Matthias Görgens

> For future reference, if you force push then the reviewer has to re-review the whole PR. If you push normally, then the reviewer just needs to check the additional...

> Thanks @matthiasgoergens Thank you for the quick review, and patience with me forgetting one rename! I'm just getting started with contributing.

@rhettinger @hauntsaninja @mdickinson You don't need a loop for this. See [this pure Python prototype](https://gist.github.com/matthiasgoergens/28e58abd6beaea72bd2e49085c435966) inspired by [an old StackOverflow question of mine](https://stackoverflow.com/questions/44991042/binary-search-bisection-for-floating-point-numbers). And here's a [prototype of a fix...

Unfortunately, `-frandom-seed` doesn't seem to be working. See how the file names differ, despite the same seed: ```console ~/m/p/r/isa (master)$ riscv64-unknown-elf-gcc -frandom-seed=1 --verbose -march=rv32g -mabi=ilp32 -static -mcmodel=medany -frandom-seed=1 -fvisibility=hidden -nostdlib...

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65015 suggests that combining `-frandom-seed` and `-save-temps` should work. And when I'm calling the compiler manually, that does seem to be the case. However, I haven't managed to convince `configure`...

Reduce still loops internally.

Thanks for noticing the bug!

The pure Python prototype would turn into a small C function. Not an entire new module.

@pochmann I implemented your suggestions in https://gist.github.com/matthiasgoergens/28e58abd6beaea72bd2e49085c435966

I overhauled the C PR https://github.com/matthiasgoergens/cpython/pull/5/files Tests pass now. I'll point it at the upstream CPython repository, after I polish the git history.