On Thu, Mar 26, 2009 at 04:30:04PM -0500, Shawn Walker wrote: > Danek Duvall wrote: >> On Thu, Mar 26, 2009 at 04:17:07PM -0500, Shawn Walker wrote: >> >>> ...there are probably a few more variations on the above theme, but >>> it should be clear that this is per-package state; not publisher >>> state. >> >> I'm glad you have the patience to work out these cases. They just bring >> out my fascist side and make me want to prevent any of these problems by >> disallowing people from configuring their publishers in any meaningful way. > > Don't forget the repository maintainers can cause us grief as well :) > > We do have a bug open to prevent users from removing publishers that > have installed packages. But that won't stop them from manually editing > cfg_cache, or changing the origin url to one that doesn't contain the > expected packages.
There are plenty more ways that repo maintainers can make life difficult for the client. The client still has no way of coping with the case where the maintainer creates a new repository at the same address, effectively blowing away the old repository. In this case, it actually might be desireable to remove the publisher, since a refresh from a publisher that's in a completely different state is likely to confuse the client even more. -j _______________________________________________ pkg-discuss mailing list [email protected] http://mail.opensolaris.org/mailman/listinfo/pkg-discuss
