Event and expected atomic behavior

[Jim Jagielski <ji...@jaguNET.com>]

What I've done is changed trunk to simply warn, but not
block startup... ideally, we should either "fix" atomics
(which is more an APR issue, of course) or fix Event.

for those on APR: The httpd Event mpm uses atomics as signed
ints, not unsigned, and for a long time trunk had a startup
check which made sure that atomics worked as "expected"...
This was just backported to httpd 2.4.x and could raise
some issues or "bug" reports...