[Freetype]Re: Encodings (was Re: dviv, a dvi viewer for freetype2)

[Juliusz Chroboczek]

C> charmap 0: platform: 1, encoding: 0

That's Macintosh Roman.  In theory.

C> charmap 1: platform: 3, encoding: 1

Microsoft Unicode.  In theory.

Let me explain.  In theory, points in (1,0) are supposed to identify
glyphs corresponding to characters in the MacRoman encoding.  Points
in (3,1) are supposed to identify glyphs corresponding to characters
in the Unicode BMP, with whatever additions Microsoft deem desirable

However, the TrueType format doesn't provide a way to identify
encodings that are non-standard but are text encodings.  For that
reason, many font designers misuse these encodings by:

  - using the MacRoman encoding for non-standard encodings;
  - using the Microsoft Unicode encoding and putting their
    (non-standard) glyphs at the positions where CP 1252 (current or
    older version!) glyphs ought to be.

There is therefore no good answer to your question.  In my experience,
Microsoft encodings tend to be more reliable than Apple ones (due to
more testing on Windows, I suspect).  What the right approach is, only
experimentation will show; perhaps you should check what pdftex does,
and behave the same.

For Type 1 fonts, you should probably obey the dvips encoding files
scheme rather than the FreeType encoding scheme.  Using the custom
encoding will only work for fonts installed without the `-p' flag to
`afm2tfm'.

Regards,

                                        Juliusz Chroboczek