| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
| |
Reviewer: Renato Araújo <renato.filho@openbossa.org>
Luciano Wolf <luciano.wolf@openbossa.org>
|
|
|
|
|
| |
Reviewer: Renato Araújo <renato.filho@openbossa.org>
Luciano Wolf <luciano.wolf@openbossa.org>
|
| |
|
|
|
|
|
| |
Reviewer: Luciano Wolf <luciano.wolf@openbossa.org>
Marcelo Lira <marcelo.lira@openbossa.org>
|
|
|
|
|
| |
Reviewer: Hugo Parente Lima <hugo.pl@gmail.com>
Luciano Wolf <luciano.wolf@openbossa.org>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Instead of the awkward:
<enum-type name="Foo::Bar"/>
<value-type name="Foo"/>
<value-type name="Foo::Sbrubles"/>
We now use the proper:
<value-type name="Foo">
<enum-type name="Bar"/>
<value-type name="Sbrubles"/>
</value-type>
Reviewed by Luciano Wolf <luciano.wolf@openbossa.org>
Reviewed by Renato Araújo <renato.filho@openbossa.org>
|
|
|
|
|
| |
Reviewer: Renato Araújo <renato.filho@openbossa.org>
Luciano Wolf <luciano.wolf@openbossa.org>
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
| |
Now returning true *only* when exactly all overloads
accept exactly 1 argument.
Fixes compilation of QAudioInput and QAudioOutput from
QtMultimedia. Both have start() and start(QIODevice*) overloads,
and the previous code was generating 'arg' instead of 'pyargs'.
Reviewer: Renato Araújo <renato.filho@openbossa.org>
Luciano Wolf <luciano.wolf@openbossa.org>
|
|
|
|
|
| |
Reviewer: Luciano Wolf <luciano.wolf@openbossa.org>
Lauro Moura <lauro.filho@openbossa.org>
|
|
|
|
|
| |
Reviewers: Renato Araújo <renato.filho@openbossa.org>
Hugo Parente Lima <hugo.pl@gmail.com>
|
|
|
|
|
|
|
| |
files to allow multiples version installed simultaneous.
Reviewer: Hugo Parente Lima <hugo.pl@gmail.com>
Luciano Wolf <luciano.wolf@openbossa.org>
|
|
|
|
|
| |
Reviewer: Luciano Wolf <luciano.wolf@openbossa.org>
Hugo Parente Lima <hugo.pl@gmail.com>
|
|
|
|
|
| |
Reviewer: Luciano Wolf <luciano.wolf@openbossa.org>
Hugo Parente Lima <hugo.pl@gmail.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The new algorithm do the following:
- Try to use type_info on the object the get the object real name.
- Try to find a type resolver with the name returned by type_info.
- If a type resolver was found, get the python type.
- Else, ask binding manager to resolve the type walking on all possible
subclasses found in the inheritance tree.
The binding manager has a graph representing the class inheritance tree.
Note: This commit break the libshiboken ABI, but not the API.
Reviewer: Luciano Wolf <luciano.wolf@openbossa.org>
Renato Araújo <renato.araujo@openbossa.org>
|
|
|
|
|
| |
Reviewer: Luciano Wolf <luciano.wolf@openbossa.org>
Renato Araújo <renato.filho@openbossa.org>
|
|
|
|
|
| |
Reviewer: Hugo Parente Lima <hugo.pl@gmail.com>
Luciano Wolf <luciano.wolf@openbossa.org>
|
|
|
|
|
| |
Reviewer: Hugo Parente Lima <hugo.pl@gmail.com>
Luciano Wolf <luciano.wolf@openbossa.org>
|
|
|
|
|
| |
Reviewer: Hugo Parente Lima <hugo.pl@gmail.com>
Luciano Wolf <luciano.wolf@openbossa.org>
|
|
|
|
|
| |
Reviewer: Marcelo Lira <marcelo.lira@openbossa.org>
Luciano Wolf <luciano.wolf@openbossa.org>
|
|
|
|
|
|
|
|
|
| |
Similar to 83c272, avoid using assertNotEqual which in Python 2.7 uses
the != operator; that operator however is intentionally left undefined
in Point class.
