| Commit message (Collapse) | Author | Age | Files | Lines |
|---|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The PySide project has been split into three pieces, including
Shiboken. This had far-reaching consequences for the signature project.
Shiboken can be run together with PySide or alone,
with tests or without. In every configuration, the signature
module has to work correctly.
During tests, the shiboken binary also hides the shiboken module,
and we had to use extra efforts to always guarantee the accessibility
of all signature modules.
This commit is the preparation for typeerrors implemented with the
signature module. It has been split off because the splitting
is not directly related, besides these unawaited consequences.
I re-added and corrected voidptr_test and simplified the calls.
Remark.. We should rename shiboken to Shiboken in all imports.
I also simplified initialization. After "from PySide2 import QtCore",
now a simple access like "type.__signature__" triggers initialization.
Further, I removed all traces of "signature_loader" and allowed
loading everything from PySide2.support.signature, again. The
loader is now needed internally, only.
Also, moved the type patching into FinishSignatureInitialization
to support modules with no classes at all.
The "testbinding" problem was finally identified as a name clash
when the same function is also a signal. A further investigation
showed that there exists also a regular PySide method with
that problem. The test was extended to all methods, and it
maps now all these cases to "{name}.overload".
Updated the included typing27.py from https://pypi.org/project/typing/
from version 3.6.2 to version 3.6.6 .
Task-number: PYSIDE-749
Change-Id: Ie33b8c6b0df5640212f8991539088593a041a05c
Reviewed-by: Cristian Maureira-Fredes <cristian.maureira-fredes@qt.io>
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
| |
Once the qtbase commit
(22c1a46a03bc3347afc0e7462e19558283d0e1b7) gets integrated into
the qt/qt5 repo for dev branch, we can revert the terrible
workaround fix, because it won't be needed anymore.
This reverts commit 5cd52cda24bf70bf99f4deec037039df0ab928f5.
Task-number: PYSIDE-724
Change-Id: I627c7ec945b864a1c16ba6cd7288807591f70140
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Starting with Qt 5.12-ish (qtbase commit
c538a333db4b7526fb4d9a82c06296d2492bf000) the macOS QPA cocoa plugin
switched to using layer-backed NSViews.
For some unknown reason when building PySide2
with a python.org official interpreter and the above commit, none
of the QWidget / NSViews get drawn (-[QNSView drawRect:] is not
called).
Using an older qtbase with the same official interepreter works.
Using the same qtbase with a custom built interpreter works.
Using the same qtbase with a custom built interpreter which is the
same version as the official one (3.6.5 at the moment) works
(built from v3.6.5 source package published on the downloads page).
Using the same qtbase with official interpeter does NOT work.
Python2 vs Python3 does not matter.
Release vs Debug does not matter.
I have not been able to build an interpreter with the minimum
deployment target set to 10.6 as the official package does (but
rather 10.12), so this is still an avenue to try.
The terrible workaround is to disable layer-backed views via an
environment variable when running testrunner (QT_MAC_WANTS_LAYER).
All users would have to do the same for their own applications,
otherwise they will not work.
Change-Id: Ia9a34f1e7b0b1b4807030cc7acecbefc8f344e84
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This change introduces tracking of modules for which bindings will not
be built. This allows us to add conditions on disabled modules,
and thus disable certain tests.
Thus we disable pysidetest test when its dependencies are not met.
Due to this, we can now builtd bindings only for QtCore,
allowing faster development iteration when touching only QtCore.
Note that this only affects which module bindings will be created.
Shiboken itself still requires QtCore, QtXml and QtXmlPatterns,
as does pyside2-rcc.
Change-Id: I75084a1741e7f4c3594e43af0fd9668a0e969c56
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
|
| |
|
|
|
|
|
|
|
|
|
|
| |
Setting owner as default to not allow Python to create a copy
of the QPixmap associated with the QLabel.
The C++ object pointer is acquired through the pixmap() method.
A test case was included.
Task-number: PYSIDE-150
Change-Id: Ie6975c39cbf49a59ebd478db0e1a0c30fc14864a
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
When Qt launches the QXcbEventReader thread, by default the created
thread dispatcher will call g_thread_init_glib to initialize the glib
thread. When libqgtk2 plugin is loaded, the plugin calls gtk_init
which also needs to initialize the glib thread library.
This can cause a race condition where the xcb thread might not finish
initializing all of the glib thread library, but the main thread
believes that initializing is done, and thus ends up dereferencing
null pointers.
Specifically when the glib function g_slice_alloc is called in the main
thread, which calls allocator_categorize, the glib initialization flag
'sys_page_size' is checked. This flag can already be set by the call
to g_slice_init_nomessage in the xcb thread, but magazine_mutex might not
yet be allocated (in g_slice_thread_init_nomessage), and the main
thread ends up dereferencing a null pointer mutex.
Relevant code can be found at
https://sourcecodebrowser.com/glib2.0/2.27.4/gslice_8c_source.html
The workaround is to set the QT_NO_GLIB environment variable
to 1 when running the tests, so that a regular
QEventDispatcherUNIX is used. Thus only the gtk plugin will call the
glib initialization function, eliminating the race condition.
Note that the issue probably happens only for glib versions < 2.32.
The g_thread_init function is deprecated since 2.32, and glib thread
initialization is done at the start of the program, as referenced at
https://developer.gnome.org/glib/2.32/glib-Deprecated-Thread-APIs.html#g-thread-init
Task-number: QTBUG-64716
Change-Id: Ibcccf8f6e0a3299e61dd320eb6d08e29658298e2
Reviewed-by: Christian Tismer <tismer@stackless.com>
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
With the signature module, it is now a straight forward task
to generate a registry of all known function signatures.
We check that these signatures all exist.
One file contains all signatures for one platform and version.
The test is only activated when run in the CI system.
An initial call creates the expected file as output and raises
an error. The result can then be picked up from the error log
and added to the repository.
Done: linux2 5.6.4
Done: darwin 5.6.4
Done: win32 5.6.4
Task-number: PYSIDE-510
Change-Id: I4f406cf72d25fdd2336814f6f20129079b8be54f
Reviewed-by: Christian Tismer <tismer@stackless.com>
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This is achieved by registering a qt.conf file with a Prefix pointing
to a directory relative to the loaded PySide2 module (e.g. QtCore).
Thus Qt does not crash due to not finding platform plugins.
Because this change would affect tests, which are ran before the
PySide package is installed, a new environment variable called
PYSIDE_DISABLE_INTERNAL_QT_CONF is introduced. This variable disables
the registration of the internal qt.conf file, thus it will not point
to a not yet created location, which will allow tests to run as
before.
Change-Id: I5a96037adfafe1f08ea57535aa4a2a0d1660dfaf
Task-number: PYSIDE-558
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
Reviewed-by: Christian Tismer <tismer@stackless.com>
|
|
|
in preparation for a subtree merge.
this should not be necessary to do in a separate commit, but git is a
tad stupid about following history correctly without it.
|
