ConfigParser module in 1.52 is buggy?

David Bolen db3l at fitlinxx.com
Wed May 30 17:51:28 EDT 2001


Thomas Wouters <thomas at xs4all.net> writes:

> Okay, all valid reasons for staying at 1.5.2, and I don't blame anyone for
> doing so. The question at hand is, though, whether you would uprade to a
> 1.5._3_ ?

Fair question.  In terms of a general upgrade (and in the absence of a
specific enumeration of chnages), my inclination would probably be no,
since the fixed overhead for upgrading the installed base is largely
independent of what I'm putting out there.

But there's a big caveat - if the changes were well contained and/or
specifically addressed something I needed fixing, it might well be
something I would use.  Pushing out a small delta to the installed
version (and in my case, the actual number of bytes to distribute come
into the picture too) is often more attractive than switching major
releases if only because there's less change and risk involved.  Or if
it fixed a problem without invalidating other extensions (which might
not yet be ready for the latest 2.x).

So in the context of having an updated ConfigParser (where this all
started I guess), yeah, I'd probably push that out on top of 1.5.2
before jumping to 2.x if that's where my problem lay.

But there's no simple answer - most of the imputs vary based on the
amount of change (so moving to 2.x will have higher costs and risks),
but that has to be weighed against eventually going there, so at some
point you cross the line to just taking the leap to the latest version.

I just wanted to raise awareness that it's not always as easy as
saying "1.5.2 is ancient - why haven't you upgraded" as if that was
the simplest thing to do :-)

--
-- David
-- 
/-----------------------------------------------------------------------\
 \               David Bolen            \   E-mail: db3l at fitlinxx.com  /
  |             FitLinxx, Inc.            \  Phone: (203) 708-5192    |
 /  860 Canal Street, Stamford, CT  06902   \  Fax: (203) 316-5150     \
\-----------------------------------------------------------------------/



More information about the Python-list mailing list