Re: Multiple pattern-specific variable assignments

[Paul D. Smith]

%% "Adrian Ashley" <address@hidden> writes:

  aa> Tried that:

Oh right.  Duh.  Need to remember to drink my morning tea before
responding :).

  aa> I hesitate to call this a bug, as the documentation does say that
  aa> pattern rule matching proceeds until the first match and then
  aa> stops.  But perhaps there's a case for arguing that if the effect
  aa> is only to set target- or pattern-specific variables, then all
  aa> patterns which match should do so cumulatively (perhaps excepting
  aa> terminal ones?)

Not sure what you mean by "terminal ones" in this context, but I agree
that stopping at the first pattern is perhaps not useful behavior for
target-specific variable patterns.  I'll think about this.

  aa> Or is there a Cunning Trick?

Well, you can always fall back to old-fashioned recursive variables :)

    CC_GCC=gcc
    CC_OTHER=another-cc
    CFLAGS = -DBASIC

    foo-%.o : CFLAGS += -DVARIANT=FOO
    bar-%.o : CFLAGS += -DVARIANT=BAR

    gcc-CFLAGS = -DGCC
    other-CFLAGS = -DOTHERCC

    CCTYPE = $($(patsubst $*-%.o,%,$@)-CFLAGS)

    %-gcc.o: %.c
        $(CC_GCC) $(CCTYPE) $(CFLAGS) -o $@ -c $<
    %-other.o: %.c
        $(CC_OTHER) $(CCTYPE) $(CFLAGS) -o $@ -c $<

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