Request for help to make mariadb-10.3 in Debian reproducible

[Chris Lamb]

Hi Otto,

> > > For Mroonga I nor the upstream developers have no clue yet of the
> > > root cause

Thanks for your comments. I'm replying quickly on two points as I may
not get a chance to get back to your mail in full until Wednesday.

> As a side note, the search engine groonga itself is also currently
> unreproducible in a similar way (build-id changes, root cause
> unclear): 

Heh, so sometimes I get the impression that some maintainers see the
build-id differences and try to address this directly. In the land of
ELF reproducibility, this is almost akin to asking how to "force" two
files to have identical SHA checksums rather than working out why the
files "actually" differ. Not saying that's the case here, but just
being 100% clear for all readers of this mail. :)

> https://tests.reproducible-builds.org/debian/rb-pkg/unstable/amd64/diffoscope-results/groonga.html
> BUT it is reproducible in testing

The main difference between unstable and testing is that we do not
vary the build path in the latter, so this would be my first avenue
when trying to fix this.


Regards,

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