[Bioperl-l] Bioperl versioning

Dave Howorth dhoworth at mrc-lmb.cam.ac.uk
Mon Oct 23 11:55:42 EDT 2006


Sendu Bala wrote:
> Chris Fields wrote:
>>> Dave Howorth wrote:
>>>>>> That's the user point of view - how does the developer actually tell
>>>>>> CPAN that something is a developer release so that normal users don't
>>>>>> automatically install it?
>>>>> I found this:
>>>>> http://www.atrixnet.com/PM/So-You-Want-to-Contribute-to-The-CPAN.ppt
>>>>>
>>>>> Is says that $VERSION should simply be changed from a naked number into
>>>>> a single quoted number and this should be recognized by the CPAN
>>> indexer.
>>>> <http://search.cpan.org/~nwclark/perl-
>>> 5.8.8/pod/perlmodstyle.pod#Version_numbering>
>>>
>>> Thanks for that.
>>>
>>> I guess from that the 1.5.2 version number should be:
>>>
>>> $VERSION = 1.05_02

I believe so - the underscore is key. Look at your favourite CPAN
modules and see what they do.

>>> And 1.6 would be
>>>
>>> $VERSION = 1.06
>>>
>>> But will this cause a problem wrt 1.4? 1.4 has:

I think it will cause a problem, yes. 1.4 > 1.06 As a workaround, you
could remove 1.4 from CPAN and require everybody who installs from CPAN
to uninstall it before installing 1.06.

>>> $VERSION = 1.4;
>>>
>>> Is 1.4 lower than 1.06? Should we keep to a single digit version, so
>>> 1.5_02 and 1.6? Does this really not work with CPAN?

I think that would work but see at the end.

>> Should we call them
>>> version fifty and version sixty? 1.50_02, 1.60?

Then you can count 1.50_02, 1.50_03, 1.52, 1.53_01 ... if you wish.

>> Doesn't perl 5.6.1 and up use the 'x.y.z' versioning syntax?  It would be
>> much simpler to use that. 
> 
> That does not present us with a way to have 1.5.2 marked as a developer 
> release in CPAN.
> 
> Also, see the discussion here: 
> http://perldoc.perl.org/functions/require.html
> 
> Since we require 5.6.1 the backwards-compatible issues maybe don't apply 
> to us, but do these ideas work with modules, or just Perl itself? Is 
> CPAN et al. happy with this form of versioning?

I'm not an expert :( It's my understanding that there is an awful lot of
flexibility in Perl module version numbering (as you might expect :)
However, I believe there are some gotchas. So I would recommend (a)
finding an expert and (b) trying an experiment!

> /Something/ needs to be done about Bioperl versioning, because the 
> current 1.4 or 1.5 is completely inadequate.
> _______________________________________________
> Bioperl-l mailing list
> Bioperl-l at lists.open-bio.org
> http://lists.open-bio.org/mailman/listinfo/bioperl-l
> 
> 



More information about the Bioperl-l mailing list