The handling of QSslOptions is complicated not only by the subject, but
also by the fact that some of the openssl directives are negatives. This
commit tries to separate the inherent complexity from the complexity of
the api by allowing us to test them independently.
Change-Id: Ieb9386c69dd9b0b49dc42e26b2878a301f26ded1
Reviewed-by: Peter Hartmann <peter.hartmann@nokia.com>
GL headers should not be included inside of the Qt namespace.
Change-Id: Ib19cbd3ab7b071a96e0816798f8ebc3fd4e28448
Reviewed-by: Toby Tomkins <toby.tomkins@nokia.com>
Reviewed-by: hjk <qthjk@ovi.com>
The removed code was commented-out before the tests were imported into
the Qt repository in June 2006. This code is no longer valid as the
warning messages it relates to are not produced by any code in qtbase.
Change-Id: Iac63b27387b51a71c567386ca61faf5fb00c3364
Reviewed-by: Rohan McGovern <rohan.mcgovern@nokia.com>
The availableWithoutDebug() test was trying to force QT_NO_DEBUG to be
defined, but was actually doing the opposite. While I'm not satisfied
with the way this test works (it would be better to test with
QT_NO_DEBUG already defined when Q_FUNC_INFO is included). at least this
test now does what it says it does.
Additionally, the test has been improved by checking that Q_FUNC_INFO
returns a non-empty string, as opposed to just compiling.
Change-Id: I9175e3a7406152b6250b507a61b5062e3cfabeb8
Reviewed-by: Rohan McGovern <rohan.mcgovern@nokia.com>
Previously, autotests and benchmarks were disabled on Mac OS X. This is
wrong, and needs to be corrected. The 'make check' target is disabled
for the autotests on Mac OS X for now, but all tests will now be built.
The check target does not actually run the benchmarks, so this can be
safely re-enabled without breaking the CI system.
There will be commits to follow which re-enable the 'make check' target
on Mac OS X for each module.
Change-Id: I1347da46e3688b48e8cc5d933ee5188ea3723639
Reviewed-by: Rohan McGovern <rohan.mcgovern@nokia.com>
The existing code doesn't work on 64bit machines. We must first convert
the list into a QVector, which guarantees that elements are laid out
correctly as an array.
Merge-request: 1467
Reviewed-by: thiago
(cherry picked from commit 00020eed3fa948f69cfa776e92121edec6f975cc)
Change-Id: I00020eed3fa948f69cfa776e92121edec6f975cc
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
This feature is only available to advanced qmake clients who have their
own defaults_pre.prf, where QMAKE_SYNCQT_OUTDIR is set before loading
Qt's own default_pre.
This is useful to be able to put the sync.profile file in a different
level in the source tree than where you want the output in the build
tree.
An alternate/complimenting solution would be to allow the output path
to be overriden from within the sync.profile file.
Change-Id: I3cdef82dbe492877507e269e18571456beeb2b7d
Reviewed-by: Marius Storm-Olsen <marius.storm-olsen@nokia.com>
This reverts commit f109f5b8a9.
Instead of disabling the asserts, rather enable the debug code
so the asserts actually trigger.
Change-Id: I6f62588d836de5c926294241e8d7196a963e4aac
Reviewed-by: Eckhart Koppen <eckhart.koppen@nokia.com>
"qsslconfiguration.cpp:204:73: warning: suggest parentheses around
comparison in operand of ‘|’ [-Wparentheses]"
Change-Id: I887ffdf3ef8263c35a8f391b3fc97faee41b7dab
Reviewed-by: Martin Petersson <Martin.Petersson@nokia.com>
This removes a change to libpng that was only relevant for building
a certain Qt 4 tool in certain environements without system libpng.
This is no longer relevant, so this kludge should now be removed.
Change-Id: I9a04ab0224f9e84ccab3ec27762582be439fe3ef
Reviewed-by: Kim M. Kalland <kim.kalland@nokia.com>
This is a source-incompatible change.
