On Sat, 2006-09-16 at 16:36 -0500, Michael G Schwern wrote: > --->>>*** The primary purpose of the install chain is to install modules > ***<<<--- > > Not to test the module.
Is it too late to speak out against this? As a 'packager' of CPAN modules for the consumption of a linux distro, we rely, even depend, on CPAN authors writing good tests. Not all do - some check that you can load the module or that their signature works (woohoo) and move on. But I cannot stress enough the value of good tests and how much a part of the install chain they need to remain. (Mind you, 'good tests' are those that actually test the functionality of the module to some base degree.) It is through the tests that we locate modules that may compile and install fine, but that don't actually function on amd64's, or sparc's (running linux, not to confuse with traditional solaris sparc's), or hppa's. If you need the example RT tickets for these I can dredge them up, but please do not encourage the abandonment of testing. It's at this point that someone points out that Michael (Schwern) was targeting TODO tests, but with such an open statement I couldn't let it just sit out there. Now, encouraging module writers to include actual functionality testing or no testing at all, that I could agree with. Thanks for the eyes, ~mcummings > > > -----o()o---------------------------------------------- > Michael Cummings | #gentoo-dev, #gentoo-perl > Gentoo Perl Dev | on irc.freenode.net > Gentoo/SPARC > Gentoo/AMD64 > GPG: 0543 6FA3 5F82 3A76 3BF7 8323 AB5C ED4E 9E7F 4E2E > -----o()o----------------------------------------------
signature.asc
Description: This is a digitally signed message part
