Mail Archives: cygwin-developers/2002/07/29/17:27:38
Conrad Scott wrote:
>"Nicholas Wourms" <nwourms AT netscape DOT net> wrote:
>>Perhaps we could implement a counter to ascertain exactly how
>>cycles until dies? Either that or I could pipe netstat to a
>>have nano count how many open sockets there are... That figure
>>not be taken as exact, due to the fact I didn't run 'time' and
>>guessing when you asked.
>You could run the client through (something like):
> client <xxx> | grep 'connect' | wc
>But I was more interested in the error codes than in the exact
>number of sockets. That can be affected by lots of things, such as
Well I am configured for standard 100bT access, so it might be possible
that the MTU's are different.
>Also my guess is that on your system, the last connect was to
>socket 4999 or 5000? Especially with the EADDRINUSE error that
>you got at another point, which probably indicates that all the
>temporary sockets (1024-5000 ?) were in TIME_WAIT state. I can't
>ever hit that problem, 'cos my computer's not quick enough :-(
>[just guessing here].
The machine runs slow as dog s**t though after running the tests. Any
attempt to generate a sound results in it hanging for ~45sec. *Sigh* I
guess that's what you get from software driven soundcards...
>BTW, what was the machine set up for this? (windows version,
WindowsME, Athlon-4 1.1GHz, 320MB
- Raw text -