don't alert user on worker restarts caused by hitting max requests

[prymitive]

When worker restarts due to hitting request limit uWSGI generates warning that might be confusing to users since it looks like something bad has happened while everything is under control:

Nov  7 22:01:42 localhost app: ...The work of process 99 is done. Seeya!
Nov  7 22:01:43 localhost app: DAMN ! worker 1 (pid: 99) MISTERIOUSLY died :( trying respawn ...
Nov  7 22:01:43 localhost app: Respawned uWSGI worker 1 (new pid: 111)

With this patch uWSGI handles such condition and doesn't show any warning:

Nov  7 22:03:26 localhost app: ...The work of process 99 is done. Seeya!
Nov  7 22:03:27 localhost app: Respawned uWSGI worker 1 (new pid: 111)

[prymitive]

Maybe there is more obvious way to detect such restarts, but it seems that master doesn't mark worker in any specific way during restart so I'm using delta_request. I'm not sure if this is 100% correct way.

[unbit]

it is a bug in the condition (it should check for termination signal as the "misterious" death). HEAD should be fixed

[prymitive]

It works with latest master