[Bioperl-l] LocusLink IO

Paul Boutros pcboutro@engmail.uwaterloo.ca
Tue, 10 Dec 2002 15:59:49 -0500 (EST)


Hi,

I've tried the diff process entirely on a windows box, and entirely on a
Unix box, and there don't appear to be any real differences between the
files.  The ordering of lines in my version (from NCBI) is slightly
different than in the test version, and the version I used has some
additional information (not *new* tags, just more of the same ones) than
the test one.

But, as best as I can tell, this isn't an \r\n problem -- that was an
artifact of me extracting one on a Windows machine and the other on a Unix
server when I did the diff.

I also re-ran my minimal test-case today with a current CVS tarball, and
no changes.  I've also verified it with updated downloads from NCBI and on
a couple of different windows machines.

Any other test I can run to help with this?

Paul

On Tuesday December 3rd You Wrote:

looks like a \r\n problem.  did you happen to untar this file on a windows
or non-os-x-mac box, or do something else with it on one of these
platforms?  i got nipped by this problem using winzip to decompress a
.tar.gz once.  when i transferred the contents (a bunch of c source files)  
over to a linux box, the compiler had problems reading them b/c the
newlines were mangled.

i would try doing the whole decompress/test process on a linux box, or see
if you can find a way to do a binary decompress on the files if you're
stuck doing it on old-mac/windows. good luck.

-ad

On Mon, 2 Dec 2002, Paul Boutros wrote:

> I followed the suggestion (I think from Allan Day) of extracting &
> diff'ing record 27 from the file.  This is what I got:
> 
> ===========================
> pcboutro@engmail[5] diff testLL.txt LL-sample.seq | more
> 43a44
> > UNIGENE: Hs.75741
> 46d46
> < UNIGENE: Hs.75741
> 52a53,54
> > BUTTON: homol.gif
> > LINK:
> http://www.ncbi.nlm.nih.gov/HomoloGene/homolquery.cgi?TEXT=26[loc]&TAXID
> =9606
> 133a136
> > UNIGENE: Hs.121521
> 136d138
> < UNIGENE: Hs.121521
> 144a147,148
> > BUTTON: homol.gif
> > LINK:
> http://www.ncbi.nlm.nih.gov/HomoloGene/homolquery.cgi?TEXT=27[loc]&TAXID
> =9606
> ==========================
> 
> The file does indeed terminate with a >> but I didn't see any empty
lines
> after that.  I'll submit this as a bug report along with everything I've
> tested so far.
> 
> Paul
>