Since patching The Library <https://github.com/behdad/harfbuzz/pull/299> appears to be a nonstarter, then what is the correct, *workable* way to unpack the <hb_glyph_info_t> and <hb_glyph_position_t> objects given by the harfbuzz PyGObject buffer? As I have mentioned many, many times, performing dynamic lookups on four struct attributes is too inefficient to be useful for extracting shaped text from “the real world”. It is over ten times slower than the original C function.
Unless an alternative method exists, I’m afraid the harfbuzz python bindings are only useful for shaping trivial amounts of text.
_______________________________________________ HarfBuzz mailing list HarfBuzz@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/harfbuzz
