<div>
I'm trying this with a font that has the kerning data in the "kern" table as well as in the "GPOS" table. (Is that possible?)
</div><div><br></div><div>Can you show me a font that this feature works with?</div>
<div><div><br></div><div>-- </div><div><p style="margin: 0px; font-size: 12px; font-family: 'Trebuchet MS'; color: rgb(35, 35, 35); "><b>Lóránt Pintér</b></p>
<p style="margin: 0px; font-size: 12px; font-family: 'Trebuchet MS'; color: rgb(35, 35, 35); ">Developer at <a href="http://prezi.com"><span style="color: rgb(18, 85, 204); ">Prezi</span></a></p></div></div>
<p style="color: #A0A0A8;">On Wednesday, February 20, 2013 at 10:26 AM, Jonathan Kew wrote:</p>
<blockquote type="cite" style="border-left-style:solid;border-width:1px;margin-left:0px;padding-left:10px;">
<span><div><div><div>On 19/2/13 23:35, Lóránt Pintér wrote:</div><blockquote type="cite"><div><div>Using ZWNJ would be a great way of fixing it. And indeed it splits the</div><div>ligature all right, but it also destroys the kerning. Here's a little</div><div>test I did with HarfBuzz @ e0486fc1affd3796fb8f664e2e7fc208f1d2106c:</div><div>(the font has an "fi" ligature, and has some kerning between "F" and ",")</div><div><br></div><div>Splitting the ligature with the ZWNJ works:</div><div><br></div><div>Shaped "fif\u200Ci" with features: [ ]</div><div>Glyph: #682, x_advance: 652</div><div>Glyph: #71, x_advance: 380</div><div>Glyph: #3, x_advance: 0</div><div>Glyph: #74, x_advance: 292</div><div><br></div><div>Kerning across the ZWNJ does not:</div><div><br></div><div>Shaped "F,F\u200C," with features: [ ]</div><div>Glyph: #39, x_advance: 483</div><div>Glyph: #13, x_advance: 242</div><div>Glyph: #39, x_advance: 563</div><div>Glyph: #3, x_advance: 0</div><div>Glyph: #13, x_advance: 242</div><div><br></div><div>Is it possible that doing this will be supported later in HarfBuzz?</div></div></blockquote><div><br></div><div>Hmm - is that font using a legacy 'kern' table, or a GPOS 'kern' </div><div>feature? For the former, I can understand that kerning would break, as </div><div>it's a naïve glyph-pair lookup, but for the latter, I thought we should </div><div>now ignore the ZWNJ.</div><div><br></div><div>JK</div></div></div></span>
</blockquote>
<div>
<br>
</div>