Re: VPATH rebuilding of gawk

[Ralf Wildenhues]

* Andrew J. Schorr wrote on Mon, Jan 22, 2007 at 09:57:20PM CET:
> 
> Yes, well I guess I see no reason to include the 'eg' tree in the
> distributed tarball (or in CVS for that matter).  Given that 'eg'
> is generated from the .texi docs, why bother to distribute it and/or
> maintain it?

To facilitate cross-compilation?

> On Mon, Jan 22, 2007 at 08:56:41PM +0100, Ralf Wildenhues wrote:
> > The `$<' macro is portable in inference rules (like .c.o) only.
> > Solaris `make' will barf otherwise.  Several instances in the patch.
> 
> My mistake, thanks for catching that, you seem to be correct, even though
> the Sun documentation seems to suggest that $< ought to work.  The man
> page says:
> 
>         $<    The name of a dependency file, derived as if  selected
>             for use with an implicit rule.

Oh.  The Sun docs call inference rules "implicit".  Hmm, the GNU make
manual uses that notation, too.

> But in fact, it does not seem to work in this context, as you suspected.

Oh, no suspicion involved here:
<http://www.gnu.org/software/autoconf/manual/html_node/_0024_003c-in-Ordinary-Make-Rules.html>

I agree with the rest you wrote, I merely think that changing the
current state isn't a clear win-only situation, see above.  But
certainly I don't care much.

Cheers,
Ralf