TlsV1 is ambiguous; what is actually meant is TLS version 1.0. There are
also TLS versions 1.1 and 1.2; we might want to add options for these
once OpenSSL supports them (apparently they will be with OpenSSL version
1.0.1).
Change-Id: I940d020b181b5fa528788ef0c3c47e8ef873796a
Reviewed-by: Lars Knoll <lars.knoll@nokia.com>
If subdir.CONFIG contains no_check_target, we do not recurse into
that subdir when generating the 'check' target.
This will be used to selectively disable crashing tests on Mac OS X.
Currently, all autotests are disabled on Mac OS X. The goal is to re-
enable them, but not all at once. This approach allows us to get each
change through the CI system individually.
The first step is get all tests to build, but not run, then we will
re-enable the tests for each module, possibly disabling individual
tests at the same time.
Change-Id: I69f62c238f381ae0315d414cd71d76bc88b088dd
Reviewed-by: Bradley T. Hughes <bradley.hughes@nokia.com>
Use QAtomic*::load() and ::store() instead of the deprecated
cast, assignment, and comparison operators. These will be removed
in the near future.
The tests for these particular operators have not been changed,
though, as the change to remove the operators will also remove
the respective tests.
Change-Id: I2f24d18992af0c6e0f487d707218e4e84f4bdd12
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@nokia.com>
Currently QTest::newRow function expects to have correct tag string for
each test data. This was enforced by a simple assert. By adding a
an assert with message we might simplify debugging of a test that
generates test cases.
Change-Id: I84ebb397264c99e7bb46d281fd38583dd2d6964c
Reviewed-by: Jason McDonald <jason.mcdonald@nokia.com>
Certain Q_ASSERT usages are not compiling with -force-asserts since they
are checking for instance variables or other items which are not
available in release mode.
This fix guards the asserts to only be enabled in debug mode.
Change-Id: I2ed74e3d5118608361e8c924e93d8ff0efdc8eab
Reviewed-by: Lars Knoll <lars.knoll@nokia.com>
maybe if your trying to instansiate a QWindow before the QApplication has
been instansiated.
Change-Id: I68ed4ad1dbdfedc242a7cb27c1381d50a3d3dfa4
Reviewed-by: Samuel Rødal <samuel.rodal@nokia.com>
Move the QDockWidget close and float buttons from the right end of the
dock widget title bar to the left end to avoid Mac App Store rejection.
Task-number: QTBUG-19493
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@nokia.com>
(cherry picked from commit 3b2bd144b7873daf2e7119248dbac7c0f0253993)
Change-Id: I3b2bd144b7873daf2e7119248dbac7c0f0253993
The animations are now deleted in destruction of the Vista style.
With the previous implementation the animations were not deleted for
instance if you move mouse cursor off from a widget with hover
animation (like QPushButton).
Task-number: QTBUG-21532
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@nokia.com>
(cherry picked from commit 261bbb12003dab0e45b5814f85dd74aa64bcfb79)
Change-Id: I261bbb12003dab0e45b5814f85dd74aa64bcfb79
So that it won't just return "Helvetica" that may not exist in
target system. Providing it seems to fix QTBUG-21833 as well,
thus we can remove the QEXPECT_FAIL statement in qlistview test.
'_' appears to reach the minimum right bearing in some fonts.
Task-number: QTBUG-21833
Change-Id: I2340af6ee736083ec4f6575e469840ebec4e545e
Reviewed-by: Jiang Jiang <jiang.jiang@nokia.com>
Moving these classes to an unnamed namespace reduced chances of
a name conflict and size of executable. GCC 4.6 in release mode produces
a 5kB smaller binary.
Change-Id: Ie1819a88180f9db1cd32cc59ea6fe268ecbd5813
Reviewed-by: Olivier Goffart <ogoffart@woboq.com>
When a GL context group is destroyed, all multi-group shared
resources associated with the group should be cleaned up.
Otherwise we could get a double deletion in the resource's
destructor, because it still retained a pointer to the
deleted group.
