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

Re: Perl policy / Plans for the upgrade



Le Wed, Jun 16, 1999 at 03:14:41PM -0400, Andy Dougherty écrivait:
> It's just that on a philosophical level, a policy that says you
> are not allowed to do something (in this case have 5.00503 and 5.00504

The policy doesn't forbid you to install both. But for the Debian specific
part of perl, it's much more simple to have only one perl-5.00X package.

> simultaneously) rubs me the wrong way.  If we're going to allow
> simultaneous versions to be installed, I'd prefer to be able to avoid
> limitiations on which versions can be installed simultaneously.

The problem is that you need to change the name of the package for
each bugfixe releases, eg perl-5.00503 and so on. That's not possible
if you want to be able to depend on perl. Because versionned provides are not
yet supported. And also because different version of the same package
cannot be installed at the same time.

> If you assume that 5.00503 and 5.00504 can be installed simultaneously,
> then such a vendorarchlib _does_ become useful, since modules wouldn't
> (usually) have to be recompiled for these versions.

That would make sense yes. However as Debian perl uses an archlib directory
with $apiversion (and not $]) in it, the binary modules would still be found.

> I'm afraid I can't answer that since there are still unresolved issues
> with how 5.006's default libraries will be structured.  For example, one
> thing to consider is that even a pure perl module that works under 5.006
> might not work under 5.005 (e.g. it might use a new language feature or a
> new standard module or a new function in an existing standard module,
> etc.)

That's ok as long as backward compatibility is assured. The modules that
uses 5.006 specific functions could be installed in a versionned directory
that will only be found by perl-5.006 ...

> > We shouldn't wait anymore. Otherwise perl-5.005 will get rejected from
> > the release-goals for potato. :) 
> 
> Well I wouldn't want it to get rejected, so forge on ahead :-).

I already missed it for slink (but it was /really/ too late
when I suggested it). :-)

Cheers,
-- 
Hertzog Raphaël >> 0C4CABF1 >> http://prope.insa-lyon.fr/~rhertzog/


Reply to: