This one time, at band camp, Matthias Klose said: > Debian Bug Tracking System writes: > > As far as I can tell there is no actual bug here: the lib64c-dev > > provides exists in unstable and the dependency appears to do what I was > > expecting it to do. > > yes, but you are required to depend on a real package as well, not > just only on a virtual package.
I sort of diagree here. AIUI, the reason for the rule to depend on real || virtual is to have consistent behavior when doing automated packaging work - the autobuilders will install the same package from a given pool of choices and so on. This is to cover things like awk, which is provided by gawk, mawk, etc. This is not that sort of situation - if I read things right, there is only one package on each architecture providing this virtual package - the virtual package is only there to provide an easy way to get that functionality across architectures without hard-coding all of the architectures individually. If there were more than one package per architecture providing this virtual package, then the dependency would need to be adjusted to provide consistent behavior. But at first blush, we don't seem to be there. But I may be missing something. I'm willing ot be convinced, but this seems like one of those corner cases where following policy to the letter is actually less helpful than the current arrangement. -- ----------------------------------------------------------------- | ,''`. Stephen Gran | | : :' : [EMAIL PROTECTED] | | `. `' Debian user, admin, and developer | | `- http://www.debian.org | -----------------------------------------------------------------
signature.asc
Description: Digital signature
