On Fri, Dec 27, 2013 at 09:02:49PM +0100, Jerome BENOIT wrote: > thanks for the reply. > On 27/12/13 20:45, Steve Langasek wrote: > > On Fri, Dec 27, 2013 at 07:53:32PM +0100, Jerome BENOIT wrote: > >> Hello List, > >> I am maintaining a package, FireHOL not to name it, which basically > >> contains bash sources. So it Architecture was set up to all by one of > >> my predecessor. Meanwhile, kfreebsd support emerged. > >> As FireHOL is meant to manage iptables, it is de facto meant for linux: > >> http://qa.debian.org/debcheck.php?dist=unstable&package=firehol (bottom) > >> Therefore, may I restrict Architecture to linux-any ? > > You *may* do so, but why bother? > Because debcheck complains. Then perhaps debcheck should be fixed. > The package already depends on the > > architecture-dependent iptables, and is therefore uninstallable on kfreebsd. > > So there doesn't seem to be any harm to having the package be Architecture: > > all. > Will setting Architecture to linux-all create more harm ? It will increase the size of the archive with redundant packages, for very little reason. In the case of firehol, the harm is small, of course. But we wouldn't want to do this /as a rule/. Consider the not-improbable case of a large Arch: all data package for a game that fails to build on kfreebsd. Should the data package be marked Architecture: any, so that it's not available on architectures where the game engine package is absent, if this would cost us a gigabyte (or more) of space in the archive? No, it shouldn't; which demonstrates that debcheck-cleanness is a wrong metric to use. -- Steve Langasek Give me a lever long enough and a Free OS Debian Developer to set it on, and I can move the world. Ubuntu Developer http://www.debian.org/ slangasek@ubuntu.com vorlon@debian.org
Attachment:
signature.asc
Description: Digital signature