[Bioperl-l] progress for 1.2.2 bugfix releae

Jason Stajich jason at cgt.duhs.duke.edu
Sat Jun 28 19:46:12 EDT 2003


I've cleaned up a load of things which had basically already been fixed on
the main trunk by needed to be merged over.  A lot of things were revealed
when running perl5.8.0 so I've tried to clean them up as well.

The weirdness we saw with t/DB.t failing during make test (but succeeding
when you run perl -w -I. t/DB.t) can be avoided by explicitly calling
$seqio->close in the last test block.  Don't ask me why, I don't know.
I've fixed this on the trunk and branch.

Also merged the -noclose option for Bio::Root::IO (which when set will not
close a filehandle upon object DESTROY) for the cases when we have shared
filehandles.  This also prevents STDIN/OUT/ERR from being closed if they
are used in an object which uses Bio::Root::IO for IO.  Without this we
get some problems with some of the Indexing code and since it is an extra
option not a major API change I decided it should be on the stable branch
anways.

I also fixed bugs 1458,1459 on both the main trunk and branch - thanks for
the reports Mike Olsen!  Give it a try with current code in CVS and we'll
close the bugs out if you agree they are fixed.

Also added some notes in the Changes changelog. Checklist for the release
need to make sure we up the version number- how are we doing that now.  I
think the plan is to have all the changes in by Monday.  Once that happens
we need to tag the release, do a clean cvs export checkout of the code, do
a make manifest, and tar & zip up the whole thing.

I also did some cleanup on Bio::DB::Flat::BinarySearch to make it more
uniform between perl versions by ussing
sysseek($fh,0,SEEK_END);
$len = sysseek($fh,0,SEEK_CUR) instead of
$len = tell($fh) in order to get the length of the file because tell
deals with buffered io while sysseek does not.  I was having problems on
5.8.0 and applied the code replaced above where appropriate.

Hilmar- I think there are still some warnings being spewed by some
Ontology related code which we should probably get to before the code is
released.

t/GOterm.....................ok 1/59Use of uninitialized value in
concatenation (.) or string at
./branch-1.2/blib/lib/Bio/Ontology/GOterm.pm line 285.

t/OntologyEngine.............ok 2/22Use of uninitialized value in hash
element at ./branch-1.2/blib/lib/Bio/Ontology/SimpleOntologyEngine.pm
line 441.


Also added some notes in the Changes changelog.

Checklist for the release:

  Need to make sure we up the version number in the right place. How are
  we doing that now?

  Brian will move the FAQ doc/faq/faq.xml and regenerate the FAQ
  (bumping version number too - let me know if you can't do it).

  Anything else?

-jason

--
Jason Stajich
Duke University
jason at cgt.mc.duke.edu


More information about the Bioperl-l mailing list