On 22/01/2010 15:13, Frank Mittelbach wrote:
> if the two would work together by just using \cs_set_eq:NN for the moment then
> as a short measure we could do that until mhsetup gets sorted out. However, it
> might be the case that more is needed given that the mhsetup code is only
> "based on" expl3 concepts but was used to run standalone and was developed
> when expl3 was in an earlier stage.
>
> as Joseph said, we should sort it out one way or the other soon.

A fix at the mhsetup end seems to be to move the definitions to 
\AtBeginDocument:

\AtBeginDocument{
   \@ifundefined{if_num:w}{\MH_let:NwN \if_num:w =\ifnum}{}
   \@ifundefined{if_dim:w}{\MH_let:NwN \if_dim:w =\ifdim}{}
   \@ifundefined{if_case:w}{\MH_let:NwN \if_case:w =\ifcase}{}
   \@ifundefined{or:}{\MH_let:NwN \or:=\or}{}
}

(lines 93 onward).

Philipp, can you test this works in the "real world": if so something 
can go to CTAN today.
-- 
Joseph Wright