Hi, OpenMP is currently more or less unusable with the C++ front-end because of EH issues (see PR26823). This unfortunate situation is dragging on for more than two months now and makes further testing impossible.
Some of the problems were fixed in PR26084, some still remain: The original (unreduced) testcase from PR26084 still causes an ICE. This is probably the same problem as in PR26823. We also have a third PR about an EH problem (PR26913) which has probably the same cause. Alas there seems to be no activity in that direction since PR26084 was closed. Could somebody of the gomp-maintainers please have a look? Thanks a lot in advance, Volker
