Web lists-archives.com

Re: RFC: Support for selective usage of (fake)root during package build (R³)




Hi Adam,

Thanks for the feedback. :)

Adam Borowski:
> [...]
> 
> Have you perhaps gathered some data about how many packages would build
> reproducibly wrt the old state after a blind mechanical change to R³?
> I'd guess it's the vast majority (especially if you change MakeMaker).
> 

I have not, but my gut feeling does indeed agree that it will the vast
majority that "just works(tm)" or FTBFS.  That said, checksums +
diffoscope should assist with anything in between.

(Reminder: static ownerships are still not covered at the moment, so
packages needing those must be in the FTBFS bucket with R³=no).

> 
> I do have a concern: this adds a yet another field every new package needs
> to manually set.  And even if in 10 years from now, the Policy will require
> R³ to be set (even to "yes" ("binary-targets")), it'd be a manual effort for
> every single package.
> 
> Thus, what about making R³ default to "no" in a new dh compat level?
> 

I do understand your concern. However:

 * This proposal is primarily an interface between the package and
   buildpackage/the builders tool.  The debhelper part is just gluing
   this together so it becomes easy to adopt in the average case.

 * The compat level is an interface between the package and debhelper.
   IOW, it would be ill-suited to support R³ as it is not an interface
   to one of the most important members of the specification.

 * Even if we could use a debhelper compat level to introduce R³, it
   would have neutered the adoption rate.  We got plenty of packages,
   where maintainers want to have their package backportable to stable,
   oldstable and even oldoldstable (not to mention various Ubuntu LTS
   releases).

 * For those concerned about mirror metadata (was not mentioned in
   Adam's mail), the R³ field is not inserted in the .dsc with dpkg
   1.19.0.  Accordingly, this has no affect on the size of the
   Sources files.


Thanks,
~Niels