Re: [Groff] grog and mom.

[Werner LEMBERG]

> After fixing that, I found some existing problems, e.g. the first egrep
> passes through many more lines than intended.  However, on testing my
> changes I find grog.sh and grog.pl give different results on the same
> file, so they clearly don't encompass the same logic.
> 
> Can we pick one, preferably the more portable grog.sh, and work on
> that one being correct?  What advantage does the perl version have?

I'm a bad shell *and* a bad perl programmer, so I really can't
comment.  I fully agree with you that having two versions is not
necessary.  The advances in autoconf show what is really possible in
shell programming, so perhaps it is a good idea to concentrate on
perl.sh, using only standard tools (as documented in autoconf).

Will you do that?

> As for adding -mom detection, that's easy enough by checking for
> .PRINTSTYLE or .START.  But -mom documents use .PS and .TS so that
> triggers pic and tbl command line options.

Oh, oh.  This has slipped my first reading of the mom docs.  Using .TS
and .PS in mom is not a good idea...  Peter, any alternative ideas?

> I know "Mom has not been tested with any of the groff preprocessors
> (eqn, grap, grn, pic, refer soelim, tbl), although she works fine
> with .PSPIC). Interaction with them is not really in her design
> mandate", but re-using well-known pre-processor `tags' presumably
> causes problems unless GNU pic's behaviour of only matching /^\.PS$/
> is being relied upon?

I fully agree.

> What do people think grog's route should be?  If -mom then ignore .TS,
> .PS, etc.?

My preference would be a change in the mom macros.


    Werner