Re: Order of static pattern rules is significant while IMHO it should not be

On Thu, 2025-06-05 at 21:43 +0200, Denis Excoffier wrote:
> jupiter% touch foo100.z0 bar200.z0; rm -f foo100.z1 bar200.z1
> jupiter% make
> foo100 (foo100 is expected and foo100 is obtained)
> 200 (bar200 is expected but 200 is obtained instead)
> jupiter% make --version | head -1
> GNU Make 4.4.1
> jupiter%
> 
> The output of 'make -p' seems to infer that
> '$*' is filled with the last pattern rule encountered,
> even if this is not the one with the recipe.
> 
> Is it intended behavior?

I think you are correct that this is a bug.  Certainly, it doesn't
accord with my understanding of how GNU Make should work nor the
documentation as I read it.

I think static pattern rules without recipes are rare (at least I've
never seen or used one) and so the details of handling them might have
been overlooked.