Some help with nroff_output_varies_by_locale_or_utf8

[Chris Lamb]

[keeping pollo in CC as previously requested]


Hi Louis-Philippe,

> > In order that I don't duplicate your efforts though, how much
> > investigation have you done already?
[…]
> This is my first introduction to the wonderful world of
> reproducible-builds debugging and I don't have tons of experience with
> your tooling :)

Well, then this looks like a great introduction. :)  Anyway, what I
might suggest you do in the first instance is try and recreate this
outside of reprotest. For what its worth I am regrettably not at all
familiar with this tool so I am pretty much in the same position as
you right now regarding the num_cpus (etc.) weirdness.

> Looking at the rst2man code, it also seems it is using:
> 
> "locale.setlocale(locale.LC_ALL, '')" [2]

I would guess at a consistent locale rather than patching it out but
I would avoid jumping to any premature conclusions when attempting
any reproducible issue yet alone a locale one. It could be a lack
of installed/configured locales, just as another random example; I
have seen this a quite a few times...

Anyway, best to try and create a minimal testcase outside of
reprotest in the first case. Oh, is the version of rename-flac on
salsa in unstable (and thus on tests.reproducible-builds.org)?


Regards,

-- 
      ,''`.
     : :'  :     Chris Lamb
     `. `'`      lamby at debian.org 🍥 chris-lamb.co.uk
       `-