httpd-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Dean Gaudet <>
Subject Re: apache/linux modules
Date Tue, 03 Feb 1998 01:54:19 GMT
The same is true of our GuessOS and Configure.  If we don't know the
platform we stop dead.  So this isn't an argument against autoconf in any


On Mon, 2 Feb 1998, Alexei Kosut wrote:

> On Mon, 2 Feb 1998, Dean Gaudet wrote:
> > If 2.0 is a rewrite (I'm beginning to agree with Roy and Rasmus and
> > whomever else on this one) then starting with autoconf means that for
> > those platforms that we have *active* support we'll still be sure to get
> > it right.  For those platforms that we don't have active support, oh well. 
> > We try our best (autoconf is a really good way to try to support posix
> > platforms you don't have access to).
> But what about platforms the autoconf maintainers don't have access to?
> That's a problem I sometimes have with autoconf.
> I mean, we can use the config.guess stuff to determine an OS name (which
> are actually pretty close to our GuessOS), but last time I looked at
> autoconf, if you do that and come across an OS it doesn't know, autoconf
> basically stops dead in its tracks. And there's nothing that can be done 
> about that without rewriting the way autoconf invokes config.guess
> (e.g., you can't even do "./configure --host=some-unknown-host"
> because it checks that against the known types.) And the config.guess
> maintainers don't seem to be as up-to-date as they possibly should.
> If this has changed since last I looked (about a year ago), then that's
> something...
> -- Alexei Kosut <> <>
>    Stanford University, Class of 2001 * Apache <> *

View raw message