https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82258
--- Comment #10 from Paul Thomas <pault at gcc dot gnu.org> --- Created attachment 42254 --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=42254&action=edit tree dump for testcase of comment # 9 The output in comment #9 has left me totally perplexed. I cannot for the life of me understand why the lbounds that should be positive are wrong while everything else is OK. I have attached the tree dump (-fdump-tree-original). Could you compare it on your system, please? If they are the same, I would guess that we have uncovered some weird backend bug. If they are different, we have something to work on. Note that the computation of the lower and upper bounds is pretty similar in respect of what is extracted from the descriptor. Also, there is no sign of the new field 'span' being addressed at all in this testcase. Cordialement Paul
