[commit: ghc] master: Change how we handle libffi (b30015e)

[Edward Z. Yang]

Excerpts from Ian Lynagh's message of Mon Mar 25 04:29:05 -0700 2013:
> On Mon, Mar 25, 2013 at 02:20:18AM -0700, Edward Z. Yang wrote:
> > 
> > ezyang at javelin:~/Dev/ghc-build-tick$ find -name *Cffi*
> > ./rts/dist/build/libCffi_p.a
> > ./rts/dist/build/libCffi_thr_p.a
> > ./rts/dist/build/libCffi_debug_p.a
> > ./rts/dist/build/libCffi_thr_debug.a
> > ./rts/dist/build/libCffi_debug.a
> > ./rts/dist/build/libCffi_l.a
> > ./rts/dist/build/libCffi_thr_l.a
> > ./rts/dist/build/libCffi_thr.a
> 
> You should have a rts/dist/build/libCffi.a
> 
> Perhaps v isn't in your GhcLibWays?

Shouldn't it be added by default?  I'm using the new build.mk.sample,
which has this line:

    GhcLibWays = $(if $(filter $(DYNAMIC_GHC_PROGRAMS),YES),dyn,v)

DYNAMIC_GHC_PROGRAMS is YES by default on Linux; so it would seem that
we are not getting v in GhcLibWays...

Edward