Meant to reply-all, but.... ---------- Forwarded message ---------- From: Phillip Moore <w.phillip.mo...@gmail.com> Date: Wed, May 30, 2012 at 2:38 PM Subject: Re: Issues with VERSION migration in a module suite To: Marvin Humphrey <mar...@rectangular.com>
On Wed, May 30, 2012 at 11:05 AM, Marvin Humphrey <mar...@rectangular.com> wrote: > On Wed, May 30, 2012 at 7:40 AM, David Golden <xda...@gmail.com> wrote: > >>> I really wish that versions and dependencies were expressed at the distro >>> level rather than the module/package level. Allowing multiple version >>> numbers >>> within an atomic unit of installation doesn't make sense. >> >> I agree. Now all we need is a DeLorean, a flux capacitor and 1.21 gigawatts. > > *chuckle* > > It's possible to overlay a package management system on top of Perl5/CPAN > which uses distro-level dependency specification. I don't know the gory > details, but isn't that what apt-get, rpm, etc do? Funny you mention that -- that is *precisely* how my own package management system for perl works :-P > It would also be easy for a build system to enforce unified versions at distro > build time. It's not possible to prevent tarballs from appearing on CPAN > which violate this provision, but the issue is not so hopeless or insoluble > that we try to understand the problem better. > >> I outline the concept of a well-formed distribution here: >> http://www.dagolden.com/index.php/308/packages-modules-and-distributions/ >> >> I would go further and say that I recommend that whenever possible, >> all modules in a distribution should have the same version number, to >> avoid the sort of distro/module version sync problems that people have >> experienced. And the irony is that I am trying to *convert* this distro to the very convention(s) that David recommended. Deleting the distro and waiting might let me get away with fooling PAUSE, but someone, somewhere will get burned by the individual module versions decrementing, so... Major version 500 it will be. Thanks for verifying my sanity, or lack thereof....
