On Tue, Feb 17, 2015 at 01:30:31PM -0500, Mark H Weaver wrote: > Where is this documented? I looked for it in the "Packaging Guidelines" > section of our manual, and also the subsection "Package Naming", and > couldn't find it. Anyway, even if such a guideline were documented, I > would oppose it as a general rule.
What I meant is the following paragraph, which is indeed ambiguous: "Both are usually the same and correspond to the lowercase conversion of the project name chosen upstream, with underscores replaced with hyphens. For instance, GNUnet is available as @code{gnunet}, and SDL_net as @code{sdl-net}." The "upstream project name" is not clearly defined. I wrote this paragraph quite some time ago. Normally, I understood the "upstream project name" as the tarball name; usually, both are the same so there is no problem. The idea was to have a mainly mechanical process to avoid per-package discussions. So in general, I think one _can_ choose the tarball name, but if it is too weird, that may not be a good idea. For the package in question, it is referred to everywhere on its homepage http://dynamo.iro.umontreal.ca/wiki/index.php/Main_Page as "Gambit". For instance at the bottom: "Gambit is Copyright ...". In the middle, it speaks of "Gambit-C", as in "Gambit-C is a version of the programming system ...". There does not seem to be any other version, though, so in practice, "Gambit" and "Gambit-C" appear to be synonyms. All in all, I think that all three package names "gambc", "gambit" and "gambit-c" are defendable. We may follow the path of lowest resistance and keep the name as it is now. Andreas