On Sat, 19 Jan 2008, Mark Brown wrote: > On Sat, Jan 19, 2008 at 08:30:36PM +0100, Raphael Hertzog wrote: > > On Sat, 19 Jan 2008, Mark Brown wrote: > > > > I wonder if the header line could > > > even be autogenerated since I'd expect 90% of packages are going to want > > > to use the same template? > > > If people grab symbols files from qa.debian.org/cgi-bin/mole/seedsymbols > > then they get the right the package already. The whole header line can't > > be autogenerated, you need indicate the library soname somewhere and > > currently the parser requires a dependency after the library name. > > The library soname can be figured out from the libray itself in most > cases.
Yeah, but many packages provide multiple libraries in the same package... and I don't want too many special case. > > > > shouldn't be used, I'd rather make that explicit than let them generate > > > > a > > > > dependency on a recent version of the lib which might go unnoticed. > > > > Yes, me too but I didn't see any way to do that. > > > At least, you know how now. > > Hrm? I thought you said the support for doing that (except by > specifying an invalid version) wasn't implemented? I explained you how to associate a symbol to an alternate unsatisfiable dependency. This is what I would do if I had public symbols that I didn't want people to use. You can even generate refer people to the README.Debian of the the library... with an unsatisfiable dependency like "use-bad-symbols-see-readme-debian-of-lib64z1". Cheers, -- Raphaël Hertzog Le best-seller français mis à jour pour Debian Etch : http://www.ouaza.com/livre/admin-debian/
