[Biopython-dev] Python 2.3 support

[Peter Cock]

On Tue, Apr 21, 2009 at 1:35 PM, Brad Chapman <chapmanb at 50mail.com> wrote:
> Hi Peter;
>
>> As we've been warning for the last couple of releases, Biopython 1.50
>> should be the last release to officially support Python 2.3.  No one
>> has complained yet, but they may not have noticed. I suspect there may
>> be people out there using a local Biopython installation on an old
>> Linux/Unix computer where the system Python is rather old. For
>> Biopython 1.50 I added a warning to setup.py when run on Python 2.3 so
>> that may get more attention.
>
> Are we getting a lot of feedback that we need to keep supporting these
> old versions? 2.3 was released in 2003, 2.4 in 2004, and 2.5 in 2006.
> This means people who need anything prior to 2.5 haven't updated in over
> 3 years. I understand the problem of non-responsive sysadmins and what
> not. However, we only have so many cycles for testing and coding; is it
> worthwhile spending some on these problems?

Until recently I have a very strong personal interest in keeping
Biopython running on Python 2.3, so I never regarded this as "wasted
cycles".

My personal Windows machine ran Python 2.3 and MSCV 6.0.  In order to
update the python version and continue to compile Biopython, I would
also have had to replace the compiler etc. and the hard drive was
pretty full so this didn't appeal.  I have recently been trying Ubuntu
on this machine instead (on a second hard drive).

For reference, my current (only) Windows machine (at work) has Python
2.3, 2.4 and 2.5 for which I use mingw32 to compile Biopython (same
setup as Michiel), plus Python 2.6 for which I'm using Microsoft's
free VC++ 2008 Express Edition from
http://www.microsoft.com/express/download/

> Practically, I'd be for dropping 2.4 support in the next release and
> being a bit more aggressive in general on moving upwards and onwards.

I wouldn't support that.  I would insist on giving at least one
release's notice as a minimum.

Peter