Re: automake 1.7.1 dependency tracking regression?

[Matthias Andree]

On Mon, 02 Dec 2002, Alexandre Duret-Lutz wrote:

> Just a few words about this:
> 
>  Matthias> you don't have access to the $(sysconfdir) like
>  Matthias> thingies in configure[.ac] as the ultimate and "real"
>  Matthias> expansion will only take place in the Makefile.
> 
> Several answers to this issue are presented in the Autoconf FAQ,
> at the end of the Autoconf (>=2.54) manual.  There is a pointer
> to that in the second revision of the text I sent.

I've seen these all, and I've been through these. Compilation flags are
certainly one aspect I've not used so far, but if you happen to need
more of then, the command line tends to get so long that "make -s" is
the only way to make some sense of the output. 

We have discussed the other approaches and know their disadvantages. The
header approach breaks dependency tracking on "make foo", the next is
against the standards, and adding more bloatware^Wmacros is not my
favourite either.

I find the "generate a .c file" approach quite attractive at the moment,
it works well for us, does not violate coding standards, does not
require additional macros and that's it.

Bonne nuit.