Tilde/asciitilde, circumflex/asciicircum

[Lars Hellström]

At 04.48 +0200 2000-09-07, Pierre MacKay wrote:
>You asked about why Xdvi would show different results from
>dvips.  The reason is that Xdvi does not send the dvi through
>a PostScript translation, and only a PostScript translation
>makes use of "virtual Font" techniques.

That last claim simply isn't true. Virtual fonts are a way of substituting
a sequence of DVI commands for a simple SETCHAR command and have nothing to
do with postscript as such. One can design a DVI driver to understand
virtual fonts no matter what its target is, and current DVI drivers
(regardless of their target) generally do.

I also find your claims about `fi' and `fl' being in encoding positions 256
and 257 rather remarkable, as postscript encoding vectors should only have
entries from 0 to 255 (except for base font for composite fonts, the manual
says, but I doubt that is what the case here). Not that there would be much
point in adding extra entries either, as postscript characters can only
have codes in the range 0--255. Could it perhaps be the case that you're
used to looking at fonts using Fontographer (or some similar program)? I've
noticed that it shows unencoded glyphs as having codes 256--, but that
isn't what really happens in the font.

VPtoVF shouldn't like character codes above 255 either (although I haven't
checked this); in any case the TFM format only accepts 8 bit character
codes.

Lars Hellström