Re: [PATCH] Fix expansion ICE on store to CONST_DECL (PR middle-end/77959)

2016-10-14 Thread Richard Biener
On October 14, 2016 7:20:43 PM GMT+02:00, Jakub Jelinek wrote: >Hi! > >The following (invalid) testcase ICEs, because we try to store into >CONST_DECL's FIELD. Normally in GIMPLE we have MEM_REF[] and >writes to that expand gracefully into a MEM, but as soon as we use

[PATCH] Fix expansion ICE on store to CONST_DECL (PR middle-end/77959)

2016-10-14 Thread Jakub Jelinek
Hi! The following (invalid) testcase ICEs, because we try to store into CONST_DECL's FIELD. Normally in GIMPLE we have MEM_REF[] and writes to that expand gracefully into a MEM, but as soon as we use get_inner_reference in expand_assignment (even if the MEM is just reverse order, or we just want