This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [patch] combine.c if_then_else_cond() subreg modes
> After some discussion I came up with a different patch which fixes
> the testcase as well, but I couldn't convince myself yet if it's OK
> to compare the mode sizes or if I must compare the modes themselves
> for equality.
const_ints don't have modes (well, they're VOIDmode) so there is no
mode to compare. That's why everyone has to do special things to keep
track of the mode a const_int is supposed to be.
The bug I'm tracking down doesn't go anywhere near cse.c, though.
It's all in combine.c. Even if some other patch "fixed" my testcase,
I still think the code in if_then_else_cond() is wrong.