Hi Ryan,

As I could easily get clang compiling I did a scan through your log
and have concluded that almost all the warnings are either false
positives, i.e. NodeVisitor,  or not critical such as unused
parameters.

In two cases I spotted genuine problems
osg::OcclusionQueryNode::releaseGLObjects(..) wasn't const when it
should have been, and osgViewer::Viewer::take(View) should have
qualified the View's namespace to make sure it overloaded the take
method correctly.  I've fixed this two problems and checked the fixes
into svn/trunk and the OSG-3.0 branch.   In both cases these were bugs
and in the case of  OcclusionQueryNode it will have resulted in a GL
resource leak and potential errors on closing and opening contexts.

There is chance that amongst the hundreds of warnings there are some
other genuine bugs lurking but with so many false positives it's hard
to spot them.  For the 3.0.0 release I'll now do any further
investigation into the clang results/testing with clang as I've
probably got the most out of it in the time available.

Robert.
_______________________________________________
osg-users mailing list
[email protected]
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org

Reply via email to