Can't compile with MSVC (and therefore install with pip) due to reusing of built-in min/max function name

Copied from Can't compile with MSVC (and therefore install ... (lscsoft/ligo-segments#17) • Unassigned.

---

Dear experts,

We rely on ligo-segments as a dependency of GWpy, which itself is a dependency for our package. We use that outside of LIGO computing environment which complicates things somewhat. Specifically, we found that one can't do pip install ligo-segments on windows in anaconda environment where MSVC is the default compiler (typical case), which leads to users being unable to install our package with pip as well¹.

The issue itself is with following lines: https://git.ligo.org/lscsoft/ligo-segments/-/blob/master/src/segmentlist.c#L199, https://git.ligo.org/lscsoft/ligo-segments/-/blob/master/src/segmentlist.c#L218

For which compiler says:

      src/segmentlist.c(199): error C2059: syntax error: '<parameter-list>'
      src/segmentlist.c(218): error C2059: syntax error: '<parameter-list>'

The message is a bit cryptic, but I think the problem is with function names on these lines, min and max - these keywords have a special meaning, or at least MSVC assumes so here. I was able to rectify the issue locally renaming them to minimum and maximum and correspondingly changing the calls on lines https://git.ligo.org/lscsoft/ligo-segments/-/blob/master/src/segmentlist.c#L343 and https://git.ligo.org/lscsoft/ligo-segments/-/blob/master/src/segmentlist.c#L350. I did not do thorough testing, but if this is the only place where these functions are used then the change should be rather straightforward and not affect any other usage/environments? In that case I'm happy to implement the fix myself and submit a merge request.

1. Yes, GWpy can be installed with conda (and will contain already compiled binaries of ligo-segments), but not our package, and there are more problems for us when it comes to development dependencies which we manage with poetry - it only supports pip. Long story short would be great to be able to do everything with pip, especially if the fix is as easy as described here. ↩