Re: [Groff] numeric chars are broken

[Werner LEMBERG]

> And here is the trivial patch:
> 
> $ diff -u groff_char.man  groff_char_new.man
> --- groff_char.man      2002-07-20 03:17:28.000000000 -0500
> +++ groff_char_new.man  2003-01-04 14:12:19.000000000 -0500
> @@ -600,7 +600,7 @@
>   .
>   .P
>   In groff, each 8bit input character can also referred to by the construct
> -.BI \en[char n ]
> +.BI \eN'n'
>   where
>   .I n
>   is the decimal code of the character, a number between 0 and\ 255

This is wrong.  In groff, each 8bit *input character* can also be
referred to by the construct \[charX] where X is the decimal code
of the character, a number between 0 and 255.  Glyph indices are not
limited to 256 positions.  Remember this scheme:

  input char -> internal char/glyph entity -> output glyph index/name.

For example, input character `A' and `\[char65]' are the same.

For latin-1, the input character `ü' is mapped to entity `\[:u]' which
in turn is mapped to glyph name `udieresis' for devps, to SGML entity
`ü' for devhtml, to output character code `0xFC' devlatin1, etc.

The mapping between latin-1 input characters and groff entities is now
in file latin1.tmac.  After loading this file input character
`\[char252]' (which is latin-1 `ü') is mapped to `\[:u]', using the
`.trin' request.

Anyway, I've fixed the typo, thanks.


    Werner