[Bioperl-l] Thoughts on some test reorganization

[Chris Fields]

On Nov 17, 2008, at 11:30 AM, Sendu Bala wrote:

> Hilmar Lapp wrote:
>> In other words, not everyone out there upgrades the OS eagerly.
>> I agree it's reasonable not to put a lot of energy into fixing bugs  
>> that only show up under Perl prior to 5.8.x. But if BioPerl refuses  
>> to even work (or spit out ugly warnings) under 5.6, isn't that a  
>> bit too much of forcing upgrades on people who may not necessarily  
>> need it?
>
> My thoughts as well.
>
> Chris, did you see something specific to justify a change? Like, for  
> 1.5.2 there were specific modules/pragmas only first included in 5.6  
> that motivated the change.

No, hence my bit indicating that 5.6 should work, and whether or not  
we want to make the 5.8 requirement explicit.  I also don't think we  
should be fixing bugs or making changes to deal with a 5-yr-old perl  
release when upgrading one's local perl is a much better option (not  
to mention the benefit of bug fixes, more cohesive core, better  
security, etc).  As I mentioned, even 5.8 has effectively been 'end-of- 
lifed', so why actively support a version that is even older than that?

> I don't think requiring people upgrade their perl just so we can  
> enjoy some entirely /theoretical/ benefit really makes much sense.


We can leave the indication that we require 5.6.1 and up but recommend  
5.8.x (already in place) and will only support fixes for perl 5.8 (not  
made explicit).  99.9% of the time when a bug is reported the perl  
version will not make a difference, but I don't want to be shoehorned  
into supporting an old version of perl when push comes to shove.

chris