Re: Greek and Russian in Groff

[G. Branden Robinson]

Hi Deri,

At 2023-03-26T00:37:24+0000, Deri wrote:
> Aah, the gropdf -yU flag is a bit misleading! It does switch gropdf to
> embed the URW fonts, but it has no effect on groff itself, so a call
> to .ft TR will cause troff to look at the default foundry troff font
> TR, which lacks the cyrillic glyphs which are defined in the U-TR
> font.

Yup.  The light bulb is on now.  :)

> The special character "uxxxx" not defined message comes from troff
> when it is searching the TR font.

Agreed.

> A little background. While writing gropdf talking to Werner about the
> best way to utilise the extra glyphs available in the newer versions
> of the URW fonts, he came up with the concept of defining foundries as
> a way of switching between two sets of compatible fonts. So I
> implemented the -y flag in gropdf, the equivalent -y flag for troff
> never got implemented so the only way to tell troff to use the U
> foundry is to use tricks such as -f U-T on groff command line, or .fam
> U-T or .ft U-TR in your document. Incidentally, using any of these
> "tricks" makes the use of -P-yU redundant since groff will pass the
> full font name "U-TR" in the groff_out(5) stream passed to gropdf.

Yes.

It's difficult for me to see how we can seriously expand groff's font
versatility without tackling the foundry issue much as you have outlined
above.  But I've looked at the code for resolving fonts names, and it's
not a trivial amount of work.  That's probably why it was never done.

Thanks for clearing this up!

Regards,
Branden

signature.asc
Description: PGP signature