Re: Failure in test silent5.test with heirloom make

[Ralf Wildenhues]

Hi Stefano,

* Stefano Lattarini wrote on Sat, May 01, 2010 at 06:40:41PM CEST:
> OK, I must admit to be stuck here.  How can I know what objects files 
> are expected to be compiled from, say, a bar.c file, in turn generated
> from a bar.l file?
> 
> Or (e.g.) how can I distinguish the following situation:
> 
>   foo_SOURCES = bar.l
> 
> (where bar.c is generated from bar.l, bar.o is generated from bar.c,
> and no explicit dependency rule is generated), from the following
> situation:
> 
>   foo_SOURCES = bar.l
>   foo_CFLAGS =  $(AM_CFLAGS)
> 
> (where bar.c is generated from bar.l, and foo-bar.o is generated from 
> bar.c, with an explicit dependency rule of foo-bar.o from bar.c)?

The logic for this happens in handle_single_transform, the variable is
$have_per_exec_flags, and is passed to lang_*_rewrite.  The logic
applies equally to the lex language (with foo_LFLAGS) as well as C (with
foo_CFLAGS).  The *target_hook functions can infer this information from
the $aggregate argument passed to them (again, see
handle_single_transform).

Hope that helps.

Cheers,
Ralf