On Mon, Apr 01, 2002 at 09:07:28PM -0500, Jeff Trawick wrote:
> it isn't called in the parent... that's why I wanted to use
> -DNO_DETACH :)

by "parent" I meant "apache parent", not just "that process that forked
other processes".

> note that -DNO_DETACH from gdb is doomed...  setsid() will fail (at
> least with my normal environment)...  -DNO_DETACH doesn't work for you
> from gdb, you just hit your breakpoint before the failure point
> 
> I'll forget about it for now (but I will commit a cleanup of
> apr_proc_detach() which won't change the semantics).  For the long
> term I guess I want -DFOREGROUND and appropriate support in
> apr_proc_detach().

I thought about adding a -DFOREGROUND but I guess I convinced myself
that it was still just a replacement for -DNO_DETACH. Given this clear
use case I'm totally for -DNO_DETACH simply not calling apr_proc_detach()
(as before) and -DFOREGROUND calling apr_proc_detach() with the foreground
flag. Better to get it right now rather than later.

-aaron

Reply via email to