Raphael Geissert <[email protected]> writes (in September of 2008): > I've just implemented another check, this time for the copyright file > which basically tries to detect the ones that don't provide "any" > licence information at all. > > I've tried to make it as less prone to FPs as possible and even ran > lintian on over 16410 packages which are either available at sarge, > etch, testing, unstable, or experimental and reviewed some of the > packages for which the tag was emitted and so far have not found any > false positive [1]. > > [1] There are some packages that just quote a file found on the orig > tarball like (but please note that it is all what they say): "you may > copy, modify, and redistribute freely." - evolver_2.30c-1 And there are > some gitwares around being matched too: flying, allegro-demo, > allegro-demo-data, allegro-examples.
I must admit that seems like a false positive to me. I don't recall how much we discussed this patch originally when you first posted it, but it was sitting in my pending mail folder waiting for me to look at it. Looking it over, it makes the assumption that any package with a license will mention the word license or one of the standard license names. But while it lacks proper legal wording, I believe that the project has historically accepted packages with license statements like the one you quote above, even if it doesn't have the proper phrasing for a legal license. Basically, I would prefer to follow the lead of ftpmaster here. If they think that a test like this would catch more issues before they see them, or would be useful to them in checking packages, we should tackle it. But since this is the sort of test that tends to produce the most push-back (non-technical, matching English text, a heuristic), I'd prefer to have confirmation from them that this looks good before we add it. -- Russ Allbery ([email protected]) <http://www.eyrie.org/~eagle/> -- To UNSUBSCRIBE, email to [email protected] with a subject of "unsubscribe". Trouble? Contact [email protected]
