I just noticed a new Data::GUID/UUID release to CPAN. Since I use it heavily in Handel so tend to look at such modules change list just to keep up with things.
1.148 Thu Nov 16 10:21 2006 - Debian has chosen to distribute their own Data::UUID, which has a different interface and breaks other modules. They also use a grossly-inflated version number, which means that this version number must be inflated to allow modules to rely on the CPAN Data::UUID properly.
Tests added to EXPLICITLY assert the one known difference between genuine Data::UUID and Debian's ersatz version in libossp-uuid-perl.
Thanks to ADAMK for bringing this to my attention.
Re:It's the License
stu42j on 2007-03-08T19:19:51
Actually, the bug in question was supposed to be fixed in the Debian package. There is a patch in the source file but somehow it doesn't make it into the final package.
It should really be fixed upstream.
This does sound like a very sucky situation, but I think it's hard to characterise it as the fault of the Debian maintainers.
Try as I might, I can't construe this situation in such a way as to apportion blame to the Debian maintainer. You might claim that the Debian maintainer should have avoided packaging the Perl module part of ossp-uuid, but that would break the expectations of people who use ossp-uuid. As it happens, the Debian package instead breaks the expectations of people who use CPAN Data::UUID. But the precise breakage is one that wouldn't have been predictable to the Debian maintainer.
What do you think the Debian maintainer could or should have done differently?
Re: Why do vendors do this crap?
jk2addict on 2007-03-08T20:37:17
I just want to know who to flog heavily with a wet noodle when someone files an RT about "uuid generation not working, I'm getting errors in DBIx::Class::UUIDColumn", and wasting time going "Is Data::UUID installed? Yes. Hmmm....why is it not working...oh wait...which Data::UUID...the Debian version or the CPAN version..."
I realize it's not totally the maintainers fault...but at the same time, some red flag should have gone off when someone says "well, we can't package the CPAN version for licencing reasons, so lets just use this one that's name the same thing instead". Brilliant!:-)
Hey, while you're at it, just rename the Ruby interpreter to Perl. That should work....it's named the same and mostly compatable right?
It just feels like a Redshat-ism. It would be even worse if the @INC is jackered, and installing the CPAN version still doesn't get used because it falls too late in @INC compared to their version.Re: Why do vendors do this crap?
sigzero on 2007-03-09T12:51:09
Did they try and contact the author?Re: Why do vendors do this crap?
rjbs on 2007-03-09T16:20:03
No one contacted me. They may have tried to contact Alexander G., whom I've been unable for years.