Re: [PATCH] fvwm reproducibility patch

[ng0%n0.is@localhost]

Hi,

Greg Troxel transcribed 388 bytes:
> I wonder what our current situation and approach is for reproducible
> builds.  I am sympathetic in general, but am also wary of engaging in
> large scale pushing of wet noodles uphill if upstreams don't want to
> take patches.  (Any one package of course doesn't hurt.)
> 
> Are we able to build a package reproducibly if there are certain
> properties of the upstream package?  How many do this?
> 

If upstream doesn't take it, and sometimes this happens for irrational
reasons, we can still share it with other projects. At least this was
what was supposed to happen and what I tried to push for in the last
project (Guix). The amount of projects which refused patches like this
were low.

I can give a longer reply later today.

If we do this with more packages, we might want to look into where
the line is drawn. There could be variations of packages if someone
does not want this / resources are not enough. One example I could
go into is how the bootstrap of rustc improved over time (it involved
lots of builds at some point starting from mrustc). I've had
some exposure to the good and bad parts of this. If you want to,
I could try and come up with some reasonable guidelines for the
documentation in the next months for review?