httpd-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Jim Jagielski <>
Subject Re: question about the STATUS entry for lingering close
Date Tue, 11 Apr 2000 11:59:54 GMT
Jeff Trawick wrote:
> Where is the knowledge of when it is safe to use SO_LINGER?  I think 
> this question applies to various configuration issues.  In 1.3, there
> was a lot of hard-coded configuration based on the platform, so there
> was a place to say "NO_LINGCLOSE" or "USE_FCNTL_SERIALIZED_ACCEPT" or
> whatever.  We can *usually* probe for the ability to do this or that,
> but we can't always probe for when something is broken and we don't
> try to probe for the best possible mechanism.  We have the policy to
> prefer one type of interprocess lock over another when it exists, but
> there is no platform-specific policy for anything, right?

Well, I'd prefer that if we know something about a platform because
of 1.3, that we keep that knowledge in 2.0. Just because an OS
has some capability, as you say, doesn't mean it's not seriously
damaged. I brought this up with the NO_USE_SIGACTION as well,
where the platforms in question _have_ sigaction() but it's
not reliable enough that we should depend on it. So we're looking
at some sort of osconf.h file that replicates in some small way
some of the entries in 1.3's ap_config.h file. Or hardcode it

   Jim Jagielski   [|]   [|]
                "Are you suggesting coconuts migrate??"

View raw message