[Biocorba-l] SeqFeature and Id

[Brad Chapman]

Hey guys;
Thanks for the feedback.

[problem with recursive structs in ORBit]

Juha:
> > > I have reported the bug some time ago. I will check what is the status.

It's supposedly fixed in the ORBit 2.x series, but I tried out the
stable CVS branch of ORBit 0.5.x series and the fix wasn't there.

Ewan:
> > Can we restructure the IDL not to have the bug. We don't expect the
> > SeqFeature struct to be endlessly recursive (do we?) I hope not.

Juha:
> I do not find clean way to do it... We may need to couple sub_locations
> into SeqFeature interface directly or something if we decide not to use
> recursion.

If we did this, it would take care of most cases (ie.
join(51941..52048,52136..52432,52640..52885)), and would just get
stuck on more complicated cases with nested joins and things. This
tradeoff seems okay to me for now, since it would let most things
work.

me:
> > > > --> The inclusion of CORBA.idl causes lots of problems for me.

Juha:
> > > The typedefs should be part of ORB, but it seems that there are no
> > > standard way to include those. I think we should use existing
> > > definitions and make the CORBA.idl interchangeable, or something. Any
> > > other comments?

Ewan:
> > Let's just build a typedef.idl and distribute that separately...

This sounds good to me, and is actually what I've done on
biopython-corba just to have something that works to code off of:

http://cvs.biopython.org/cgi-bin/viewcvs/viewcvs.cgi/biopython-corba/idl/types.idl?rev=1.1&content-type=text/vnd.viewcvs-markup&cvsroot=biopython

(sorry about the wrapping on that URL). Does this seem like a good
enough working solution for people? (even if it's called types
instead of typedef? :-).

> BTW Currently our own sequences are declared in modules just after
> corresponding items. Also typedefs for constants (e.g SeqTypeDef) etc
> are declared in "appropriate" modules. I guess that this is OK? I.e. no
> need to move those into separate file/module.

This is fine by me -- I'm definitely not too worried about
reorganizing things; just in having all of the definitions needed to
make the IDLs work :-).

Brad