[sf-perl] Mouse, again?
doom at kzsu.stanford.edu
Fri Mar 2 11:34:33 PST 2012
Dodger <el.dodgero at gmail.com> wrote:
> I've never found a reason to use either.
> But even if I did I would never use them in anything intended for sharing on CPAN.
> It's one of those "do into others" things.
Well, I know what you mean. It's a design decision you've got to make
up your mind about when you work on a CPAN module. I've put some
stuff out there with a heavy stack of dependencies on the theory that
we all know how to use CPAN.pm, CPANPLUS, and/or cpanminus these days,
and then gotten complaints from users who were apparently into doing
it all manually, downloading tarballs and typing "perl Makefile.PL"
It's something you need to at least think about: all perl users are
not exactly members of "the perl community" and we often don't have a
good grasp of what their lives are like.
(I'm also an advocate of a fat perl core... The perl porters seem to
have convinced themselves that "let them eat CPAN" is a valid response
to user problems.)
> (Or ten thousand different incompatible test harnesses that I end up with installed because people randomly decided all their tests for a pure perl module needed to be done using Test::Wankish::Obscure::Nonsense)
Well, it's at least possible for cpan authors to classify a dependency
as a test requirement, so that your installation tools will have it to
run the tests, without actually installing it.
("Incompatible"? They don't actually break each other, do they?)
More information about the SanFrancisco-pm