[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

1Pipelining streams0

    Date: Fri, 23 Oct 87 14:25 EDT
    From: Daniel L. Weinreb <DLW@ALDERAAN.SCRC.Symbolics.COM>

    Yes, the pipe implementation could be changed to grow up to a limit, but
    it doesn't seem very important, and it certainly doesn't seem like a
    distinction between "true" and (presumably) "false" pipes.  [...]

At least we agree on that -- the pipe implementation should start with a
minimum buffer size and then expand to a maximum buffer size.  These
sizes might be options to the pipeline routine.  As for the use of
"errors" when the pipe becomes filled, well, I was trying to think
of a mechanism for detecting a full pipe.  Since a programmer might only
use this error condition during development, as in an experiment to
determine the "optimal" buffer size, I suppose that it really isn't
necessary for day-to-day use.