httpd-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Ben Laurie <...@gonzo.ben.algroup.co.uk>
Subject Re: snprintf and release schedule
Date Sun, 19 Jan 1997 17:18:45 GMT
Jim Jagielski wrote:
> 
> Ben Laurie wrote:
> > 
> > Jim Jagielski wrote:
> > > 
> > > Jim Jagielski wrote:
> > > > 
> > > > Marc Slemko wrote:
> > > > > 
> > > > > If people agree with my last comments on the lingering close stuff,
I
> > > > > think those changes (ie. rip out all lingering close stuff, add a
> > > > > shutdown() before close() in child_main) should also go in the next
beta. 
> > > > > 
> > > > 
> > > > I think that we should change things so lingering_close is NOT used
> > > > by default, but still leave the code in. There may be some OSs
> > > > (SysV4-based) that would benefit).
> > > > 
> > > > I'm curious why shutdown would be needed though, before close...
> > > > 
> > > 
> > > Working backwards on reading Email, I see the quote from the sockets
> > > FAQ. With that in mind, I'd say that lingering_close() is unneeded
> > > and should be scrapped. Also, that the close be wrapped by a shutdown
> > > iff SysV4-based.
> > 
> > According to that FAQ you need _either_ shutdown or SO_LINGER. I suppose it
> > is vaguely possible that you may need both. I'd be inclined not to scrap
> > lingering_close() instantly, but to provide options to do any combinations
> > of the following:
> > 
> > 1. USE_SO_LINGER
> > 2. USE_SHUTDOWN
> > 3. USE_LINGERING_CLOSE
> > 
> > And recommend that people test in the following order:
> > 
> > 1. No options (which we make the default)
> > 2. USE_SO_LINGER
> > 3. USE_SHUTDOWN
> > 4. USE_LINGERING_CLOSE
> > 5. USE_SO_LINGER and USE_SHUTDOWN
> > 6. USE_SO_LINGER and USE_LINGERING_CLOSE
> > 7. USE_SHUTDOWN and USE_LINGERING_CLOSE
> > 8. USE_SO_LINGER and USE_SHUTDOWN and USE_LINGERING_CLOSE
> > 
> > Of course, USE_SHUTDOWN and USE_LINGERING_CLOSE are nearly mutually exclusive
> > - I'd propose that USE_SHUTDOWN controls the final shutdown() in
> > lingering_close() when it is enabled.
> > 
> > Then, if any of these solve the problem, we need to know which, and the
> > platform.
> > 
> 
> Wow... that's a lot of options. As I recall, the "old way" worked
> fine for all systems other than some/all Sysv4 ones. If it
> ain't broke, don't fix it. Fix the specific problem with Sysv4,
> which sounds like a needed shutdown/close and leave the rest
> alone.

Exactly. That's why the default (1) is "no options". The point about listing
the rest is to nail down, in one go, precisely what, of the options thought of
so far, is needed for those platforms which don't work.

This is not what is happening in 1.2b4 or 1.2b5-dev, though.

Cheers,

Ben.

-- 
Ben Laurie                Phone: +44 (181) 994 6435  Email: ben@algroup.co.uk
Freelance Consultant and  Fax:   +44 (181) 994 6472
Technical Director        URL: http://www.algroup.co.uk/Apache-SSL
A.L. Digital Ltd,         Apache Group member (http://www.apache.org)
London, England.          Apache-SSL author

Mime
View raw message