[Debian-science-sagemath] sagemath testing migration

[Ximin Luo]

Tobias Hansen:
> Hi list,
> 
> on the tracker page of sagemath this appeared in the testing migration box:
> 
> sagemath-jupyter/i386 unsatisfiable Depends: sagemath (>= 7.4-3)
> sagemath-jupyter/i386 unsatisfiable Depends: sagemath (<< 7.4-3.1~)
> 
> This kinda looks as if missing binaries on i386 are a blocker. Is that
> the case?
> 

Yes:

<adsb> the issue is an arch:all package. and yes, they're expected to be installable on i386 and amd64
<adsb> your choices are make that possible or persuade the release team to tell britney to ignore it

> If yes, I would say we wait for the piuparts run and then do another
> upload that:
> - lowers the barrier for the tests to pass so that the i386 build succeeds
> - fixes the piuparts checks, if necessary
> 
> Because I don't have time to look properly into the i386 segfaults in
> the next two days.
> 

I'll see if I have a few hours for this today, otherwise I'll do as you say.

If we can't fix the segfault by Jan 25 however, I think the better result is to ask the release team for an exception to the above rule (i.e. keep amd64 but drop i386 for stable), because I don't see the point of releasing a piece of software to i386 that segfaults for two years.

However, I understand to hit the Jan 5 deadline we do have to "game the system" a little bit here. So I'll do what you suggested, to give us some breathing space for a better outcome in a month.

X

-- 
GPG: ed25519/56034877E1F87C35
GPG: rsa4096/1318EFAC5FBBDBCE
https://github.com/infinity0/pubkeys.git