On Mon, 2004-05-10 at 12:50, Jan Nieuwenhuizen wrote: > I'm having trouble to get gnomecanvas/pango to draw a character from > our customly-encoded feta font. [...] > What could happen, is that this custom encoded-font gets mapped (or > mapped partly, the `a' and `b' are still/also at latin1 positions) > onto the private unicode area (0xe0000 and up); so I've tried several > shots at that, remapping characters and converting to utf8 sequence, > but to no avail. Things that I would expect to work: - If you use a postscript font with the standard Adobe Unicode character names for PUA codepoints - If you use an OpenType/CFF font (or TrueType font, but you seem to have Type1 outlines) with a Unicode character map. If your font is not encoded in either of these two ways, then Pango has no way of knowing how to get from Unicode codepoint to glyph in the font. Chris Blizzard is working on some API enhancements to Pango to allow the application to control how this mapping works; this is needed for MathML support in Mozilla which uses various misencoded and custom encoded fonts. So, hopefully we'll have some more flexibility here for Pango-1.6. But in the short term, you'll certainly have the easiest time if you can adjust the encoding of the font. > The observed behaviour suggests that pango looks up (or remaps) > characters by name, which is just how I'd like to refer to them again. Pango has no code here other than what FreeType does internally. > So what I'm actually/probably looking for is something like: > > pango_addchar (&string, "font", "lilypond-feta, r16", > "char-name", "noteheads-0", 0); Generally, I think we'd leave this up to the application; the application needs to know what Unicode code points they want from the font. Pango only needs to know how to map Unicode code point to glyph in the font. Regards, Owen
Attachment:
signature.asc
Description: This is a digitally signed message part