------- Comment #5 from rth at gcc dot gnu dot org 2010-05-27 15:44 ------- The context in which stmt_can_throw_internal is being called is different pre- and post-inlining, so of course the answer could be different. This is neither surprising nor a bug.
I even tend to doubt (without checking of course) whether this is a regression with the EH rewrite, since I would expect the same new EH edge to have appeared with the old EH code -- anything less would have been a bug in the old EH code. However, in order to see the bug one also needs the VTA code present, which means looking at the redhat-4.4 branch. So yes, Jakub, I think the inliner needs to do something to clean up the debug statement that would appear alone in that new block. I suppose such debug statements would have to be dropped to avoid changing the code due to changes in the CFG. -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=44182
