[Biopython-dev] ApplicationResult and generic_run obsolete?

[Bartek Wilczynski]

On Tue, Jul 7, 2009 at 12:04 AM, Brad Chapman<chapmanb at 50mail.com> wrote:

>> I'm fine with deprecating ApplicationReslut.
>
> Bartek, you just won the typo of the month contest hands down.
>
Well, It's always motivating to see that people actually read your
posts carefully ;)

> My vote is to document using subprocess and avoid creating our own
> wrapper. No one has to learn a Biopython specific API for running
> programs, and subprocess provides plenty of flexibility to get stdout,
> stderr and return codes. For places where we feel like using subprocess
> is tricky, additional documentation within Biopython should help those
> encountering it for the first time. This gives us more time to work
> on biology problems, and leaves the running programs problems up to
> the greater Python community.

 well, having such a documentation would be a great thing. I've just gone
through the docs for subprocess module and it seems to be the layer unifying
all those crazy different ways of spawning processes. It's a shame  I somehow
missed that it's there since python 2.4... So now, after doing my homework and
checking what has been going on in python since 2004, I think that Brad's idea
is better. We have dropped support for 2.3, so we can try to move from
Application.generic_run to subprocess.Popen instead of trying to
provide our own
wrapper.  We just need good docs.

cheers
  Bartek