On Wed, 15 Sep 2004, Felix Winkelmann wrote:
What if file a.scm defines a macro that file b.scm uses (by refering
to syntax defined in a module declared in a.scm)?
I regard the difficulty compiling this construction separately,
as evidence from God that Scheme, having abandoned the first-class,
mutable, storable, runtime code transforming macro, has fallen
short of the grace of LISP and uses the wrong macrology.
But we don't want to get into *that* mess, so I'll just point
to Flatt's "composable and compilable macros" as a tolerable
compromise - although it breaks true separate compilation.