[Bug rtl-optimization/37262] Two branches of the same condition being emitted

[segher at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=37262

--- Comment #6 from Segher Boessenkool <segher at gcc dot gnu.org> ---
(In reply to Martin Sebor from comment #4)
> I built the gcc-5-branch on powerpc64-linux and the duplicate instructions
> are still there with the second test case compiled at -O2 (see below).
> 
> $ /build/gcc-5.x/gcc/xgcc -B /build/gcc-5.x/gcc -O1 -S -Wall -Wextra

That's -O1.  Is it actually still there at -O2?

For the other case (both bgt and beq), at expand time there is a switch
which could just be a single conditional (only case 2 is not the same
as default, not empty in this case).  This doesn't seem to be powerpc-
specific; could you open a new PR for this?  Thanks.