On Tue, Aug 12, 2003 at 04:43:49PM -0700, Brad Fitzpatrick wrote: > I was thinking: how about I just ditch c->write_and_uncork and check > instead of c->write_and_go == conn_read? If the state machine is moving > from writing to reading, it's done writing. > > I'll do that. Reply if you can think of a case where that's the wrong > thing, let me know. No, I can't; that seems safe enough. -- avva