On Wed, Dec 08, 1999 at 09:00:09AM -0500, Ben Collins wrote:
> On Wed, Dec 08, 1999 at 09:40:21PM +1000, Anthony Towns wrote:
> > + Since dpkg will not prevent upgrading of other packages
> > + while an <tt>essential</tt> package is in an unconfigured
> > + state, all <tt>essential</tt> must supply all their core
> > + functionality even when unconfigured. If the package cannot
> > + satisfy this requirement it should not be tagged as essential,
> > + and any packages depending on this package should instead
> > + have explicit dependency fields as appropriate.
> Sorry that I missed most of this, but...
> I think this will make the dependency chain even more complex. I agree
It doesn't actually do anything, it just documents existing caveats.
> with the idea that essential packages should be functional even without
> being configured, but then you will need to make all essential packages
> static binaries to satisfy this.
ldso and libc6 are already Essential, so the dynamic linker, and libc6 are
guaranteed to be available.
> Remember that dpkg allows them to be
> unpacked even when pre-deps aren't satisfied, just not configured. So if a
> library dependency isn't met, and it is unpacked, then there is no way to
> have it be functional.
Erm. A Pre-Dependency must be satisfied when a package is being unpacked,
although it doesn't have to be entirely configured at this point.
It does mean Essential packages need to have all their dependencies listed
as Pre-Depends, but that's okay.
I'm not really clear on what you're saying here, but I'm fairly sure there
isn't an actual problem.
> Maybe dpkg _should_ be changed to check the status of essential packages
> before installing things. It seems like that is the core problem and is
> much easier to solve. Plus it makes sense, since essential packages are
> supposedly required for a minimaly functional system (a lot of which dpkg
> depends on anyway).
It's not necessary though, and that'd give you much stricter ordering
requirements than dpkg currently needs. (Although Apt currently enforces
this --- see Santiago's and Jason's comments in -policy previously)
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:
pgp299J7BpVPr.pgp
Description: PGP signature