On 15-05-06 04:14 PM, Richard Wordingham wrote: > On Wed, 6 May 2015 23:30:20 +0200 > "Adam Twardoch (List)" <[email protected]> wrote: > >> 2. For the language tags, one should use high-level ISO-style >> language tags, while the low-level OT languagesystem tags "might" >> work, but it's not recommended to use them. > > My mistake. The code I suggested shouldn't even compile. One should > use hb_language_from_string(). However, I don't know what one should > use if there isn't a language associated with the 'language system'. I > suppose a good example would be special glyph shapes for IPA or APA.
That can be done using the "x-hbotXXX" extension. See my other messages. > Incidentally, looking at hb-ot-tag.cc, what's the big problem with the > Chin and Cree languages? The Cree styles are heavily split, while a > single style is assumed to suffice for the various Chin languages and > the northern Athabascan languages. The standard has not been fully > updated for splitting off some of the Athabscan languages! Or is one > supposed to look for a Navajo language-system for Navajo, and on not > finding it, look for a general Athabascan language-system? I did note > that there are some problems where Microsoft splits languages more > finely than ISO 693 - translating language naming systems has some of > the problems of translating natural languages! Since we have everyone's attention, lets do a cleanup of that list in light of new OpenType 1.7 language tags. Arthur already volunteered (off-list) to send a Pull Request, and we can all refine it there. Thanks, behdad _______________________________________________ HarfBuzz mailing list [email protected] http://lists.freedesktop.org/mailman/listinfo/harfbuzz
