Rainer Weikusat <[email protected]> writes: [...]
> The downside of this approach is that this means adding > future connect timeouts will become proportionally more expensive as > more and more connections are being established This is an understatement. Based on a somewhat simplified model of the event list (initially empty, only connect timeouts), adding n timeouts in a row would have a cost cost 2 * (n - 1) + 1 (assuming 'insertion cost/ empy list' 1, same for removal cost) when adding and removing the events but would be (n * (n + 1)) / 2 for the other case (sum of 1 + 2 + ... + n).
