httpd-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Greg Stein <gst...@lyra.org>
Subject Re: RANT: Absolute Paths and configure
Date Mon, 03 Apr 2000 09:01:07 GMT
On Mon, 3 Apr 2000, Manoj Kasichainula wrote:
>...
> +1 on replacing the current setup with something far better. I agree,
> it sucks.  It's a pain to follow even for me, since there have been so
> many changes required in the past few months to get everything to
> work, eliminate automake, and so on.

Lots of change, sure. But that doesn't mean that the basic approach (use
autoconf) is wrong. It simply means that we have thrown a very complex
piece of software, with the intention of being very widely portable, at
autoconf. Do we get it correct the first time? Feh. Anybody who believed
that should talk to me about a bridge that I own.

The stuff can certainly be refined and improved over time. Heck, we just
got rid of the 'apaci' file and streamlined that mechanism. We can
continue that over time.

Python touches more points on the underlying OS than Apache, yet it has a
very clean configure.in. I see no *inherent* reason for ours to be super
complex.

Just give it some time... don't look at the initial versions, decide it is
hopeless, and go and build Yet Another Mechanism. We already had a custom
mechanism (Configuration and APACI) -- part of the autoconf selection
process was to get a standardized tool that (more) people are familiar
with. We can then leverage that experience and improve our config process.

> Unfortunately, +1ing something better is the easy part; writing it is
> harder. It'll either have to be based on autoconf or written from
> scratch. Last I checked, requiring Perl, or any non-standard packages,
> on users' machines was out of the question.

I totally agree. I find it hard to believe that we need so much complexity
that autoconf can't do it for us (which, in turn, means we only need an
"sh" program).

> And until the Software
> Carpentry guys come through (assuming that their implementation
> doesn't require Python on users' machines), there is nothing better
> out there that I know of.

It will require Python, by definition.

www.software-carpentry.com for those who haven't heard of this.

Cheers,
-g

-- 
Greg Stein, http://www.lyra.org/


Mime
View raw message