Bug#612610: lintian: should suggest switching to 3.0 (quilt)
Ron <ron@debian.org> writes:
> Since there were strong arguments against this in the past, from more
> than one person, and since nothing has changed since, including members
> of the release team from time to time still voicing their
> dissatisfaction with the 3.0 quilt format and intent to continue not
> using it, it doesn't seem like simply "having the same debate all over
> again" is going to be very productive unless "hope the people who
> objected last time can't be bothered repeating themselves" is considered
> "productive".
> Maybe it would be better if the people who want to push this actually go
> and review the objections from last time, and come up with a new format,
> or with changes to this one that actually ameliorate the objections that
> were already made.
I believe this has already been done to some extent, speaking as one of
the people who stayed with 1.0 at the time of the previous format. The
addition of single-debian-patch and local-options resolved my concerns.
I would like to see this raised on debian-devel again because I think two
things have changed with time, not because I'm just hoping that people
won't repeat the same arguments and a different argument might win by
default. (And, besides, how long have you watched debates in this
project? :) Since when do we ever leave any argument to win by default?)
First, the addition of those options, which may not be common knowledge,
makes 3.0 (quilt) fairly close from a packager's perspective to 1.0,
except with cleaner separation in the source package format between the
Debian files and the upstream files, support for good compression formats
(I'm now using xz for basically everything, since it's pretty much
superior to gzip and bzip2 in all respects and is being actively
recommended by at least one upstream maintainer of gzip), and a fix for
the ambiguity between native and non-native packages.
Second, some (not all, but some) arguments against 3.0 (quilt) seemed like
arguments against how Raphael presented it, arguments against the way it
was being discussed, or arguments from a position of lack of benefit.
Those things tend to change over time. We've now had it for some years,
more people have had a chance to try using it, some of the edges have been
smoothed over, it's clearly not going anywhere (whether one was
unconvinced of the initial benefits or not), and irritation from previous
discussions has subsided. I think there's a substantial possibility that
this could have resulted in people quietly changing their minds, or at
least being open to changing their minds.
--
Russ Allbery (rra@debian.org) <http://www.eyrie.org/~eagle/>
Reply to: