Interesting, I see the difference also, ./hb-view UthmanicHafs1\ Ver09.otf --direction=ltr ١٢
▁▁▅▅▄▁ ▂▃▊▂▃ ▂▇ ▃▂ ▁▖▅▃▖▉▅ ▗▖ ▃▆▟▖ ▗▟ ▆▂ ▙▗▃▙▙▄▃▟ ▂▆ ▙▚ ▗▟▖▙▆ ▇▅▗▁▙ ▄ ▗ ▖ ▙ ▏ ▊ ▅▅▟ ▋▌ ▋▍ ▙ ▉ ▉ ▙▍ ▋▌ ▚ ▟ ▙ ▃ ▙▖▅▟▃▁ ▁▃▖▅▗▟ ▙▄ ▗▂ ▚▅▆▚▞▅▆▌ ▂▃ ▗▟ ▄▗▁▅▙▄▁ ▚▁▅▟▄▆▃▙▇ ▇▅ ▆ ▇▅ ▄▇ ▇▆▆▆ and ./hb-view /Library/Fonts/Tahoma.ttf --direction=ltr ١٢ ▖ ▙ ▂▁ ▁▃ ▖ ▙ ▄▄ ▄▄▅▆▇ ▙ ▖ ▋ ▍ ▖ ▉ ▋ ▎ ▊ ▍ ▊ ▍ ▊ ▍ ▋ ▊ ▍ on harfbuzz master, wanted to say font bug but I see CoreText is also deals with this better: [image: Inline image 1] data:text/html;charset=utf8,<div style="font-family: 'KFGQPC Uthmanic Script HAFS'; font-size: 900%">١٢ On Sat, Jun 24, 2017 at 1:07 AM, Eric Muller <emul...@amazon.com> wrote: > It seems that when: > > hb_buffer_set_script (fBuffer, HB_SCRIPT_ARABIC); > hb_buffer_set_direction (fBuffer, HB_DIRECTION_LTR); > hb_buffer_add_utf32 (fBuffer, "١٢"); // pseudo code, of > course > > then the glyphs are presented in the wrong order to the font features. You > can see this with the font available at <http://fonts.qurancomplex.gov > .sa/?page_id=42>. It's intended for the Quran (and only for the Quran; > word doc at the same URL), and ligates sequences of digits for the end of > ayah. The example above displays "end of ayah 21". > > I am using a slightly old version of Harfbuzz, so may be this has been > fixed? > > Thanks, > Eric. > > _______________________________________________ > HarfBuzz mailing list > HarfBuzz@lists.freedesktop.org > https://lists.freedesktop.org/mailman/listinfo/harfbuzz >
_______________________________________________ HarfBuzz mailing list HarfBuzz@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/harfbuzz