Stephen J Turnbull <turnbull(a)sk.tsukuba.ac.jp> writes in xemacs-beta(a)xemacs.org:
>>>>> "N" == N J Doye <nic(a)niss.ac.uk>
writes: 
N> How hard (he says na,Aovely) is it to keep a mule package,
N> non-mule compliant?
 If it's a package that we synch from the FSF distribution of GNU
 Emacs, that's going to depend on the maintainer.  If the maintainer is 
 not an ISO-8859-1 kinda guy, I would suspect that Mule dependencies
 will start crufting on pretty quickly. 
Yes, despite any of our potential efforts to head it off too, I'm
afraid.
 For various reasons (including the instability of the API) it's
 difficult to try to automatically do something sensible; the only way
 is to do the synching manually. 
There currently is no API, documented or otherwise.  It's more a
distributed set of various functions thrown together.  I will be
writing up formal Lispref-level documentation on what exists in the
coming weeks.  I see this as being our best chance to have any input
into where Mule goes from here.
 Unfortunately, unless somebody wants to take that on, the likely
 result is that the last nomule version will freeze and bitrot. 
Right.  The process has already started to some extent.
-- 
板挟み