[Bioperl-l] bioperl-dev or branch?

[Chris Fields]

On May 21, 2009, at 10:55 PM, Robert Buels wrote:

> Mark A. Jensen wrote:
>> relatively well-tested functionality. So I do conceive of bioperl- 
>> dev (rightly
>> or wrongly) as a parallel branch of bioperl-live, but not a temporary
>> "feature" branch as such. It can and prob should be pretty  
>> persistent.
> Branches can be persistent, if there is a really good reason to keep
> them as such.  But this instance does not seem to be one of them.
>
>> This is the way I'd love it to work, modulo the exptl core changes  
>> mentioned
>> above. My own tendency in development is to make stuff as separable  
>> as
>> possible (by specifically overriding core methods in 'Helper'  
>> modules, for example),
>> and if this were part of the bioperl-dev rules of engagement (i.e.,  
>> no core modules,
>> only overrides), then users could count on the behavior you  
>> describe. Mirroring
>> the existing core paths is also key for this expectation.
>
> The problem with trying elaborate ploys to avoid developing on  
> branches
> is the 'trying to keep stuff separable' invariably fails in some  
> cases.
> It is far better to rely to on your modern version control system to
> sanely manage your changes.  I think you guys are pretty scarred from
> years of CVS, a version control system which (by modern standards) is
> laughably broken.  SVN is no shining jewel either, but at least it
> understands the concept of file trees.
>
> Rob

Well, there is the issue that we *dont't* want to pollute core anymore  
than it already is.  I would rather not dump any/all code directly  
into core de novo.  That practice has created code bloat, and  
unmaintained code bloat at that.

chris