Bernd Fondermann wrote:
> Noel J. Bergman wrote:
> > Perhaps we should just test that there IS a backlog, and not try to be
too precise.
> Probably. If this is OS or TCP specific, we should only test this if
> we would really like to unit test the OS or TCP. :-) We can take all
> things for granted out of scope of the unit we are testing.
Please review the current test, and let me know what you think. The test
presumes that there is a fixed number of connections that will be accepted,
a hard minimum of backlog connections, an upper bound of backlog connections
that can be measured, and well-defined behavior once that upper bound has
been reached.
--- Noel
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
