httpd-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From (Ralf S. Engelschall)
Subject Re: APACI 980313: Give it a try
Date Sat, 21 Mar 1998 18:05:36 GMT

In article <Pine.LNX.3.96.980321174857.18192B-100000@ecstasy.localnet> you wrote:
> On Fri, 13 Mar 1998, Ralf S. Engelschall wrote:

>> Now that we have complete shared object support for all modules I've enhanced
>> my APACI stuff to support shared object building, too. A lot of other
>> enhancements were done, too (--add-module=mod_xxx.c to add a module on-the-fly
>> to the build process, --enable-module=most for only enabling those modules
>> which are really usable anywhere, etc.).
>> URL:

> Ralf, this is very nice stuff. I love being able to build different
> configurations so easily from simple command lines. Since it doesn't upset
> the current way of configuring and compiling, it might be worth adding it
> to 1.3. 

Thanks, that's exactly what I said in the past: we can obly gain from
including it, because no existing stuff is changes. But some others perhaps
still think it would break something or at least could create too much new
problems. But perhaps these statements are from the first initial version. The
current APACI stuff is now very complete and clear and as portable as the
src/Configure script. So perhaps we really should give it a second try for
1.3b6 or at least 1.3b7. I'll think about it when there is a real chance that
there aren't again such massive -1 because we are late in the release cycle.

> But there is one thing I don't like. It always sets SERVER_CONFIG_FILE to
> an absolute path. I like to build Apache with a particular prefix, then
> run it with multiple server roots with -d. Setting SERVER_CONFIG_FILE to
> an absolute path means that -d does not work. I think sysconfdir should
> default to the current value, "conf", for maximum compatibility and so -d
> works again. Perhaps there could be a test in configure to see if the user
> is trying to set --sysconfdir, and if _not_, do not define

Hmmmm... good point, Paul. I personally always use -d in conjunction with -f,
so totally ignored the point that just using -d alone can be very useful.
Yes, APACI should support this, you are right. I'll try to add some more
intelligence to the option parsing, so we can support this. I take your
suggestion serious and will add support for this way of configuring.

Thanks for your feedback, Paul.
                                       Ralf S. Engelschall

View raw message