Thanks for the suggestion. Thanks too for using the word "sleep", because I realized I could just use that. Created a trigger that sleeps for 60 secs, then tried a "/kill" during the sleep. Death doesn't come until the sleep is done. Then I realized that only proves it's one thread per client, not necessarily one process for all clients. So I ran two clients (I didn't know if you *could* run two clients on the same machine, but you can), started the sleep with the first guy, then did a "/kill" with the other guy. The kill doesn't come, and both clients wind up getting booted with "[user] could not respond" appearing on the soldatserver console.
This *seems* to confirm that it's only one process sans threads. Unless it's one thread per ip (which doesn't make much sense), or threads that wait in turn (possible, but acceptable for my work).
Date Posted: June 15, 2007, 04:51:17 PM
P.S.: Someone should really put together a list of functions and procedures we have access to and that are not being ripped out of the server in the near future. It's stuff like Sleep(), ThreadFunc(), SetArrayLength(), etc. that aren't documented to be available but just happen to be there if you somehow figure out they exist.