The check doesn't even provide the intended protection against races completely, as I noted earlier. I think the discussion has shown that that particular attempt to detect the error wasn't a good one, so I've reverted it.
I'll instead be looking into moving the callback code completely to the Stdio.Fd class, so that the race is avoided better. Nothing except Stdio.File should be depending on the Stdio.Fd interface, hopefully.
/ Martin Stjernholm, Roxen IS
Previous text:
2003-03-19 16:02: Subject: I/O callbacks in blocking mode
No, I don't feel like implementing such a check, I feel like unimplementing masts check. I think the usefulness of read callbacks on blocking fds outweighs the "potential danger" (which is rather limited).
/ Marcus Comstedt (ACROSS) (Hail Ilpalazzo!)