[Bioperl-l] Really weird bl2seq behavior

[Stefan Kirov]

I know this is off topic, sorry about this...
I jus wonder if anyone else has run into the same or similar problem:
I am bl2seq from perl, using system. Done from perl I get slightly 
different results each other time, i.e. even runs give me the right 
alignment, odd ones- no. The difference is not huge, but still 
significant. It is also important that the output file is temporary and 
I write multiple times and delete it afterwards. I guess this is might 
be part of the problem.
Also I am using -o to create the report. Doing exactly the same thing 
from command line always gives the proper alignment. So it has to be a 
perl or bl2seq bug (perl 5.8.1) or some kind of incompability...
If I use redirection instead of -o option it is OK, so I guess it is a 
bl2seq bug.
I am completely confused and I hate to let this go... Any ideas?
Thanks!
Stefan