Hi Spencer,
"King, Spencer" <spencer.king@wustl.edu> skribis:
Toggle quote (22 lines)
> I agree that longer-term maintenance is a point of concern. Perhaps
> that will change if Julia begins to see more widespread adoption. I
> think one of the biggest issues facing Julia packaging is that there
> currently isn't the same level of interest as languages like Python. I
> also know that there have been some issues with packages not building
> reproducibly due to upstream issues with the internals of the Julia
> compiler, but I'm not going to pretend I have an in-depth
> understanding of that issue.
>
> I have seen your thread about your efforts to package a new version of
> Julia. It looks pretty complex and is definitely a major step up in
> complexity from any packages I've written so far. I think this comes
> back to the issue of interest, Julia just doesn't have the same level
> of interest as other languages in scientific computing so it can be
> challenging to find collaborators.
>
> Like you said, individual package upgrades have gone fine so far. I've
> done a few myself without issues. I agree that it probably won't last
> as the package set grows and we end up with more complex dependency
> trees. However, I imagine that is a similar issue faced by other
> package sets in Guix.
Fortunately, Julia is less of a nice today than it was a few years back.
Hopefully we’ll find more people to help.
Regarding long-term maintenance, have you looked at the importer that
was proposed a while back? (Cc’ing Nicolas.)
I think we try and have an importer and updater in place to ease package
maintenance. Maybe the ‘juliahub’ importer (it seemed almost ready to
me), or maybe something else. It’s been instrumental in maintaining
other package sets.
Thanks for your feedback,
Ludo’.
