On Wed, 2005-10-12 at 18:12 +0200, Ludovic Courtès wrote:
> "Jonathan S. Shapiro" <[EMAIL PROTECTED]> writes:
> 
> > Autoconf is an abomination for many reasons beyond that one.
> 
> Warning: you're on a GNU mailing list.  ;-)
> 
> Besides, I'd be glad to hear about your precise griefs against Autoconf...
> 
> Autoconf is a good tool to help port things across Unix-like platform...

Conceptually, autoconf was a good idea, and I agree that it was needed
at the time. 

Setting aside the use of M4, which is another abomination, the
fundamental design flaw in autoconf is that it tests compatibility by
executing programs on the target platform rather than consulting a
database. The consequence is that Autoconf is effectively useless for
cross-builds. Vestigial support for cross compilation exists,
sporadically, in some of the macros, but it is not really used or
understood by autoconf users, and its design is flawed. Given the model
for testing, how exactly do you test the integer size on a cross target?

In consequence, the *users* of autoconf are just amazingly sloppy about
not doing clean configuration. The failures of autoconf are becoming the
failures of applications more generally. Have you tried building a cross
compiler lately with GCC?

Finally, the design of autoconf does not promote portability. It
promotes platform-specific compilation. This was progress in the
negative direction.

shap



_______________________________________________
L4-hurd mailing list
[email protected]
http://lists.gnu.org/mailman/listinfo/l4-hurd

Reply via email to