Actually, I fixed the old nroff/troff back in 2002 so that it supported 
most of the
groff extensions (except .while).  I have the source code if someone 
wants it,
unfortunately I based it on 4.4 encumbered, but it should be trivial to 
be constructive
with diff to remove those additions.

-- Ragge



Thorsten Glaser wrote:
> Joerg Sonnenberger dixit:
>
>   
>> On Sat, Sep 15, 2007 at 01:23:58PM +0200, Hubert Feyrer wrote:
>>     
>>>  * this is only C, no C++, right?
>>>       
>> Yes. C++ is a mess to implement. Note that I think the only piece of C++
>> in base is groff.
>>     
>
> Well, GNU groff is replaceable.
>
>   
>> Another thing to buy from SCO now :-)
>>     
>
> Putting something like this one:
> | http://cvs.mirbsd.de/src/usr.bin/oldroff/
> into base and making the full GNU groff suite a package was the
> thing I thought of as sensible. Benefit would be the ability to
> upgrade GNU groff more often (without syspkgs, even), less GPL’d
> crap and especially no C++ crap in the base system. Another thing
> from SCO (that code stems from 4.4BSD-alpha from TUHS).
>
> The code¹ is a mess, and there are bugs (sometimes, a line ending
> is missing, rarely on i386, more often so on sparc), but – except
> for terminfo(5) which has too many diversions for tbl(1) – it’s
> building all of our manpages rather finely.
>
> Just in case someone cares ☺ Theo would be interested in it too
> if there were someone to clean up the code and make it legible.
> I’m not too keen on that, though…
>
> bye,
> //mirabilos
>
> ¹) I have ditroff sources flying around here, but they don’t bear
>    either a copyright notice or a licence; mtimes are pre-Berne
>    convention, but as a non-US citizen I guess I still can’t use
>    it, and BWK didn’t respond to my enquiting email, so I stuck
>    with nroff (and ignored troff, no C/A/T typesetter here ☻).
>