Keith Marshall:

> > I agree, but I didn't find a public mechanism to
> > do that.
>
> Does the  attached  achieve  the  desired  effect?
> Agreed,  it isn't intuitive: it's very picky about
> the placement of the reassignment of register 'P',
> and  you  have to set it to less than zero for it,
> and the effect of the assignment to the 'N' regis-
> ter,  to  be  honoured, but this is achieved using
> only publicly documented features, (discovered  by
> a search for 'page number' within groff_mm(7)).

Yes, and this is exactly what hd@set-page does.  The
documentation doesn't say that P is  not  read-only.
I  consider  hd@set-page  as a setter of the field P
and therefore preferable to setting P directly.   If
it was not so trivial my method would be less likely
to break anything.  But you are right that your ver-
sion does not require knowledge of mm's internals.

Thanks for pointing it out, but I frankly am  unsure
as to which is better.

--
()  ascii ribbon campaign - against html e-mail
/\  www.asciiribbon.org   - against proprietary attachments