Joey Hess writes... > This seems less than optimal to me. If lsb packages were required to end > with ".lsb", then there would be no worries about source packages being > confused with lsb packages, and no worries about alien or another tools > generating something that looks broadly like a lsb package but was not > intended to be one, and there would be less special-purpose code in > alien.
You have a really good point. Can someone please explain why we currently do it the way we do? Another thing I've been thinking about is transition to a future lsb package format. Does naming packages .lsb make that transition any easier/harder? If this issue is holding up the adoption of lsb on non-rpm systems we need to address it ASAP. > Really I wish you could go one step further and munge the format in > some unique but harmless way[2] such that file(1) could recognize a lsb > package no matter what the filename, but I suppose it's too late for > that. I don't think it's too late for future versions of the spec and this seems like a really good idea. It would be nice for file to be able to determine the package format version of lsb packages. Knowing what version(s) of lsb they claim to be certified for would be nice too(but maybe more difficult). The distribution package tools could decide what they want to do with the package. -- Matt Taggart Linux Development Lab [EMAIL PROTECTED] HP Linux Systems Operation
