Mailinglist Archive: opensuse (1658 mails)

[Dave Howorth <dhoworth@xxxxxxxxxxxxxxxxx>]

David Haller wrote:
> Hello,
>
> On Wed, 07 Dec 2011, Dave Howorth wrote:
> > Greg Freemyer wrote:
> > > I think the real issue is this line from the perl specfile:
> > >
> > > Provides:       perl-IO-Zlib perl-IO-Compress-Base perl-IO-Compress-Zlib
> > > Obsoletes:      perl-IO-Zlib perl-IO-Compress-Base perl-IO-Compress-Zlib
> [..]
> > I believe *you are in a pit. Please stop digging it deeper!*
>
> Stop digging yourself ;)
>
> > IO::Zlib and IO::Compress::Base are core modules  since Perl 5.10.0.
>
> CPAN disagrees.

Hmm, well CPAN is famously unreliable! :)

I already provided the link http://perldoc.perl.org/index-modules-I.html
that says it is part of core. I already said that YaST says it is part
of core and here is command-line output from 11.2:

  # corelist IO::Compress::Base
  IO::Compress::Base was first released with perl 5.009004
  # corelist IO::Zlib
  IO::Zlib was first released with perl 5.009003

So I have three sources that say it is core and you have something that
says there is a module on CPAN. If you're still doubtful, the changelog
for the incorporation of those modules into the core are at:

http://perldoc.perl.org/perl593delta.html#New-Core-Modules

or on your own machine:

  # perldoc perl593delta

and here's the full list of modules in the current perl (you can go find
the equivalent for previous releases)

http://perldoc.perl.org/perlmodlib.html

On any given machine

  # perldoc perlmodlib

That command shows a file that lists all modules that were core for the
system perl that is installed.

> If they were core-modules, they'd have the perl-tarball as CPAN_FILE,

I thought that too, but apparently it ain't so.

> That openSUSE packages those in the perl-RPM is completely irrelevant,
> they are maintained as seperate modules by seperate people outside of
> perl-core. IMO, the perl.spec needs versioned

I absolutely agree that the complexities of the Perl packaging mechanism
are completely separate from the complexities of the RPM packaging
system and the complex decisions of the opensuse packagers.


Again, I restate my opinion that trying to apply theories as to what
ought to be the case is an unreliable way to determine what is
needed/desirable to package a Perl application. (unless you are already
a Perl Pope). An empirical test, backed by cross-checking the documents,
is a safer approach IMHO.
-- 
To unsubscribe, e-mail: opensuse+unsubscribe@xxxxxxxxxxxx
To contact the owner, e-mail: opensuse+owner@xxxxxxxxxxxx