Florent Angly | 29 Oct 03:50 2012

Re: Bio::DB::Fasta changes

Hi Chris,

I have done a bunch of work on Bio::DB::Fasta and associated modules 
lately, with the goal of making them more useful and less redundant. 
There was no intentional change of interface that would make it backward 
incompatible and all bioperl tests passed fine. In fact, if I recall 
correctly, I even added more tests because I noticed some holes in test 

I suspect that this bug report you cited unveiled a corner case that was 
not covered by the tests. I'll assign the bug to myself and investigate.


On 29/10/12 08:24, Fields, Christopher J wrote:
> There have been a number of significant changes to Bio::DB::Fasta recently that haven't been discussed
on the bioperl list.  In particular, much of the code has been moved to Bio::DB::IndexedBase.  At the moment
some of these changes are breaking compatibility with other tools (namely MAKER, see:
https://redmine.open-bio.org/issues/3389).  In the latter case I would consider this a significant
API change that needs to be addressed.
> Generally, whenever we make significant changes to modules these should always be run on a branch first
(see recent changes to Bio::Tree), and we should at least discuss this on-list prior to merging with
master.  I do think we should start that discussion now and decide what to do, e.g. roll back changes and push
these to a branch, allow these to stay in master with revisions, etc.
> chris