Using a single handler for both timeouts and SIGPIPEs is a really bad idea,

but I didn't change that part.  Calling shutdown is unreliable, so we avoid
it when possible.  Using soft_timeout requires that the sender check for an
aborted connection rather than continuing after an EINTR, which was almost
never the case.  Timeouts that used to be initiated before send_http_header
(and never killed) are now initiated only within or around the routines that
actually do the sending, and not allowed to propagate above the caller.

Reviewed by: Chuck Murcko, Jim Jagielski


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@77800 13f79535-47bb-0310-9956-ffa450edef68