🚀 go-pugleaf

> 			....  I should mention that the reason (I surmise)
> it is not printing the floating keep is that the last bit of `normal' text
> does not fill the last page.  The vicinity of the last page is approximately 
> where the floating keep was entered in the original ditroff -me input. 
> The floating keep should ideally float all on its lonesome to the next page. 
> However, since there is no text on the next page - except for the floating
> keep itself -  `-me' says, "No point printing that diagram".

What you describe could well be the problem.  If the floating keep is
defined too close to the end of the output, and doesn't fit on what would
otherwise (without the keep) be the last page, it will never get output.
The problem is that the output of the keep is tied to a break set at top-
of-page, and the break is never getting triggered.  This is a bug in -me's
end macro, which should test for such conditions.  (It does test for hanging
footnotes, but evidently not for things waiting for the next page.)

A work-around is to force a new page by ending your input with a .bp (break
page) request.  I simulated your problem (with version 2.15 of -me) and this
does indeed fix it.  The only disadvantage is that if there is nothing to go
on the last page (say, the keep fits on the bottom of the last text page) then
you will get a blank page (except for headers and/or footers).  Unless you are
doing something with the number of pages emitted, or are seriously concerned
about paper waste, this won't be a problem.

-- 

Ted Hopp	{seismo,umcp-cs}!nbs-amrf!hopp