[Open-bio-l] biodata (was Re: bioperl commit)

[Steve Chervitz]

Thanks a bunch for setting this up Jason. 

Regarding "cross-platform", biodata files can't truly be cross-platform because
of the variation in end of line characters on unix/win/mac. Ideally, our code
should work regardless of EOL variation, but it's a good idea to test it to be
sure. 

Since it's possible to re-generate a file with different line endings, I think
the way to go here is to declare that all files in biodata should be
unix-style. We could then have script (within biodata/bin maybe) that can
generate win and mac versions on-the-fly as needed for testing. Would be cool
if we could integrate this in the make test so that it win-ifies or mac-ifies
all biodata files as necessary before starting the test. Or perhaps it could
work as an filtering layer when reading the file.

Maybe this isn't that big of a deal since most modern-day archiving utilities
will fix line endings as appropriate when uncompressing the tarball/zip file.
However, it would still be worth having a way to generate different versions of
a file for testing purposes since it's possible to write code assuming files
have unix-style line endings.

Steve

--- Jason Stajich <jason@cgt.mc.duke.edu> wrote:
> 
> > download from CVS need to download a second module -- although
> > couldn't this be solved through some aliasing or so (Jason?).
> >
> Done.  Each project now has a new cvs module called 'biodata' which is a
> symlink to /home/repository/biodata/biodata
> 
> Each project also has a new entry in CVSROOT/modules which allow you to
> checkout all the necessary modules in a single cvs checkout.
> These are all named bioXX_all so:
> 
> % cvs -d:ext:USERNAME@open-bio.org:/home/repository/bioperl co bioperl_all
> % cvs -d:ext:USERNAME@open-bio.org:/home/repository/biopython co
> biopython_all
> % cvs -d:ext:USERNAME@open-bio.org:/home/repository/biojava co biojava_all
> 
> One of the dirs in there is called biodata.  Which you can alternatively
> checkout with the cvs module name 'biodata'.  It is also accessible in
> from /home/repository/biodata.   Chris -- we'll need to add it to the
> rsync list and/or verify that rsync is respecting the symlink.  If it
> isn't we'll have 3 copies of this dir on the anonymous CVS machine - but
> perhaps that isn't a big deal anyways...
> 
> All projects can make changes to the biodata directory (note the need for
> a common group below) and the data will be shared across the projects.
> 
> Note 1:
> 
> People must be in the obfcommon group -- we'll have to make a merged group
> at some point or else you'll get warning message when you try and checkout
> this dir.  Someone else can sort this out please.  Too bad unix doesn't
> allow groups to be made up of other groups...
> 
> Note 2:
> Additionally,  emails will only get sent to the lists if the toplevel
> CVSROOT for the modules (i.e. /home/repository/bioperl/CVSROOT) has been
> configured to send mail (as in cvs commit mails to bioperl-guts-l).
> 
> Note 3: Whenever Chris switches the CVS repository machine open-bio.org
> will become dev.open-bio.org
> 
> Note 4:
> project admins may want to edit the cvs module list (do this:
> % cvs -d /home/repository/bioXXX co CVSROOT/ modules
> % cd CVSROOT
> % vi modules
> % cvs commit -m "added XX to the list"
> % cd ..
> % cvs release -d CVSROOT
> )
> 
> -jason
> -- 
> Jason Stajich
> Duke University
> jason at cgt.mc.duke.edu
> 
> _______________________________________________
> Open-Bio-l mailing list
> Open-Bio-l@open-bio.org
> http://open-bio.org/mailman/listinfo/open-bio-l


=====
Steve Chervitz
sac@bioperl.org

__________________________________________________
Do You Yahoo!?
Yahoo! Finance - Get real-time stock quotes
http://finance.yahoo.com