httpd-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Jorge Schrauwen" <jorge.schrau...@gmail.com>
Subject Re: mod_define ported to httpd 2.0/2.2
Date Fri, 25 Aug 2006 14:35:29 GMT
On 8/25/06, Rainer Jung <rainer.jung@kippdata.de> wrote:
> Thanks for the feedback. I'll post URLs during the weekend. Of course
> you'll need to look at it before making decisions, I was simply
> interested in a first basic opinion.
>
> Concerning mod_define and mod_macro: they are a good fit together.
> mod_macro gives you the ability to factor out repeating parts of the
> config, but you have to explicitely call the macros with values.
>
> With mod_define you can set the values from the outside using
> environment variables. That way you can start similar instances with the
> same configs by setting params values during startup - without the need
> to copy and patch the config files.
>
> Sometime one should fuse the two together, but I had the impression,
> that the code of mod_macro first needs some reworking to make it fit
> better into the 2.0/2.2 configuration hooks.
>
the latest version runs quite well, afther some minor tweaks it seems
to compile on mac, win32 and win64 without a problem...

I'm share a config that makes use of mod_macro between my windows and
mac machine atm.

> Regards,
>
> Rainer
>
> Nick Kew schrieb:
> > On Friday 25 August 2006 14:05, Jim Jagielski wrote:
> >> I'm +0 about it, but I agree that I'd like to see
> >> the package 1st :)
> >>
> >> I've never been a fan of mod_define, seeing mod_macro (as
> >> Jorge seems to think as well) as much more useful...
> >
> > Agreed, mod_macro is *the* configuration module:-)
> >
> >> check
> >> out some of my OLD Apache PPTs from the 1st ApacheCons ;)
> >> My main issue with mod_define is that there always seemed to
> >> be such potential for issues and conflicts, esp when you
> >> have rewrite rules, etc...
> >
> > The argument in favour is that we do periodically get requests
> > for variable interpolation in various directives.  A general framework
> > would have some value.  But to be really useful, it'll have
> > to deal with both config-time and request-time interpolation
> > without confusing the hell out of the lusers.  And probably
> > other issues I haven't thought about.
> >
>


-- 
~Jorge

Mime
View raw message