From mboxrd@z Thu Jan  1 00:00:00 1970
Message-ID: <003201c3c2c1$fe6340a0$c48058db@laptop>
From: "Chris Hollis-Locke" <chris@hollis-locke.com>
To: <9fans@cse.psu.edu>
References: <1ced80d01a18ec54a5b7a944fddd9ca4@orthanc.cc.titech.ac.jp>
Subject: Re: [9fans] catenating subfont files, catsf
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: 7bit
Date: Mon, 15 Dec 2003 17:14:51 +1300
Topicbox-Message-UUID: a3be9740-eacc-11e9-9e20-41e7f4b1d025

----- Original Message -----
>From: "YAMANASHI Takeshi" <uncover@beat.cc.titech.ac.jp>
>
> Ttf2subf is great, but the ouptut subfont files have
> been sliced into very small pieces.
>
> For example, I tried ttf2subf on the Cyberbit font and it
> produced a font(6) file of almost 2000 lines.  Furthermore,
> in cerntain ttf, the result was almost 4000 lines and it
> slows rio considerably using the font.
>
> So I made a program catenating multiple subfont files
> into a single file.
>

Wouldn't it be better to fix ttf2subf rather than add more utils to fix
problems with existing ones?

I showed how to do it in a limbo prog I posted here a few months ago.
You map subfonts to standard unicode ranges as defined by unicode.org.
A simple scan determines which pages are present, you then produce subfonts
with 0 width glyphs for
those in a range that do not exist in the source font - causing draw to
render a peterface.
(there are other obvious optimisations to minimise the amount of glyph data
per subfont).

I haven't looked at ttf2subf but the Limbo prog was very simple, it cannot
possibly be that difficult.
Alternatively you could always use my Limbo app.

 Chris.