On Thu, 08 Feb 2001, Wichert Akkerman wrote:
> FWIW, if this change gets accepted all my currently Debian native package
> will suddenly no longer become Debian native and get an empty diff file.
Yes, that would be expected.
Personally, I'd rather see empty .diff.gz's than people giving up their
debian revision fields when they are using them properly (which you do).
> I think trying to encode Debian nativeness in the version number is
> trying to put information in there that it is not supposed to handle,
> and needlessly complicated.
Well, it is really a matter of choosing the lesser of various evils.
1. Leave things as is (bandwidth waste and archive mess, and maybe
license violations, as some require diff + pristine distribution)
2. Forbid debian revision in native source (empty .diff.gz)
3. Just detect a non-native->native transition (cannot catch
bogus first uploads)
I've not written much about 3 at all, basically because I have no idea how
complicated it is to add that to katie (and lintian won't be able to help in
that case, as well).
Still, it is highly probable IMHO that first uploads in the wrong source
format are also an issue, given how easy is for our tools to generate native
packages because the maintainer is not paying attention to what he's doing.
If this discussion shows that it would be better to go with 3 than 2, I'll
remove the proposal.
There is also a fourth possibility. Fix the tools so that one is forced to
tell them what format must the source package be (and it must have NO
default, otherwise we'll have the problem again). I think this one is too
painful and disruptive to consider.
--
"One disk to rule them all, One disk to find them. One disk to bring
them all and in the darkness grind them. In the Land of Redmond
where the shadows lie." -- The Silicon Valley Tarot
Henrique Holschuh
