https://gcc.gnu.org/bugzilla/show_bug.cgi?id=125442

--- Comment #6 from Souradipto Das <souradiptodas6 at gmail dot com> ---
Ah, looking closely at my test file again, you're absolutely right about the
testsuite coverage—the macro loop handles both cases under the hood, but my
actual test file was only running the bitwise OR path through the compiler. 

My mistake! I'm updating the testsuite locally to include both the `src_and`
and `src_or` verification functions, and I'll clean up the match.pd logic to
reuse the captured zero operand as well. I'll upload a v2 patch once the tests
finish. Thanks for the catch!

Reply via email to