https://gcc.gnu.org/bugzilla/show_bug.cgi?id=127262

--- Comment #2 from Drea Pinski <pinskia at gcc dot gnu.org> ---
There is a GIMPLE_SWITCH:
  switch (_4) <default: <L35> [80.00%], case 1: <L36> [4.00%], case 9: <L37>
[12.00%], case 11: <L37> [12.00%], case 13: lbl_f2_b9L.6 [4.00%], case 25:
<L37> [12.00%]>




The type of the cases don't need to match exactly the type of index. And that
is the cause of the issue.

The code starting around line 1895 in gimple-predicate-analysis.cc needs to be
fixed to handle that.

Reply via email to