Werner LEMBERG wrote:
I just upgraded to Fedora 5, and now I get CJK problems too. What
do you suggest as a solution? A font is accessed as CIDFont if the
font doesn't support glyphnames.
Two issues.
. If we include a TrueType font as Type 42, it must contain a
/CharStrings dictionary which assigns glyph names to all glyphs.
With other words, there are always glyph names. Since lilypond
builds the Type 42 by itself, it should automatically have access
to this data. Such a font is to be accessed with the default
`findfont' interface, AFAIK.
How do I get the Unicode codepoints for a TT glyph from freetype? In
pango-font.cc we only get a glyph index. Also, on windows there are
further problems with fonts being embedded in duplicate (blowing up PDF
file sizes) if we embed type42 fonts, which is why the windows platform
uses GS' loadfontfile routine by default.
Looking into xiao-haizi-guai-guai.ps I see that the /CharStrings
dictionary is empty for the CJK fonts. This is a bug, but the
solution is simple: Just convert the data in the font's Unicode
`cmap' table to
/CharStrings XXX dict dup begin
/.notdef 0 def
/uniXXXX 1 def
/uniXXXX 2 def
....
Using `/uniXXXX' is recommended to make the ToUnicode feature of
PDFs work correctly while converting from PS to PDF.
. Accessing CID-keyed fonts (this is, fonts which come as PS CIDfont
resources, not wrapped as an OTF) is not supported yet, right? It
I guess not, we just make GS read the resource, and hope everything works.
--
Han-Wen Nienhuys - [EMAIL PROTECTED] - http://www.xs4all.nl/~hanwen
LilyPond Software Design
-- Code for Music Notation
http://www.lilypond-design.com
_______________________________________________
bug-lilypond mailing list
[email protected]
http://lists.gnu.org/mailman/listinfo/bug-lilypond
