https://gcc.gnu.org/bugzilla/show_bug.cgi?id=97090
--- Comment #6 from David Malcolm <dmalcolm at gcc dot gnu.org> --- (In reply to Christophe Lyon from comment #1) > I see random results from one run to another, so it's likely that something > is not initialized correctly. I think it's due to places in -fanalyzer that iterate though hash_map and hash_set, that are thus implicitly relying on the ordering and thus on specific pointer values. Address-space randomization thus introduces a random element. I've landed some fixes for this in master recently, and am testing some more (sorting every time I use the results of such an iteration).
