Hi Carlos,
> Thanks for contacting about this. Indeed, it would be better to fix it
> upstream. I've opened a bug against jdupes [1] when I realized the
> reproducibility issue and the upstream developer acknowledge it, placing
> it in the roadmap, though I don't think he had the time to fix it yet.
Ah yes; in retrospect, I perhaps should have filed this directly
against jdupes myself (or found the existing bug). Thank you for doing
that.
> I just would like to know if you tested the patch? I've seem that it
> happens in x86_32 and armhf, but I don't have the hardware to test the
> fix.
I tested the patch, yes; does it not work for you? The difference
between architectures could simply be a difference in the
reproducibility variations applied on that particular architecture. Or
the differences could be merely an example of the nondeterminism
itself.
> Also would like to ask if I can forward your email to [1] and the
> upstream developer.
Sure, if you think it might help.
Regards,
--
,''`.
: :' : Chris Lamb
`. `'` [email protected] 🍥 chris-lamb.co.uk
`-
