comparison mercurial/commandserver.py @ 40788:41f0529b5112 stable

commandserver: get around ETIMEDOUT raised by selectors2

selector.select() should exits with an empty event list on timed out, but
selectors2 raises OSError if timeout expires while recovering from EINTR.

Spotted while debugging new chg feature.

40787:d1bda397df73

                # for requests that are queued (connect()-ed, but haven't been
                # accept()-ed), handle them before exit. otherwise, clients
                # waiting for recv() will receive ECONNRESET.
                self._unlinksocket()
                exiting = True
            ready = selector.select(timeout=h.pollinterval)
            if not ready:
                # only exit if we completed all queued requests
                if exiting:
                    break
                continue

40788:41f0529b5112

                # for requests that are queued (connect()-ed, but haven't been
                # accept()-ed), handle them before exit. otherwise, clients
                # waiting for recv() will receive ECONNRESET.
                self._unlinksocket()
                exiting = True
            try:
                ready = selector.select(timeout=h.pollinterval)
            except OSError as inst:
                # selectors2 raises ETIMEDOUT if timeout exceeded while
                # handling signal interrupt. That's probably wrong, but
                # we can easily get around it.
                if inst.errno != errno.ETIMEDOUT:
                    raise
                ready = []
            if not ready:
                # only exit if we completed all queued requests
                if exiting:
                    break
                continue