I've looked at 359 and 360 so far. > 360 recommit should check for workspace modification in the same > manner as Mq would
Looks fine. > 359 when recommit fails, cdm needs to make more effort to return the > workspace to sanity. [...] > I'd like input on whether bringing the workspace back to the local > tip or bringing it back to where it was previously is more correct. > > (there's further changes in that area I may end up making, however, > currently, the local tip is the least confusing place to be) I think we should try to return the user to where she was. If we can't do that, we should provide feedback as to where we left her. Is that possible? If not, can we require that the user always be at the local tip before running recommit? Would that be a good requirement in general? Come to think of it, can the exception handling code assume that the local tip still exists? (Is there a particular order to what active.bases() returns?) Also, what happens to the new tip if we catch an exception from q.strip()? mike
