On Thu, Oct 30, 2008 at 05:39:53PM -0500, Karl Berry wrote: > > NEWS Changelog README blank. > Optional I think. > > I agree it's optional, but it would be helpful -- it doesn't matter for > savannah, but for my GNU evaluation purposes, where this tool would also > be of great assistance, it would help :).
I think that such tests need to be done in another tool (which can be a wrapper around Savannah's). More tests entices admins and users to enforce them, and I think that's precisely what we did wrong with Savannah: introduce more and more copyright-linting tests that RMS now enforces us to do, and that nobody wants to perform, yet believe that we need to do. Looking at 2001/2002 submissions it's clear that Savannah hackers were more lax in the past, which IMHO helped reviewing projects without ridiculous delays. I'm probably responsible for introducing some of these checks. Looking at Debian it's also clear that ftp-masters introduce more and more tests for first package uploads (my package just got rejected for 2 issues that no existing package currently deal with...), and their team became particularly slow and help-begging just like we're now. So let's keep the Savannah submission tool minimal and to the point :) -- Sylvain
