On Wed, Aug 23, 2000 at 10:02:39PM +0100, Theodore Hong wrote: > Well, TCP has a flow-control mechanism built into it -- it depends on how > your TCP stack is implemented, of course, but it strikes me that if your > application stops reading data out of a socket for a while and the OS > buffers fill up, the OS should stop sending ACKs back to the sender, in > which case the sender takes this as a signal to slow down. If too much > time passes before starting to receive ACKs again, however, the sender will > close the connection.
It doesn't strike me as a very sexy way of achieving this. > Anyway, nothing an application does should be able to mess up the OS, > right? (in the ideal world) Theo, Windows is one of our target platforms... > > theo > > _______________________________________________ > Freenet-dev mailing list > Freenet-dev at lists.sourceforge.net > http://lists.sourceforge.net/mailman/listinfo/freenet-dev > -- \oskar _______________________________________________ Freenet-dev mailing list Freenet-dev at lists.sourceforge.net http://lists.sourceforge.net/mailman/listinfo/freenet-dev
