[Fontinst] Adjusting kerning information before reglyphing

[Lars Hellström]

At 21.42 +0200 2005-07-24, Stephen Hartke wrote:
>I'm working on making Arev Sans available for use as a math font in
>TeX.  I'm mostly done, and am now working on correcting the horizontal
>placement of math accents. This is implemented as a kern between the
>character and a skewchar.
>
>Here is what happens in the fontinst file:
>1. Afm data is read in and reencoded to a custom encoding.  This
>allows access to the glyphs I want (otherwise, many would be past the
>256 limit).
>2. The font is reglyphed so that names match up with the names oml.etx
>is expecting.
>3. \installfont is called.
>
>I also have an mtx file that adjusts the kern data to fix the math
>accents.  The mtx file was created by a script that pulled the info
>from FontForge; this way I can create the kerns visually, which makes
>it much faster for me.  The problem is that all of the names are the
>original glyph names, and hence I need to apply the adjusted kerning
>_before_ I reglyph.  Is there some way to apply an mtx file to another
>mtx file without calling installfont? Since this is just a temporary
>in-between font used only within the fontinst script, it seems silly
>to generate all of the things that go along with installfont.

\installfont doesn't apply metric files to each other, it just reads them
in sequence, pretty much as if there was one file which just \inputmtx'ed
them all in sequence.

What you should do is to have a second \reglyphfont applied to your kerns
file. There is no requirement that there has to be glyph definitions in
files \reglyphfont is applied to, it can be just as well be only kerns.
\reglyphfont can also be used to filter out data from this file, if there
is anything there that you don't want.

Lars Hellström