Re: Possible bug with 3.81?

[Paul D. Smith]

On Saturday, 28 October, Eli Zaretskii (address@hidden) wrote:

> > No, I think this is a bug.  Make should choose the first pattern rule to
> > build lib%.a, because both patterns can be built.
> 
> I'm confused: I think Make did choose the first pattern rule.  The
> problem was with the order of prerequisites in that rule.  Am I
> missing something?

If you unpack the example attached to the first post and annotate it a
little bit, you'll see that what actually happens is make chooses the
second pattern rule rather than the first one:

  .c.o: ; @echo '.c -> .o'

  lib%.a : %.def %.o ; @echo Rule: .def .o

  lib%.a: %.def ; @echo Rule: .def

  $ make -rR   
  Rule: .def
  .c -> .o

  $ make-3.80 -rR   
  .c -> .o
  Rule: .def .o

I think 3.80 (and below) have this right, and 3.81 has this wrong.

-- 
-------------------------------------------------------------------------------
 Paul D. Smith <address@hidden>          Find some GNU make tips at:
 http://www.gnu.org                      http://make.paulandlesley.org
 "Please remain calm...I may be mad, but I am a professional." --Mad Scientist