[pdftex] ToUnicode map and virtual fonts

Werner LEMBERG wl at gnu.org
Sun Oct 1 07:09:15 CEST 2006

> see
> http://sarovar.org/tracker/index.php?func=detail&aid=580&group_id=106&atid=495

Uh, oh, looking into tounicode3.zip, there are the following severe

  . Automatic recognition of `uXXXXX' and `uniXXXX' glyph names is
    missing.  For big CJKV subfonts, it's simply unreasonable to add
    more than 20000 calls to \pdfglyphtounicode!

  . pdftex should automatically recognize AGL ligatures like `f_i_i',
    *without* a special \pdfglyphtounicode rule.  All such rules
    should be removed from glyphtounicode.tex.

    Is there any reason why pdftex doesn't implement the complete AGL

  . I dislike the automatic handling of `small', `oldstyle',
    `inferior', and `superior' suffixes which is against the AGL
    rules.  Fonts which don't follow the AGL need a bundle of calls to
    \pdfglyphtounicode anyway, so why does this exception exist?

  . Finally, \pdfglyphtounicode can't handle Unicode values outside of
    the BMP (this is, values >= U+FFFF) -- for example, old Vietnamese
    (using CJKV glyphs) needs that...

I consider the last item a real design error in \pdfglyphtounicode.


More information about the pdftex mailing list