[RFC] Print MI prompt on interrupted command

[Tom Tromey]

>>>>> "Tom" == Tom Tromey via Gdb-patches <gdb-patches@sourceware.org> writes:

Tom> Does it seems like a reasonable idea?  It is likely to break existing
Tom> clients?  Or perhaps should start_event_loop (that's what notifies the
Tom> command_error observer) loop over all UIs and emit a prompt for each?

>> I'm leaning toward this latter plan now.

I did some experimenting with new-ui and I think that it would be weird
to emit a prompt in this situation.  So now I'm back to thinking the
original patch is probably ok.

Tom