in reply to On installing modules with CPAN.pm

I think the lesson here is this: We all need to remember that while CPAN is a very comprehensive archive, the modules within may not be entirely comprehensive in the features we may think they're supposed to provide

I would say it's more like "be sure to read the POD and look at the dependencies of a module before you go to the effort of installing it." It looks like one of the first statements in the POD is that it only supports MySQL and Pg. POD and dependencies are available for pretty much every module on either of the CPAN search engines, and I never install anything without checking it out on there first.

However, your install difficulties teach another lesson: be careful about adding dependencies on XS-based modules if you want your module to run on Win32.

The flip side of that is that you have to realize that many people are not targetting Win32. It may have a majority of desktops (though surely not 98% as you say), but Linux/BSD is the most common platform for Perl development and has been for years. Win32 support in Perl is amazingly good, but the users on that platform are still a minority, and have to expect a bit more difficulty getting some things to work. It's like being a gamer on a Macintosh: the best stuff typically gets ported, but not everything, and sometimes you have to wait a little while.

Replies are listed 'Best First'.
Re: Re: On installing modules with CPAN.pm
by Stevie-O (Friar) on Apr 22, 2004 at 04:08 UTC
    You are entirely correct. I had missed the line stating its limitation to MySQL and Postgres -- I confess I didn't read the entire POD through. In my defense, I did look -- but I was looking for a BUGS or LIMITATIONS type of section, and the limitation is listed in the DESCRIPTION section.

    Win32 support in Perl is nothing short of incredible. When I was able to follow *identical* directions (from perlxstut) on Windows and Linux to flawlessly produce an XS module from h2xs, I nearly fell out of my chair.

    And, yes, I am in the minority. But it just grates me when something won't work because they did something like use `pwd` instead of use Cwd, or something depends on a case-sensitive filesystem (there was another meditation on that recently).

    However, I wonder I would have been spared the IO::Zlib trouble I suffered on any other OS...

    --Stevie-O
    $"=$,,$_=q>|\p4<6 8p<M/_|<('=> .q>.<4-KI<l|2$<6%s!<qn#F<>;$, .=pack'N*',"@{[unpack'C*',$_] }"for split/</;$_=$,,y[A-Z a-z] {}cd;print lc