On Thu, Apr 17, 2014 at 02:33:15PM -0500, Albert Chin wrote: > On Thu, Apr 17, 2014 at 03:07:00PM -0400, John Ehresman wrote: > > On 4/17/14, 2:58 PM, Albert Chin wrote: > > >> I suspect this means the "typedef unsigned int uint" isn't being > > >> seen when shiboken scans the header files during generation on AIX > > >> while it is seen on the other platforms. I don't think the uint > > >> type is special cased inside of shiboken. > > > > > > Correct. The problem is that on AIX we have: > > > typedef unsigned int uint_t; > > > typedef uint_t uint; > > > > The question is why you don't get an alias from uint to uint_t and then > > another one from uint_t to unsigned int. I don't know the answer. You > > might try putting together a minimal test case and then seeing what the > > difference is in it. > > We do get an alias from uint to uint_t. But, oddly, not from uint_t to > unsigned int. At least, that's what I've gathered from my debugging so > far.
Testing the attached patch. Seems shiboken was ok all along. -- albert chin (ch...@thewrittenword.com)
Index: PySide/QtCore/typesystem_core_common.xml =================================================================== --- PySide/QtCore/typesystem_core_common.xml.orig 2013-08-12 21:24:05.000000000 +0000 +++ PySide/QtCore/typesystem_core_common.xml 2014-04-18 02:59:01.346768914 +0000 @@ -152,6 +152,7 @@ <primitive-type name="int"/> <primitive-type name="signed int"/> <primitive-type name="uint"/> + <primitive-type name="uint_t"/> <primitive-type name="ulong"/> <primitive-type name="unsigned int"/> <primitive-type name="signed long"/>
_______________________________________________ PySide mailing list PySide@qt-project.org http://lists.qt-project.org/mailman/listinfo/pyside
