Move bitset.pxi to bitset_base.pxd

[kliem]

We move bitset.pxi to bitset_base.pxd.

So far bitset.pxi contained mostly inline functions for the header but also a few functions that cannot be inlined due to optional arguments.

https://cython.readthedocs.io/en/latest/src/userguide/language_basics.html?highlight=pxi#cython-file-types

Depends on #30597

CC: @videlec

Component: refactoring

Keywords: bitset, header

Author: Jonathan Kliem

Branch/Commit: bc0e1fb

Reviewer: Travis Scrimshaw

Issue created by migration from https://trac.sagemath.org/ticket/30601

[tscrim]

comment:2

Modern Cython uses range, so

-    for i from 0 <= i < bits.size:
+    for i in range(bits.size):
         s[i] = one if bitset_in(bits, i) else zero

If you could make these changes where appropriate, that would be good. Beyond that, LGTM.

[tscrim]

Reviewer: Travis Scrimshaw

[sagetrac-git]

Changed commit from d5e06fd to bc0e1fb

[sagetrac-git]

Branch pushed to git repo; I updated commit sha1. New commits:

bc0e1fb

remove old style loop instructions

[kliem]

comment:4

Done.

I figured it was some old style usage.

Replying to @tscrim:

Modern Cython uses range, so

-    for i from 0 <= i < bits.size:
+    for i in range(bits.size):
         s[i] = one if bitset_in(bits, i) else zero

If you could make these changes where appropriate, that would be good. Beyond that, LGTM.

[tscrim]

comment:5

Replying to @kliem:

Done.

Thank you.

I figured it was some old style usage.

You're probably right, but it was something I noticed from the diff and I didn't look too closely if it was a hold-over or an addition. (Although I do try and change these things when I do such changes, but that is me.)

[kliem]

comment:6

Thank you.

I think range code is much more readable for a person coming from python, so I will change, when I stumble upon it.

[vbraun]

Changed branch from u/gh-kliem/properly_define_bitsets_with_a_header to bc0e1fb