🚀 go-pugleaf

We have fixed the problem with dvips and the Blue Sky type-1 fonts by
upgrading dvips to dvips v4.965 (or so . . .). Thanks again to the info-tex
crowd. Here is the low-down:

  George N. White III (Internet: GWhite@BIOnet.BIO.DFO.ca) informed me
that he was using a newer version of dvips successfully with the Y&Y type-1's.

Berthold Horn (bkph@ai.mit.edu) wrote:

>
> Also, older versions of DVIPS are unable to handle the repeated
> encoding used in the BSR CM fonts.  Because many systems cannot handle
> characters in the 0 - 31 range, these appear twice in the encoding vector.
> Which normally works fine.  Old versions of DVIPS replace T1 font
> metrics with information from the TFM file, and the TFM file knows
> nothing about the repeated encodings.
>
> There are several solutions to this latter problem, getting the latest
> version of DVIPS, or stripping out the repeated encodings from the fonts,
> or modifying dvips.pro to reverse the order in which it steps through
> the encoding.
>

  We were able to ftp and build the version of dvips now on
labrea.stanford.edu, and it works fine with the fonts we have (PC versions
of the BSR fonts).

  I looked at the fonts, and the encoding vector is in the cleartext portion
of the font. It would be simple to just delete the character encodings above 127
in the vector, as Berthold Horn suggests. (Assuming you have an editor that
won't kill files with binary data.) It seems even easier to change the
dvips prolog code (/rf in texps.pro, where the /Metrics vector is constructed
in v5.47?). This is what I would have tried if the dvips update did not work.

  Also, I may have resent the question instead of this short summary. Sorry.

  -- CHamlin, mis@aps.org