GNU autoconf versus metaconfig

[James R. <qra...@an...>]

On Wed, 17 Oct 2001, Lars J. Aas wrote:

> Yes, I remember a third one.  The Configure script.  Wouldn't it be
> better to convert to Autoconf/Automake and rather provide
> configure-script command-line options for all the different choices?

On Wed, 17 Oct 2001, Matt Ackeret wrote:

> I personally don't know much about either solution, but if you mean
> do AWAY with the "interview format" set of questions, then I'd vote
> against it.
>
> Now, I bet there are a lot of the various questions that *COULD* be
> done away with (it seems to take FOREVER to answer the questions),
> and I still had to do extra hacking to get it to work on OSX if I
> remember correctly..
>
> But in general, the interview format seemed to work well.

The GNU autoconf (configure) versus metaconfig (Configure) issue was
discussed back in June 2000.  Here's a summary:

Reasons to switch to autoconf:

    1.  Metaconfig's interactive nature makes it difficult to build trn
        using automated packaging systems (like Red Hat and Debian use).

    2.  Autoconf is in much wider use than metaconfig.  More people
        understand it, and how to use it.

    3.  Conjecture: many (most?) people who run metaconfig probably
        guess their way through the questions, or just accept the
        defaults; you have to posses a fair amount of unix development
        experience to understand all of metaconfig's questions.

Reasons to not switch:

    1.  Switching to autoconf would require a lot of effort that would
        be better spent improving other areas of trn, like the
        documentation.

    2.  Metaconfig may not be Perfect, but it's Good Enough For Now.

    3.  Some people like the interactive nature of metaconfig.

I would guess at this point that if someone were motivated enough to
convert trn to autoconf, then it would be accepted.  (But I'm not one
of the maintainers, so take what I say with a grain of salt.)

-- 
James Ralston, Information Technology
Software Engineering Institute
Carnegie Mellon University, Pittsburgh, PA, USA