On Sat, Jul 24, 2004 at 06:35:44PM +0200, Thomas Klausner wrote: > Hi! > > On Fri, Jul 23, 2004 at 03:29:10AM +0100, Tim Bunce wrote: > > > The DBI gets 9. The one failure is permissions_ok: > > > > > permissions_ok (i.e. all files are read/writable by extracting user) > > > > Why is that a kwalitee issue? I don't think it warrants impacting the kwalitee. > > First: Currently there is no real definition of Kwalitee. Now it consists of > stuff that's easy to test and that I'm interested in.
Understood. > Personally, I'm annoyed by dist that I cannot remove after installation. > If files are read-only, I'll have to do extra steps during deleting. So I > like dists which no read-only files. Which is why it's a Kwalitee indicator. > > If we (whoever is interested in this issue) deciede that containing > read-only files is not a big issue, than we can drop this indicator (or any > other...) > > But IMO a polite maintainer should make it easy to remove a dist after > installation. I wasn't aware that there was a problem. No one has mentioned in the many many years the DBI has shipped with read-only files. > > p.s. It'll cause problems for anyone using a source code control > > system that keeps files read-only - like RCS and CVS do. > > That's the reason the DBI has many read-only files. > > (I use svn now, but used to use RCS.) > > I think it was also a "problem" with Module::Build / ExtUtils::Manifest. See > here: > http://rt.cpan.org/NoAuth/Bug.html?id=4124 I interpret that as meaning the permissions_ok is primarily driven by the version of MakeMaker used by the author. MakeMaker used to explicitly make the files read-only, now recent versions don't. It's hardly fair to reflect that in the kwalitee metric. > But again: defining Kwalitee is defintly not something I want to do on my > own. I only provided first suggestions ... I expect the definition of Kwalitee to become an uncomfortably hot topic. Meanwhile, I'd suggest you remove permissions_ok :) Tim.