Re: FreeBSD patch for dpkg?
On Thu, May 01, 2003 at 06:46:47AM +1000, Rudolph Pereira wrote:
> firstly, thanks for the explanation
> On Tue, Apr 29, 2003 at 03:08:23PM -0600, Joel Baker wrote:
> > What there is, is on the Ports page, mostly. Keep in mind that there are
> > (at least) two separate efforts that are being kept track of on here: the
> > (stagnant?) FreeBSD port, and the (stalled on GCC fixes) NetBSD port.
> > There are three core decisions that are made for any possible BSD port:
> > 1) Which BSD?
> > 2) Which libc?
> > 3) Which userland?
> > These have been debated long, hard, and mostly fruitlessly; the existing
> > ports all exist because one or more people quit debating and started
> > *doing*.
Right, see the debian-bsd archives for details. It's a worthwhile read
for anyone seriously interested in this, since those issues have all
been done to death several times.
> > Personally, I wanted NetBSD (for portability), native libc (for sanity),
> > and Debian (GNU-ish) userland (for compatibility). The FreeBSD folks
> > prefer(ed) GNU libc for question #2; which is fine, since I'm not working
> > on the FreeBSD port myself. :)
> I've looked at the debian/freebsd pages, and as you say above, it seems
> stagnant - the only link to something to try is to satie.debian.org :)
> otoh, you seem to suggest that the freebsd port is preferring gnu libc,
> which may indicate that the ports page is so out of date as to be
> misleading (they say they will be using freebsd's libc).
Grr. Truth is, I prefer a working solution. glibc looked like it might
solve a lot of problems. Doesn't look like that to me now. I tried it
out, but have concluded that it was a deadend, although it reduced
certain headaches. I haven't discarded any of the libc port.
The port page, and general lack of activity have to do with the simple
fact that I don't have time, bandwidth, or even space to setup my
computers right now. Lucky to have email. :(
> > I don't care about GNU, precisely, however; the parts of the Debian
> > userland that aren't GNU should not be arbitrarily replaced by GNU tools,
> > any more than we should arbitrarily use the BSD tsort rather than the
> > Debian-packaged GNU one (for one example, that breaks some build things :)
> I agree with this too - from my pov (albeit probably an uninformed one)
> replacing the native libc with gnu libc appears almost unnecessary, and
> creates a lot of work
A great deal more than was apparent before trying it out, too. It
creates major headaches.
> ok, given the above, is there some way to work out:
> - which subgroup has what goals/ideals
> - who's working on what
> - what there is to do
> The real "problem" is that I'd like to work on a debian/freebsd port
> (mostly because I am familiar and like freebsd, and have some machines
> running freebsd that I think could benefit from it), and I have similar
> goals to you (apart from the netbsd part), but most importantly don't
> appear to have the same goals as the debian/freebsd people; lastly, I
> don't know the status of the project closest to what I'd like to achieve
> (or even what that project is), what there is to do, and what there is
> that I could do.
> Do you have any suggestions/answers? Does anyone here have them?
My goals for the freebsd port are actually pretty similar to what you
described. I would like to pick it back up again with 5.0, but can't
right now. I hope things may change in a couple months, but at present,
I'm busy trying to make enough money to survive. That has to take
priority for now.