On 2025-10-12 06:42:01, Thierry Dumont wrote:
> Today, maxima-sage was updated (or introduced ?) in Debian testing.
> 
> Will it solve the impossibility to compile sage with system maxima ?
> 
> (the actuel problem was th lisp interpreter)

Sage's version requirement for maxima is almost entirely based on
minor output differences in the doctests. Doctests compare results as
strings, so "2+2" and "4" are considered different, but Maxima often
gets smarter and the former becomes the latter in a new version.

If anyone wants to improve this, it would be really helpful to move as
many of these tests as possible to pytest; or at least, refactor them
to check bool(actual == expected) so that we look only for the result
of "True". We waste a lot of time updating our doctests for every
release, and then users waste a lot of time recompiling the brand-new
maxima that we have to require to pass the new tests.

Anyway, the good news is that if you don't care about 100% strcmp()
doctest compatibility, you can usually edit out the version bound and
the old version of maxima will work just fine.

-- 
You received this message because you are subscribed to the Google Groups 
"sage-devel" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
To view this discussion visit 
https://groups.google.com/d/msgid/sage-devel/aOuidgxN8hwUmkd3%40stitch.

Reply via email to