Re: Build-Depends listed as a should not a must?
On Sun, Feb 16, 2003 at 03:11:57AM +0100, Josip Rodin wrote:
> On Sat, Feb 15, 2003 at 07:04:32PM -0600, Jamin W. Collins wrote:
> > Non-conformance with guidelines denoted by should (or recommended)
> > will generally be considered a bug, but will not necessarily render a
> > package unsuitable for distribution.
> >
> > This means in theory that a developer could opt not to list
> > Build-Depends and this would be acceptable, right?
>
> Given that omitting non-build-essential build-depends will get them a
> serious bug filed from the first buildd maintainer to notice it (this
> is why I said normal or important _by default_ above), no.
Unless it has no build-dependencies at all, in which case andrea will
take care of it as far as the autobuilders are concerned and serious
bugs aren't needed.
> Personally I really don't see why's there all this fuss[1] about this
> particular detail. I've yet to see a maintainer refuse to list a valid
> build-dependency in their control file. Even if anyone would think of
> doing that, I'm sure they would be persuaded to change their mind very
> shortly after it was noticed. }:)
>
> [1] #87510, which will incidentally be closed with the next upload.
Uh, the objections have not been resolved. The shell one-liner Anthony
posted in http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=87510&msg=102
currently prints 246, so that'll be at least 246 release-critical bugs
automatically created by this change in policy. Can we get those fixed
first, please?
--
Colin Watson [cjwatson@flatline.org.uk]
Reply to: