Re: Help needed for writing a rule

[Patrick Begou]

Hi all,

Thanks for your suggestions.

@Paul: solution tested. Yes it works fine and is very near the initial 
syntax used in the original Makefile.
--------------------------------------------------------------------------------
.SECONDEXPANSION:

# building only the mod file
LIB/%_M.mod: $$(shell echo $$* | tr A-Z a-z)_m.f90
    @echo "building $*_M.mod from $<"
    touch $@

%.o:     %.f90
    @echo "building $@ from $<"
    touch $@

all: toto.o

clean:
    rm -f *.mod *.o LIB/*

toto.o :  toto.f90 LIB/TITI_M.mod titi_m.o
titi_m.o LIB/TITI_M.mod:        titi_m.f90
--------------------------------------------------------------------------------

@kaz: Works fine for building the application but each make rebuilds 
everything as there are no dependency
 on the target. Each time I run make, FOO.out is built again.
---------------------------------------------------------------------------------
%.out: UC = $$(echo $* | tr [a-z] [A-Z])

%.out: %.in#
        @echo "making $*.out (upcased as $(UC).out)"
        touch $(UC).out

foo.out: foo.in
---------------------------------------------------------------------------------

@David I think I'll have the same pitfall than with Kaz strategy. No ?

Patrick



Le 25/11/2022 à 17:49, David Boyce a écrit :
> Possibly this would be helpful?
>
> # Usage: $(call toupper,<string>)
> #   $1: string to be uppercased
> toupper = $(strip \
>   $(subst a,A,$(subst b,B,$(subst c,C,$(subst d,D,$(subst e,E, \
>   $(subst f,F,$(subst g,G,$(subst h,H,$(subst i,I,$(subst j,J,$(subst 
> k,K, \
>   $(subst l,L,$(subst m,M,$(subst n,N,$(subst o,O,$(subst p,P,$(subst 
> q,Q, \
>   $(subst r,R,$(subst s,S,$(subst t,T,$(subst u,U,$(subst v,V,$(subst 
> w,W, \
>   $(subst x,X,$(subst y,Y,$(subst z,Z,$1)))))))))))))))))))))))))))
>
> On Fri, Nov 25, 2022 at 11:33 AM Kaz Kylheku <kaz@kylheku.com> wrote:
>
>     On 2022-11-25 06:26, Patrick Begou wrote:
>     > Hi all,
>     >
>     > I'm stuck for several hours in trying to write a rule for
>     managing fortran modules. My difficulty is that the Cray compiler
>     requires a  module name written in upper case.
>     >
>     > example: if file is "toto_m.f90", it contains a module called
>     "toto_m" and the module file will be "TOTO_M.mod".
>     >
>     > How can I write a generic rule for building TOTO_M.mod from
>     toto_m.f90 source and put it in the LIB folder ?
>     >
>     > Of course,the following lines will not work as the "%" token
>     will be in uppercase on the right hand side.
>     >
>     > # building only the mod file
>     > LIB/%_M.mod: %_m.f90
>     >         @echo "building $*_M.mod"
>     >         touch $@
>
>     Can you make symbolic links that have the upper case, and feed
>     them to the compiler?
>
>       LIB/%_m.mod: %_m.f90
>               @echo "building $*_m.mod"
>               touch $@
>               ln -sf $@ $$(echo $*_m | tr [a-z] [A-Z]).mod
>
>     Here I have a working sample where I use a target-specific
>     assignment to define
>     a variable called $(UC) which holds the upper-cased stem:
>
>       $ make
>       make: *** No rule to make target 'foo.in <http://foo.in>',
>     needed by 'foo.out'.  Stop.
>       $ touch foo.in <http://foo.in>
>       $ make
>       making foo.out (upcased as FOO.out)
>
>     Contents:
>
>       $ cat Makefile
>       %.out: UC = $$(echo $* | tr [a-z] [A-Z])
>
>       %.out: %.in#
>               @echo "making $*.out (upcased as $(UC).out)"
>
>       foo.out: foo.in <http://foo.in>
>
>     The $* is evaluated at the time the target is dispatched, because it's
>     in a target-specific assignment. The $$ on $$(echo ...) ensures
>     that we get a literal $(echo ...) passed to the shell during
>     recipe execution.
>
>     Hope I didn't do anything wrong that makes this inapplicable to
>     your case!
>
>     Cheers ...