Re: Is Ubuntu a debian derivative or is it a fork?
On Thu, Jun 02, 2005 at 12:25:01AM -0400, Joey Hess wrote:
> Matt Zimmerman wrote:
> > I see no need to argue about whether Ubuntu should push; the patches
> > are all there in an easily accessible tree, and it would be trivial to
> > pull the patches and push them someplace else if that's desirable.
> Please take a look at the current Ubuntu 1.6 MB diff for base-config
For the record, about 93.5% (by byte count) of that consists of largely
autogenerated debian/po/ changes that are almost certainly best ignored,
apart from the odd new translation that can be picked out and submitted
to translation teams for review after merging the corresponding strings.
The parts of that that aren't s/Debian/Ubuntu/ branding would mostly go
away if I could even figure out where to start on making apt-setup not
quite so hardcoded to a single distribution. I would love to get that
> (the split diffs are useless in this case), and tell me how you consider
> this to be "easily accessible". There are some base-config improvements
> in here that could benefit others, or at least other derived distros,
> such as making it only expect one CD, but not done in a generic or
> reusable way and they're all mashed up with tons of Ubuntu specific
> For what it's worth, I've completly given up on separating the parts
> that are applicable to Debian from the parts that aren't. I have some
> hope that Colin will manage to merge some of it into the Debian package,
> since he's been doing a lot of work on merging in Ubuntu's changes to
> d-i, but if that doesn't happen soon, Ubuntu will be left with this
> massive patch to forward port as I make huge planned changes to
> base-config post-sarge.
base-config is bad, yes; it is probably my biggest headache. I've
already had to merge significant changes from Debian multiple times, and
I know exactly how painful it is.
I've been making a start on the merge today, but I need help if I want
to have a hope of ever doing anything with apt-setup.
> If Debian treated our upstreams this way, I'd be suprised if we ever got
> any patches accepted upstream.
I'm sorry that base-config has been handled poorly, and I'll try to make
more time for it. I think it's a particularly harsh example, though,
since most upstreams don't have a huge script full of code almost
entirely hardwired to the repository layout and policy (e.g. contrib,
non-free) of a single organisation. In cases where they do, we tend to
make huge changes to them or just disable them entirely, depending on
how much time we have available to make the code better and more generic
versus just making it work so that we can ship something.
Yes, fixing that situation could be very useful for other derivative
Colin Watson [email@example.com]