>>>>> "Bill" == Bill Sommerfeld <sommerfeld at sun.com> writes:
Bill> can't you block delivery of signals on the server (wrapping the hg Bill> server side?) Yeah, I was wondering that, too. Do we know what exactly causes Mercurial to go away on the server? Would it help to put Mercurial in its own process group (instead of, or in addition to, blocking signals)? Bill> can you trust all client to be running the fixed cdm at the time Bill> this goes live? (IMHO, no.) Yeah, any distributed system that cannot recover from client-side errors is flawed. I don't think it's critical for the Tuesday transition, but I'd like to see a more robust fix soon (assuming that workspaces on opensolaris.org are vulnerable to the same problem.) mike