The missing cleanup resulted in a crash when the global
static qt_gl_functions_resource was destroyed, first seen
in the tst_examples autotest in qtdeclarative. It possibly
didn't manifest before because it's event loop-dependent
(the contexts are deleted via deleteLater()).
Change-Id: I6b1e0bfdfbbb2bff8e795f545e680fcdfa094768
Reviewed-by: Samuel Rødal <samuel.rodal@nokia.com>
Forgot to add a Windows-specific source file to the project
when the new V8 version was imported.
Change-Id: Id20dcba132bf2609338845e8a769fcc8c1cbc562
Reviewed-by: Aaron Kennedy <aaron.kennedy@nokia.com>
This commit adds the ability to perform legacy SSL renegotiation as
a fallback via QSsl::SslOptions. This is something that used to work,
but has been disabled by default in newer versions of openssl. The
need for this has been reported by users (eg. in QTBUG-14983).
Change-Id: I5b80f3ffd07e0c5faddc469f6a8f857bac5740f7
Reviewed-by: Corentin Chary <corentin.chary@gmail.com>
Reviewed-by: Peter Hartmann <peter.hartmann@nokia.com>
This code is totally outdated, and has never been used in Qt4 AFAIK
Change-Id: I775fe87532807e77fc94fe661e4b659c11bfd8be
Reviewed-by: Bradley T. Hughes <bradley.hughes@nokia.com>
The moc generated code would not compile otherwise
Keep Moc::parseFunction and Moc::parseMaybeFunction in sync
(the first is used for signals and slots, and the second for normal
functions such as Q_INVOKABLE)
Last patch that introduced function pointer updated parseFunction
but not parseMaybeFunction
When a slot return a reference, moc generate code that make the
MetaObject system think it is a void, so qt_metacall and invokeMethod
do not mess with the return value.
But when we want to take the function signature, in the IndexOfMethod
call, we need to have the exact return type.
Change-Id: I4661218d7ce367ad3934e73929e7d04f0a6dbc09
Reviewed-by: Bradley T. Hughes <bradley.hughes@nokia.com>
In Qt 4.x the serial number is reported by a mixture of the hex value
and the number, The hex is what is used by other tools, and we should do
the same.
Change-Id: Ia0361d43fb5b920d053c95e932e0c8a012436e5e
Reviewed-by: Peter Hartmann <peter.hartmann@nokia.com>
This caused a "deadlock" in tst_qobject::moveToThread
The problem was that QEventLoop::quit was called from the first
prcessed event in that loop, put calling process event a second time
deadlock because there is no more event to process
(In practice, some event can come from the window manager or the
inputs, but they may as well not come)
Change-Id: Ia469110eb9c9de57669e80cf19e933f410e469a4
Reviewed-by: Bradley T. Hughes <bradley.hughes@nokia.com>
QAbstractEventDispatcher() does no longer install itself into the current
thread. Instead the new methods QThread::setEventDispatcher() and
QCoreApplication::setEventDispatcher() allow to install a custom event
dispatcher into any QThread as long as there is no default event dispatcher
created yet. That is, before the thread has been started with
QThread::start() or, in case of the main thread, before QCoreApplication has
been instantiated.
Change-Id: I7367e13d8d8aebed5a5651260bb69b8818eb1b90
Reviewed-by: Olivier Goffart <ogoffart@woboq.com>
Reviewed-by: Bradley T. Hughes <bradley.hughes@nokia.com>
privateqt_mac_p.h lives in QtWidgets/private/qt_mac_p.h now instead of
QtGui/private/...
Change-Id: I77e57de5581829e78919b13b10f7a896879670a7
Reviewed-by: Rohan McGovern <rohan.mcgovern@nokia.com>
To instrument a Qt application or library with the TestCocoon coverage
tool, do `CONFIG+=testcocoon' in the application .pro file.
To instrument Qt itself with testcocoon, use the `-testcocoon' configure
option.
Change-Id: Ie77109a078d11ea51f7a073621e0df9c752c44ae
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@nokia.com>
Reviewed-by: Rohan McGovern <rohan.mcgovern@nokia.com>
The QTest::qWait() function does not handle DeferredDelete, in this case
all deleteLater() calls before the qWait() won't be processed during the
wait. Add an explicit sendPostedEvents(0, QEvent::DeferredDelete) to fix
this problem.
