Re: [Devel] Glyph name suffixes

[Werner LEMBERG]

> It seems freetype mixes glyphes with and without suffixes when
> renders Type 1 fonts.  For example besides of "guillemotleft" glyph
> the "guillemotleft.cyr" may exist which differs from the first.  But
> when the text is rendered using SFRM1000 font from CM-Super
> collection through freetype "guillemotleft" is displayed as true
> "guillemotleft", but "guillemotright" as "guillemotright.cyr".

Which program are you using?  FreeType is just a library which
basically accepts glyph indices.  There is a function called
FT_Get_Name_Index which retrieves the glyph index for a particular
glyph, with absolutely no magic: If you input `guillemotright.cyr' as
the glyph name, you get the correct glyph index.

What you are probably referring to is the algorithm FreeType uses to
emulate a cmap.  Here FreeType correctly ignores the suffix while
looking up the AGL.  Depending on the glyph order, FreeType might pick
up either `guillemotright.cyr' or `guillemotright' -- both are equally
good candidates.  There is no means in FreeType to improve this, and
we can't implement something better: It is simply dependent on both
the application and the particular font.

SFRM1000 is just a glyph container, nothing more.  The correct means
to access this font is to use proper encoding vectors (which come with
the cm-super collection).  Everything beyond that is not possible in a
generic way with Type 1 fonts.  Vladimir Volovich would be happy, I
think, if you volunteer to convert the cm-super fonts into OpenType
fonts, adding a proper GSUB table which selects Cyrillic punctuation
if the language tag is set to `cyrl'.


    Werner