On Wed, 2009-12-02 at 15:19 -0800, [email protected] wrote: > On Wed, Dec 02, 2009 at 06:12:42PM -0500, Joanmarie Diggs wrote: > > Once the endless looping is stopped (as is the case with this patch, > > thanks!), what is the expected outcome? Should the absence of one or > > more manifests from one (admittedly borked) publisher prevent checking > > for updates for the other enabled publisher(s)? > > The transport is only doing what it was asked. If the client asks the > transport for a manifest that doesn't exist at a particular publisher, > the transport must return that it encountered an error while trying to > satisfy the request. > > Prefetch is a speculative operation, so nobody cares about the success > or failure; however, expect the transport to raise an exception if it > can't retrieve a manifest in the synchronous path -- get_manifest.
Gotcha. So let me rephrase the question and direct it towards those who implement the clients: Once transport does the right thing, stops the loop, and raises the expected exception, what is the expected behavior of the pkg cli and gui? Should they pass along the error to the user and stop, or should they attempt to proceed with the requested task for the other enabled publisher(s)? I don't have an opinion. But if the answer is that they should attempt to proceed with the requested task, I've another bug to file. ;-) Thanks again. --joanie _______________________________________________ pkg-discuss mailing list [email protected] http://mail.opensolaris.org/mailman/listinfo/pkg-discuss