Task-number:QTBUG-22633
Change-Id: I2c5574b4e7d3c070fad6382dc36ddfd1dfac50d6
Reviewed-by: Jason McDonald <jason.mcdonald@nokia.com>
When combined with http://codereview.qt-project.org/8644, the test
crashes. This appears to be due to a column of the view being deleted
in response to calling setCurrentIndex(), as documented in QTBUG-22707.
This commit must be merged at the same time as
http://codereview.qt-project.org/8644.
Task-number: QTBUG-22707
Change-Id: I075be7df5d9196e37e2286e61d317086a76a167d
Reviewed-by: Rohan McGovern <rohan.mcgovern@nokia.com>
QPlatformFontDatabase::populateFontDatabase() routine uses a different
usrPtr format which is not compatible with the one basic font db uses
(FontFile *). We shouldn't mix them together.
Task-number: QTBUG-22713
Change-Id: I239ea085e14261f6b9ba66b03710a590a44378f9
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@nokia.com>
The received window position cannot be trusted blindly, e.g. in case
of resizing a window via the mouse or calling showMaximized() the
position is bogus (for our purposes). Instead, it needs to be queried.
Before 37f338e5ed an incomplete
workaround was in place however it got removed in that commit,
resulting in weird off-by-a-certain-amount mouse position issues in
certain use cases. This patch aims to fix the issue similarly to how
it was done in Qt4.
Change-Id: I2d2a69b81a2782117b700fe366fae4c102aca1f4
Reviewed-by: Samuel Rødal <samuel.rodal@nokia.com>
Actually use the calculated size for needed space instead of just
ignoring the return value.
Task-number: QTBUG-16315
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@nokia.com>
(cherry picked from commit 0792215fa8d227617a2080af9f12fd7f486b87d6)
Change-Id: I0792215fa8d227617a2080af9f12fd7f486b87d6
The QMacStyle is not built as part of QtWidgets at the moment (see the
TODO in src/widgets/styles/styles.pri), so disable the test for now.
I've included a comment in styles.pri to re-enable the test once the
style is done.
Change-Id: Ia6ea130135586f107ba257bf2c6c2c7187a830b6
Reviewed-by: João Abecasis <joao.abecasis@nokia.com>
Otherwise the symbols used by the test aren't found when linking.
Change-Id: Ibc78ac76763d0558824b797582e16f97786c5840
Reviewed-by: João Abecasis <joao.abecasis@nokia.com>
Function pointers used as template parameters must have external linkage
according to the C++ standard. The _mm_lddqu_si128() and
_mm_load_si128() intrinsic functions are declared static in most
implementations, which means we need to wrap them in a non-static
function call to be used.
Change-Id: I4e316851df890d77999649d47654f7f9a35d4080
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
The QMAKE_COMPILER_DEFINES are used when running moc. In some cases, Qt
checks the __GNUC__ value to determine which version of GCC is being
used. In particular, the qtconcurrentfilter and qtconcurrentmap tests
fail to compile because moc ends up defining QT_NO_CONCURRENT_FILTER and
QT_NO_CONCURRENT_MAP since the condition __GNUC__ < 4 evaluates to true.
The Mac OS X development tools have not used gcc 3.x is a very long
time. The current version ships with llvm-gcc-4.2.1 and clang (which
also masquerades as gcc 4.2.1). Defining __GNUC__=4 makes sure that moc
preprocesses headers in the same way the compiler does.
Change-Id: Ie491d3f5b074fd1a509ffe6755f3308836817b6c
Reviewed-by: João Abecasis <joao.abecasis@nokia.com>
We look for a regular binary, but Mac OS X produces application bundles
by default. Instead of looking for an app bundle as well, we unify things
a bit by disabling app bundles instead.
Change-Id: I5eb173d98e893e616f23b91f13282cf95ef4f56a
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@nokia.com>
