[Bioperl-l] For CVS developers -potentialpitfallwith"returnundef"

[Heikki Lehvaslaiho]

OK. I've gone through all cases where return and undef are on the same lines.
I've done changes in 185 files.

My aims have ben the following:

1. Remove undef from return undef when not necessary.
	This will make it easier to spot cases where undef matters in the future
	Most of the changes fall into this category. The context is clearly scalar.

2. Returning undef when user expects en empty list is bad

./Bio/Tools/Est2Genome.pm fixed
./Bio/SearchIO/blast.pm:2330:  should return (undef, undef) for clarity?
                               not fixed
./Bio/Matrix/PSM/SiteMatrix.pm  fixed
./Bio/Matrix/PSM/Psm  fixed
./Bio/DB/Taxonomy::entrez.pm fixed

3. If docs say method returns nothing, explicit undef is not the right thing 
to return

4. do not return an explicit undef if the method is supposed to return false 
on failure


Before I do the commit, I'd like to see number people to do 'make test' on 
bioperl-live and report back after the commit they see changes. There are 
quite a few tests that fail currently.

I'll do the commit tomorrow Wednesday at 9 o'cock GMT.

	-Heikki






On Friday 02 June 2006 09:52, Heikki Lehvaslaiho wrote:
> I've started going through the files that have 'return undef' lines.
> I'll report back later.
>
> Initial impression is that there are a few cases where the context
> indicates list to be returned but failure returns an explicit undef. I'll
> fix those.
>
> Most of the cases are much more ambiguous. Even when documentation says the
> failure returns undef, it is clearly meant to mean false. In most cases
> documentation does not comment on return value at all. Luckily the context
> is almost always scalar and therefore it does not matter too much.
>
> I seem to be changing 'return undef' to plain 'return' a bit overzealously,
> so do not take it personally.
>
> 	-Heikki
>
> On Thursday 01 June 2006 19:46, Chris Fields wrote:
> > ....
> >
> > > > Again, didn't do that.
> > >
> > > I'm very sorry that I allowed the ambiguity, but my comments were
> > > certainly not directed at your recent changes to Bio::Restriction::IO.
> > > In fact, I put in the above * comment to exclude your changes from my
> > > discussion; you changed the docs because the code never did what they
> > > said they did (the docs were bad). That's fine (good!). My comments
> > > were a general point, slightly directed at the idea of changing all the
> > > return undef;s - changing the code so that it no longer matches the
> > > docs of a previously working method. That's what I think is bad. Though
> > > in this particular case it shouldn't make any difference at all.
> >
> > Agreed.  In any case, if tests have been properly set up then they should
> > catch problems.  This is, of course, if they are properly set up.
> >
> > Chris
> >
> > > _______________________________________________
> > > Bioperl-l mailing list
> > > Bioperl-l at lists.open-bio.org
> > > http://lists.open-bio.org/mailman/listinfo/bioperl-l
> >
> > _______________________________________________
> > Bioperl-l mailing list
> > Bioperl-l at lists.open-bio.org
> > http://lists.open-bio.org/mailman/listinfo/bioperl-l

-- 
______ _/      _/_____________________________________________________
      _/      _/
     _/  _/  _/  Heikki Lehvaslaiho    heikki at_sanbi _ac _za
    _/_/_/_/_/  Associate Professor    skype: heikki_lehvaslaiho
   _/  _/  _/  SANBI, South African National Bioinformatics Institute
  _/  _/  _/  University of Western Cape, South Africa
     _/      Phone: +27 21 959 2096   FAX: +27 21 959 2512
___ _/_/_/_/_/________________________________________________________