Reviewed by Marcelo Lira <marcelo.lira@openbossa.org>
Reviewed by Renato Araujo <renato.filho@openbossa.org>
|
|
|
|
|
|
|
|
| |
This is due to Point intentionally not having a != operator and
assertNotEqual implementation using the __ne__ operator on Python 2.7.
Reviewed by Luciano Wolf <luciano.wolf@openbossa.org>
Reviewed by Renato Araújo <renato.filho@openbossa.org>
|
|
|
|
|
|
|
|
| |
The said method retuns a void pointer never before seen on the Python
side. The code injection puts the void pointer in a BaseWrapper object.
Reviewed by Luciano Wolf <luciano.wolf@openbossa.org>
Reviewed by Renato Araújo <renato.filho@openbossa.org>
|
|
|
|
|
| |
The test stores a native Python object as a void pointer in C++ and
takes it back.
|
|
|
|
|
| |
Reviewer: Marcelo Lira <marcelo.lira@openbossa.org>
Luciano Wolf <luciano.wolf@openbossa.org>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The assertion is only used in debug mode. The duck punching test
was altered to avoid failing when run with debug. Check the
duck punching test file for a better explanation.
Also added a debug helper method "showWrapperMap" to the binding
manager, it shows the contents of the C pointer to Python wrapper
mapping.
Reviewed by Luciano Wolf <luciano.wolf@openbossa.org>
Reviewed by Renato Araújo <renato.filho@openbossa.org>
|
|
|
|
|
|
|
|
| |
Fix tests.
Created warnning about return last ref of python object.
Reviewer: Marcelo Lira <marcelo.lira@openbossa.org>
Luciano Wolf <luciano.wolf@openbossa.org>
|
|
|
|
|
| |
Reviewer: Marcelo Lira <marcelo.lira@openbossa.org>
Luciano Wolf <luciano.wolf@openbossa.org>
|
|
|
|
|
| |
Reviewer: Marcelo Lira <marcelo.lira@openbossa.org>
Luciano Wolf <luciano.wolf@openbossa.org>
|
|
|
|
|
|
|
|
|
| |
The decisor must check for the number of arguments before calling
Converter::isConvertible method, to avoid passing NULL pointers to it.
Unit tests were added.
Reviewed by Luciano Wolf <luciano.wolf@openbossa.org>
Reviewed by Renato Araújo <renato.filho@openbossa.org>
|
|
|
|
|
| |
Reviewer: Marcelo Lira <marcelo.lira@openbossa.org>
Luciano Wolf <luciano.wolf@openbossa.org>
|
|
|
|
|
| |
Reviewer: Marcelo Lira <marcelo.lira@openbossa.org>
Luciano Wolf <luciano.wolf@openbossa.org>
|
|
|
|
|
| |
Reviewed by Hugo Parente <hugo.lima@openbossa.org>
Reviewed by Renato Araújo <renato.filho@openbossa.org>
|
|
|
|
|
|
|
|
| |
calls
Author: Didier Raboud <didier@raboud.com>
Reviewers: Luciano Wolf <luciano.wolf@openbossa.org>,
Marcelo Lira <marcelo.lira@openbossa.org>
|
|
|
|
|
| |
Reviewers: Luciano Wolf <luciano.wolf@openbossa.org>,
Marcelo Lira <marcelo.lira@openbossa.org>
|
|
|
|
|
| |
Reviewer: Hugo Parente Lima <hugo.lima@openbossa.org>,
Marcelo Lira <marcelo.lira@openbossa.org>
|
|
|
|
| |
We can't run they standalone anyway, because they need some env vars to be exported like PYTHONPATH.
|
| |
|
| |
|
|
|
|
|
|
| |
This method is needed because you can not call Converter<T>::isConvertible inside
a Converter<Y>::isConvertible implementation, otherwise it'll create a 2-step
implicit conversion.
|
|
|
|
| |
Tests were also added.
|
|
|
|
|
|
| |
This fix is specific for the cases when the "protected hack" is
turned off.
Also added some tests.
|
| |
|
| |
|
|
|
|
|
| |
Reviewer: Luciano Wolf <luciano.wolf@openbossa.org>
Reviewer: Renato Araújo <renato.araujo@openbossa.org>
|
| |
|
|\ |
|
| | |
|
| |
| |
| |
| | |
This fix some problems compiling bindings for classes under C++ namespaces.
|
| | |
|