Re: [Groff] new backend for UTF-8 output

[Bruno Haible]

Markus Kuhn writes:

>   - I understand that
> 
>        An  extension  to the troff character set for Europe, E.G.
>        Keizer, K.J. Simonsen, J. Akkerhuis, EUUG Newsletter,
>        Volume 9, No. 2, Summer 1989
> 
>     have extended the troff character set to cover Latin-1. Is there
>     also a groff input syntax that allows me to enter any Unicode
>     arbitrary Unicode character by hex code?

You can already input Unicode characters using the N'ddddd' syntax, where
ddddd is the decimal value. I have two proposals (not contained in
yesterday's patch) for dealing with that.

1. Introduce a convenient hexadecimal input syntax  U'xxxx'
   where xxxx is the hexadecimal value.
   Compatibility issue: This would become one more difference between
   classical troff and groff.

2. troff and grotty need to know the metric of the character, but the
   font files don't contain all of them. I tried to make the 4 font
   files larger, by incorporating 3002 characters (those listed in Markus'
   TARGET3 file). It worked, but the processing times went up to 6 seconds
   for troff and 3 additional seconds for grotty, which is unacceptable.
   I could provide a patch which computes the metric using a portable
   unicode_wcwidth function, inside libgroff.

   But the background problem is: do we want to encourage people to use
   the whole set of Unicode characters in their groff sources, knowing
   that the results will be suboptimal for the other output devices?
   Or will someday the postscript backend be extended to support these
   many characters too?

What do you groff guys think about either proposal?

Bruno