* J. W. Jagersma: > I did some digging and found a way to solve this issue. I have > submitted a bug report with this proposed patch here: > https://gcc.gnu.org/bugzilla/show_bug.cgi?id=93981 I see. Yes, this looks like a bug. > Ideally though, with -fasynchronous-unwind-tables I think gcc could > emit just one EH region that spans the entire try block, instead of > considering for each individual instruction if it could throw or not. I believe the per-statement analysis is important for optimizing C++ code. Thanks, Florian
