On Thu, Jan 16, 2025 at 09:10:20AM +0000, Ian Jackson wrote:
> Olly Betts writes ("Re: Bug#1093153: xapian-omega autopkgtests and faketime
> on 32-bit"):
> > I think faketime() with broken support for faking time() shouldn't be
> > allowed to migrate to testing though, at which point a more complex
> > check here doesn't seem to be necessary.
>
> Yes. I will see if I can fix it.
Good luck!
Incidentally, having mulled it over it occurs to me that faketime's
handling of this is also broken upstream. I'm not sure where other
Linux distros which support 32-bit architectures are with their
equivalent to our t64 transition but this seems likely to manifest
somewhere outside of Debian so (with my xapian upstream hat on) maybe I
should look at making the faketime test in omegatest.pl detect this case
anyway. It would be more satisfactory to test time() directly as that's
what we actually need faketime to affect here.
Mostly mentioning as you may find the autopkgtest starts passing as
a result. If this bug stays RC and against faketime until fixed that
should prevent unwanted migration.
Cheers,
Olly
