atm I use groupby() to get all the contiguous runs of emoji codepoints + country code codepoints and then collapse the ZWJ’s. Still causes issues though, because multi-codepoint emojis come out like ligatures when they really should be treated like 1 atomic unit
[image: Inline image 1] On Tue, Jun 28, 2016 at 7:01 PM, Ebrahim Byagowi <[email protected]> wrote: > I think this file > <https://cs.chromium.org/chromium/src/third_party/WebKit/Source/platform/fonts/SymbolsIterator.cpp>, > specially the SymbolsIterator::consume, contains the logic you are looking > for. > (reposting directly sent mail) > > On Tue, Jun 28, 2016 at 9:17 AM, Kelvin Ma <[email protected]> > wrote: > >> Since emojis have to be shaped separately from all the other text, how do >> I separate out the emojis from the text? Previously I just had an enormous >> list of unicode emoji codepoints and I would filter them out beforehand, >> but this doesn’t work with ligature-emojis like family, flag, or skin color >> emojis, which don’t have singular unicode codepoints. What’s the preferred >> way for handling emojis in text shaping then? >> >> _______________________________________________ >> HarfBuzz mailing list >> [email protected] >> https://lists.freedesktop.org/mailman/listinfo/harfbuzz >> >> >
_______________________________________________ HarfBuzz mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/harfbuzz
