Re: CVS commit: pkgsrc/mk/flavor/pkg

[Joerg Sonnenberger <joerg%britannica.bec.de@localhost>]

On Sat, Jun 12, 2010 at 02:05:07PM -0400, Greg Troxel wrote:
> Joerg Sonnenberger <joerg%britannica.bec.de@localhost> writes:
> 
> > On Sat, Jun 12, 2010 at 01:40:13PM -0400, Greg Troxel wrote:
> >> The problem I am having is that things are being changed to support the
> >> non-replace approach with destdir that are causing regressions with make
> >> replace.  I think make replace really should work before we turn on
> >> destdir by default - but we already have.
> >
> > Have you considered that you problem you are facing with osabi is the
> > correct behavior for most other cases?
> 
> That sounds like "no one should ever use make replace".  No, it isn't
> the correct behavior, if one accepts that make replace causing an unsafe
> state and pkg_rr doing a further make replace to straighten things out
> is reasonable.

Consider things like updating from one PHP or Perl version to another.
It is not "unsafe", it is downright broken. Fundamentally broken. This
is different from osabi as it will normally continue to work.

I am using "make replace". I do know when it is sane to use and when
not. What you are doing is "I don't care about breaking dependencies".
The argument of libraries being open coded doesn't really work -- it is
a known bug and fixing it will require going to the same kind of
situation as osabi currently faces. While it can handle the case of not
having a shared major bump, it doesn't address the reverse case and from
the perspective of make replace, it would be worse. So please do us all
a favour and fix the case for pkg_rr properly -- there is no sane way to
do incremental replacements in some critical situations.

Joerg