On Wed, Feb 09, 2000 at 05:38:03AM -0600, Manoj Srivastava wrote:
> >>"Anthony" == Anthony Towns <aj@azure.humbug.org.au> writes:
> >> Horse puckey.
> >> This is a technical issue, and has nothing to do with the Social Contract.
> Anthony> It is? It doesn't?
> Anthony> ``I don't like the way people write programming languages in
> Anthony> themselves. Now, sure, for gcc, we'll have to make an
> Anthony> exception, but the rest of them can just get lost.'' doesn't
> Anthony> sound particularly technical.
> I am tempted to say, Horse puckey. I never say that gcc was
> the only program that should qualify.But if you want to be
> delibrately confrontational, bring it roght on. I can take statement
> to the extreme with the best of them.
And that you haven't been for the past year or so has made reading -devel
*much* less entertaining.
> Anthony> So they can't be automatically built for new architectures. Big
> Anthony> deal. That's no reason to get rid of them for everyone.
> You must be imagining things. Who talked about throeing the
> code out by default? I talked about having the package maintainers
> ask for dispensation, to ensure that the package are not putting in
> self dependencies for convenience.
>
> It wouyld have been far easier for me to ask that latex2html
> be installed in order to build the package, since I could just
> use the installed latex2html to build the html docs.
This isn't what it sounded like you were saying. It sounded like you
didn't want any compilers other than gcc in the system, rather than that
you just didn't want J. Random Package to Build-Depend on itself.
Having packages that can be built (maybe in one or two passes) without
having to have themselves already installed should definitely be packaged
that way.
Packages that inherently can't be built that way, should obviously just
be packaged, with a circular build-dependency, and ideally a README.Porters
or something to help new dists get the package ported.
I don't see where special dispensation comes into it though.
> Actually, *YOU* are the one trying to brush the issue under
> the carpet. *I* am the one asking that these packages be examined,
> and registered if they can't easily get rid of the self dependency.
Actually, I'd have thought that with 5 mainline architectures already,
most of these (outside the base system, perhaps) would have already been
dealt with? Are there any example packages with debian/rules files that
could use rewriting for this reason?
Cheers,
aj
--
Anthony Towns <aj@humbug.org.au> <http://azure.humbug.org.au/~aj/>
I don't speak for anyone save myself. GPG encrypted mail preferred.
``The thing is: trying to be too generic is EVIL. It's stupid, it
results in slower code, and it results in more bugs.''
-- Linus Torvalds
Attachment:
pgpiY_cgmMNsw.pgp
Description: PGP signature