[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: binNMU or reproducible builds (choose only one)



Hi,

On Thu, Sep 24, 2015 at 11:56:56AM +0100, Colin Watson wrote:
> On Thu, Sep 24, 2015 at 12:29:59PM +0200, Santiago Vila wrote:
> > But once we are able to trigger a rebuild with sourceful NMUs, as
> > Ubuntu does, binNMUs will hopefully be a thing of the past.
> 
> Amusingly, the way we do it in Ubuntu is a huge hassle in some cases,
> and at least some of us would rather have binNMUs.  (That's partly
> because it's a manual process; if it were automated it would be better,
> but it still wouldn't solve the problem that in some cases you really do
> want to do single-architecture rebuilds without having to rebuild a
> stack of packages on slower architectures entirely unnecessarily.  Hi,
> Haskell.)

Maybe one could generate a pseudo binary package for architectures
not included in the binNMU, so that all architectures provide their
binary packages with the same version.

I think it easier to understand by giving an example:

Let's assume a binNMU is needed for "demopkg" in version "23.42-1"
for i386 and amd64.

Current behaviour would result in:

[amd64, i386]: 23.42-1+b1
[armel, armhf, arm64, mips, ...]: 23.42-1

The binNMU process could now prepare 23.42-1+b1 for all unaffected
platforms automatically by downloading the existing binary package,
updating the changelog file, increasing the version and saving it
again. That would result in:

[all architectures]: 23.42-1+b1

That should make it possible to have symlinks for
/usr/share/doc/$pkgname from all to $arch (assuming "all" is also
regenerated) and fixes the version mismatch problem regarding
multi-arch.

The obvious disadvantage is, that users will have to download
a more or less unchanged package again.

-- Sebastian

Attachment: signature.asc
Description: Digital signature


Reply to: