[Bioperl-l] minicpan (CPAN::Mini) doesn't mirror BioPerl-1.6.1

[Chris Fields]

On Mar 4, 2011, at 12:16 PM, George Hartzell wrote:

> Chris Fields writes:
>> On Mar 3, 2011, at 5:04 PM, George Hartzell wrote:
>> 
>>> Hi all,
>>> 
>>> I'm using minicpan (from CPAN::Mini) to keep a cpan mirror.  My mirror
>>> does not contain BioPerl-1.6.1.tar.gz even though search.cpan.org
>>> seems to have it.  The only tarballs that end up in Chris' directory
>>> are these:
>>> 
>>> /Users/hartzell/tmp/minicpan//authors/id/C/CJ/CJFIELDS/BioPerl-db-1.6.0.tar.gz
>>> /Users/hartzell/tmp/minicpan//authors/id/C/CJ/CJFIELDS/BioPerl-network-1.6.0.tar.gz
>>> /Users/hartzell/tmp/minicpan//authors/id/C/CJ/CJFIELDS/BioPerl-run-1.6.1.tar.gz
>>> 
>>> It is listed in the 02packages.details.txt.gz file that is mirrored.
>>> 
>>> Thoughts, suggestions?
>>> 
>>> Thanks,
>>> 
>>> g.
>> 
>> That's odd; could it be the repo you created the mirror from?
>> 
>> This is my PAUSE account:
>> 
>> BioPerl-1.6.0.meta                                   78518  Sun, 25 Jan 2009 05:19:20 GMT
>> BioPerl-1.6.0.readme                                 12174  Sun, 25 Jan 2009 05:19:41 GMT
>> BioPerl-1.6.0.tar.gz                               6813396  Sun, 25 Jan 2009 17:14:12 GMT
>> BioPerl-1.6.1.meta                                   82231  Tue, 29 Sep 2009 16:33:52 GMT
>> BioPerl-1.6.1.readme                                 12174  Tue, 29 Sep 2009 16:33:49 GMT
>> BioPerl-1.6.1.tar.gz                               10462526  Tue, 29 Sep 2009 16:43:11 GMT
>> BioPerl-DB-1.6.0_001.tar.gz                         494106  Wed, 30 Sep 2009 20:49:59 GMT
>> BioPerl-Network-1.6.0_001.tar.gz                   2198338  Wed, 30 Sep 2009 20:50:18 GMT
>> BioPerl-Run-1.6.1_001.tar.gz                       4289239  Wed, 30 Sep 2009 21:07:38 GMT
>> BioPerl-db-1.6.0.meta                                 7728  Tue, 24 Feb 2009 02:58:05 GMT
>> BioPerl-db-1.6.0.readme                               7066  Tue, 24 Feb 2009 02:58:05 GMT
>> BioPerl-db-1.6.0.tar.gz                             392773  Tue, 24 Feb 2009 15:33:05 GMT
>> BioPerl-network-1.6.0.meta                            1268  Tue, 24 Feb 2009 03:00:26 GMT
>> BioPerl-network-1.6.0.readme                          3669  Tue, 24 Feb 2009 03:00:28 GMT
>> BioPerl-network-1.6.0.tar.gz                       1410782  Tue, 24 Feb 2009 15:33:16 GMT
>> BioPerl-run-1.6.0.meta                               10289  Tue, 24 Feb 2009 02:59:26 GMT
>> BioPerl-run-1.6.0.readme                               576  Tue, 24 Feb 2009 02:59:26 GMT
>> BioPerl-run-1.6.0.tar.gz                           4082430  Tue, 24 Feb 2009 15:34:43 GMT
>> BioPerl-run-1.6.1.meta                               10289  Wed, 25 Feb 2009 16:07:06 GMT
>> BioPerl-run-1.6.1.readme                               576  Wed, 25 Feb 2009 16:07:06 GMT
>> BioPerl-run-1.6.1.tar.gz                           4082463  Wed, 25 Feb 2009 16:14:37 GMT
>> CHECKSUMS                                             5035  Sun, 03 Oct 2010 19:21:56 GMT
>> 
>> chris
>> 
>> 
> 
> Ok, apparently we've earned a reputation, and it has preceeded us....
> Here's the code that Ricardo uses to filter the list of things to mirror:
> 
>  sub _filter_module {
>    my $self = shift;
>    my $args = shift;
> 
>    if ($self->{skip_perl}) {
>      return 1 if $args->{path} =~ m{/(?:emb|syb|bio)?perl-\d}i;
>      return 1 if $args->{path} =~ m{/(?:parrot|ponie)-\d}i;
>      return 1 if $args->{path} =~ m{/(?:kurila)-\d}i;
>      return 1 if $args->{path} =~ m{/\bperl-?5\.004}i;
>      return 1 if $args->{path} =~ m{/\bperl_mlb\.zip}i;
>    }
> 
>    return 1 if $self->__do_filter($self->{path_filters},   $args->{path});
>    return 1 if $self->__do_filter($self->{module_filters}, $args->{module});
>    return 0;
>  }
> 
> I'll ask him why he's filtering bioperl out and we can see if we can
> fix it.  I'm hoping that it's just an over-eager cut-n-paste of stuff
> that he thinks of as the "perl" source (which is why it's within the
> skip_perl test) and not that we've done something egregious.
> 
> g.

Could be the size of the distribution was considered egregious (we're at 10 MB compressed I think).  Most of the other dists that appear to be filtered are fairly large as well.  That's nothing really these days; would be interested to hear what Ricardo says.

chris