Jan Vroonhof <vroonhof(a)math.ethz.ch> writes:
Could you, just for the purpose of future generations (i.e. me in a
few weeks :-)) that want to fix up the modules, give a small summary
of these problems.
Basically, we'll want to revise the XEmacs C API, and decide which
functions/structs/... form the "interface", and which ones form the
implementation. Without modules, the distinction is clear: Lisp code
is interface, C code is implementation.
In the course of hacking XEmacs, I've often added or removed
parameters to functions. OTOH, some will likely never change,
e.g. Fcons(). If we are serious about the modules, we'll have to
decide which functions are safe for use without fear of them getting
changed, and document them.
Also, we'll probably need to prefix the random global objects with
`XEmacs_' to avoid collision.