Subject: Re: latest netatalk (was Re: Netatalk AFP performance?)
To: Hauke Fath <hauke@Espresso.Rhein-Neckar.DE>
From: Joe Laffey <joe@laffeycomputer.com>
List: port-mac68k
Date: 10/26/2001 08:55:56
On Fri, 26 Oct 2001, Hauke Fath wrote:

> At 14:21 Uhr -0500 25.10.2001, Joe Laffey wrote:
> >Autoconf is a great system that works fine on the BSD OSes.
>
> As long as it works. When it comes broken, it's a nightmare to see through
> and fix. Even imake work is nicer.


Well, you are right, that unless you are VERY familiar with
Autoconf/Automake (which I am gradually becoming, but now where near VERY)
it is a royal pain to see what is wrong. In this case I find it easiest to
run configure and start the compile and THEN edit the Makefile. Of course,
this is for a one-time hack to get it compiled on my machine.

When Autoconf work it does work great, though ;-) (hmm... when windows
works, it works pretty well, too. So I guess that's not saying much!)

>
> >If they have
> >broken it on OSes other Linux I doubt it has to do with Autoconf/Automake,
> >so much as their implementation.
>
> It has to do with adding an additional level of complexity, and with
> needlessly replacing Makefiles that worked - change for change's sake.

I'll buy that. Added complexity leads to a more fragile system. I have not
worked directly with imake. But I have built a number of projects with
Autoconf. For simple checks (checking for header, sizeof stuff, endian,
etc.) it works real well. But I agree it would be harder for someone not
familiar with the project to make changes!


--
Joe Laffey              |  Want to convert subnet masks between different
LAFFEY Computer Imaging |  notations, or figure the number of IPs in a block?
St. Louis, MO           |  Whatmask-It's FREE - www.laffeycomputer.com/wm.html
------------------------------------------------------------------------------