* Danek Duvall <[EMAIL PROTECTED]> [2008-09-04 21:18]: > On Thu, Sep 04, 2008 at 02:10:46PM -0700, Dan Price wrote: > > > I thought we had imagined a "broken" state? > > Stephen has a line in pkg-states.txt about damaged packages being in a > substate of installed, but I always assumed that was a state that a package > entered into when pkg verify discovered that it wasn't correctly installed. > > It could be used for this case as well, though, but once pkg has crashed, > I'd be a bit leery of doing all the work necessary to figure out what was > broken. Perhaps it fires off a pkg verify after having crashed during an > install / uninstall / image-update?
We need transitional states so that removals don't destroy dependencies believed satisfied by executing installs (transit_install); similarly we need to know what's being removed so that installs can reinstall their dependencies (transit_remove). An interrupted install could leave its affected packages in transit_install, but we'll also need to some locking around evaluations, etc., too. - Stephen -- [EMAIL PROTECTED] http://blogs.sun.com/sch/ _______________________________________________ pkg-discuss mailing list [email protected] http://mail.opensolaris.org/mailman/listinfo/pkg-discuss
