Re: Debian GNU [was: smarter way to differ architectures needed?]
Gordon Matzigkeit <email@example.com> wrote:
> [CCed to debian-devel, since people there are probably interested.]
I'm not sure that's wise, but I've left the cc in place for this
> >>>>> Brian May writes:
> BM> I am not sure why you would want to say that something depends on
> BM> an architecture (ie major change involved in the way we think
> BM> about packages) when you can do everything required the existing
> BM> mechanism. Depends is designed to depend on other packages, it
> BM> was never meant to indicate kernel,arch combination is required
> BM> to install a package.
> I know that, and I think that design decision is the cause of all our
> problems, including the schism between GNU and Linux.
> Kernel and hardware architecture are really just some of a package's
> dependencies. Think about it: some packages depend on Linux and others
> don't. Some depend on i386 and others don't. Some depend on libc, and
> others don't. Some depend on Perl, and others don't. Some depend on
> nothing at all.
Depends is really just a best-effort, first approximation solution to
a problem which I believe is NP-complete. [But that's off the cuff,
I've not sat down to verify that my axioms are reasonable.]
Most of the time we can find something about a system which we can
simplify to make things continue to work together. Where we can't,
our lives get that much more complicated.
For the case of GNU/Hurd, if it's a design goal for the Hurd to be a drop
in replacement for Linux you should probably accept the consequences of
that design goal rather than trying to get the packaging system to do
it for you.
If all you're trying to do is storage management, we can probably do
something reasonably decent with symlinks and an extra control file tag.