Open N-Coder opened 3 years ago
I changed it to ClassName
and removed the ref counting, but I'm still unsure what'll happen in case of an exception.
I'm thinking (not tested) that in case of an exception, PyErr_Clear
would need to be called for ClassName()
to succeed. I'm also thinking that if there is an exception, fetching it and using it as part of the message may be useful.
Could the
PyErr_Clear
inUtility::ClassName
, which gets called when the passed object has no__cpp_name__
, interfere with the followingSetPyError_
and thePyErr_Fetch
it calls?