On Wed, Jul 20, 2022 at 9:43 PM Nils Bruin <nbr...@sfu.ca> wrote:
>
> I've noticed that after a modest upgrade (just "dnf update"), with some 
> regularity sage breaks. In the most recent edition, "libflint.so" couldn't be 
> found.
>
> I guess there is some prerequisite that sage was relying on to be provided by 
> the operating system was changed? Is sage relying on the system python 
> nowadays?

by default, Sage will be using the system python3, sure.
And gmp, flint, ntl, mpfr, arb, pari, ecl, singular, R, glpk, etc etc etc

On a recent Fedora on a newish desktop building, docs included, from
scratch, with all the usable system packages, takes less than an hour.


>
> The only way I have found to clean this up is by doing a "make distclean", 
> which makes for quite a heavy rebuild process! Is there a more modest 
> approach?


>
> It's nice that we're integrating better with OS-supplied prerequisites, but 
> it mixes very badly with building from source on an actual system one works 
> on:
>
> I can see the nightmare scenario where you've run a small system update (as 
> one should do to keep up-to-date with vulnerabilities) in the evening (or 
> perhaps your computer did it for you!) and the next morning you arrive in 
> front of the class with a demo where you now discover that sage is broken and 
> requires an hour's worth of build time to get up to snuff again.
>
> Is there a better/safer way of having both an up-to-date system and a 
> reliable sage install?

As Matthias wrote, it's not Sage specific, and if the ABI of the
system package changes, then yes, the relevant part has to be rebuilt.


>
> [Is the problem perhaps that fedora bumps the default "python" to a newer 
> version for which sage doesn't have its libraries built/accessible? In that 
> case: fedora doesn't delete older versions. I can still run python3.9 or 
> python3.8. So in that case perhaps sage should be a little more careful about 
> what python interpreter it chooses and be a bit more explicit about it (or at 
> least have an option for that)?. I don't know if "python" is the problem]


>
> --
> 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 sage-devel+unsubscr...@googlegroups.com.
> To view this discussion on the web visit 
> https://groups.google.com/d/msgid/sage-devel/d46378df-8652-45d4-96f9-ced715d8ec6cn%40googlegroups.com.

-- 
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 sage-devel+unsubscr...@googlegroups.com.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/sage-devel/CAAWYfq2pX9AqtKk2Fgum%3Doc9b3r3qgcoOmtnqhQnwUEzm6meVA%40mail.gmail.com.

Reply via email to