
Reviewed-by: Sean Harmer <sean.harmer@kdab.com>
Reviewed-by: Samuel Rødal <samuel.rodal@digia.com>
These will be needed by the upcoming OpenGL enablers so move them
out of qopenglfunctions.h to somewhere that any opengl related file can
access them.
Change-Id: I0c788559397d446ec7210e2ad940da862179710d
Reviewed-by: Samuel Rødal <samuel.rodal@digia.com>
The motivation for this patch is twofold:
1: we need a way (for iOS/Android) to tell the current window to remove
focus from the focus object when the user hides the input panel.
Otherwise, if the focus object is e.g a line edit, the cursor will
continue to blink inside it, which is wrong. As it stands, telling
the active window to deactivate
(by calling QWindowSystemInterface::handleWindowActivated(0)), will cause
the whole application to deactivate if no windows are active, which
is not what we want.
2: Qt currently understands just two application states, Activated and
Deactivated. On mobile platforms we can have other states
as well, like "suspended" on iOS. So controlling the application
state should not depend on window activation, but instead be controlled
through a separate API by the platform plugin.
This patch will add the following function:
QWindowSystemInterface::handleApplicationStateChanged(Qt::ApplicationState newState)
that lets us control the application state from the plugin. This also
means that we factor out application state handling from window
activation, which also gives us a way to remove focus from a window while
keeping the application active.
To not break existing desktop platforms that relies on application
activation being tied to window activation, we need to make this API
opt-in by using a platform integration capability hint. This is not optimal, but
found necessary after investigating several other solutions.
Which states (other that active/inactive) it makes sense
to add to Qt::ApplicationState will be a topic for later patches.
Change-Id: Ic6fdd3b66867abb67da43eba04ec86f06d82ff94
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@digia.com>
Reviewed-by: Samuel Rødal <samuel.rodal@digia.com>
This fixes QMetaType detection of const reference arguments in signals
while connecting using the new syntax and Qt::QueuedConnection
const references should have the same QMetaType as non references.
That means we need to remove the const reference while getting the
QMetaType.
Change-Id: I9b2688da7fb9ae985aec0d8fa62a1165357ffe71
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Stephen Kelly <stephen.kelly@kdab.com>
Reviewed-by: Jędrzej Nowacki <jedrzej.nowacki@digia.com>
Now that we always build against an SDK, we can be sure that the
function declaration for CTFontCopyDefaultCascadeListForLanguages
is available in the CoreText CTFont.h header.
Change-Id: I304a701548833e5c7774b4fd2e72eb8c541dd103
Reviewed-by: Jens Bache-Wiig <jens.bache-wiig@digia.com>
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@digia.com>
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@digia.com>
When cross compiling between two systems that use different
program file formats (such as PECOFF vs ELF) -x cannot be
relied upon, so -f should be used instead.
Change-Id: I4c80646b2b1fbdd6072589d4d9e852adda83424a
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@digia.com>
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@digia.com>
This is because on some systems, /usr/bin/perl is too old to
run syncqt. MSYS is one such system.
Change-Id: I3e1aa2db557cc3919618350775e7218ae05aa93f
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@digia.com>
This change lets you call QHostInfo::lookupHost() with a null receiver
in order to warm up the DNS cache. This allows you to try to get the
DNS request in flight early.
Change-Id: Icfdd28146479aa534ae9ceb472f75e08aaa39cd2
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
It was not possible to extract data from the archive on OSX
which was created on Windows platform because of wrong
separators. Archive was created on Windows via QZipWriter and
opened on OSX with QZipReader. It consisted of a lots directories
and subdirectories with files. The solution is to use '/' separator
for internal representation.
Change-Id: Ic0837ca184bb6188129d53b587a5df2ec61e4e05
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@digia.com>
Activate the window of the widget under mouse pointer before replay mouse
press event.
Change-Id: I9e699374accf108aa49b2a3c73d5e76631100dfd
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@digia.com>
Reviewed-by: Samuel Rødal <samuel.rodal@digia.com>
Reviewed-by: Gunnar Sletta <gunnar.sletta@digia.com>
When the cursor specified is a bitmap one we cannot cache it based on
the shape as the pixmap set on the cursor may be different. Therefore
we should always create a new cursor in this instance.
Change-Id: I2c201590ff632490d76c1b423908ae32aa584eb6
Reviewed-by: Morten Johan Sørvig <morten.sorvig@digia.com>
When setFloating(true) is called before show, frame strut events
are not enabled for the native window (since there is none yet)
in QDockWidgetPrivate::setWindowState(). In that case, do
it in the show event handling.
Task-number: QTBUG-29012
Change-Id: I93b679f20200c149d608a1bcc65b4936a035c6a0
Reviewed-by: J-P Nurmi <jpnurmi@digia.com>
When QProcess->start() is called, Qt creates a pipe to the process to
get its exit value and output.
It does this with qt_create_pipe, which calls qt_safe_pipe.
qt_safe_pipe, on failure, returns 1. qt_create_pipe then
sets errno and returns void.
The calling function, QProcessPrivate::startProcess, does not check
errno, and thus continues to fork the process, assuming the pipe has been
created successfully.
The child process then has no way to pass its exit value to the calling
process, since the communication pipes it would normally use do not
exist, and thus when it exits it becomes a zombie.
As a bonus, if waitForFinished is called on a broken process, a crash
results because it is trying to wait on a pipe which does not exist.
The fix makes qt_create_pipe return an integer, and QProcess::startProcess
check the return value, set processError and not create the child
process.
Task-Number: QTBUG-18934
Change-Id: I2e1effdd0617be5b8c5492bcbcf5f2b1584b2241
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
The QAbstractSocket API has been already using this as a return type. Hence,
this has already been exposed to the public API users, anyhow.
http://qt-project.org/doc/qt-5.0/qtnetwork/qabstractsocket.html#socketDescriptor
A minor mistake has also been fixed in this commit at the quintptr section.
Change-Id: I8143b3050428548ff6baee2e3a0bce4058ea8701
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>