<div dir="ltr"><br><div class="gmail_extra"><br><div class="gmail_quote">On Wed, Aug 26, 2015 at 4:35 PM, David Carlisle <span dir="ltr"><<a href="mailto:d.p.carlisle@gmail.com" target="_blank">d.p.carlisle@gmail.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><span>On 26 August 2015 at 15:13, Hans Hagen <<a href="mailto:pragma@wxs.nl" target="_blank">pragma@wxs.nl</a>> wrote:<br>
<br>
<br>
><br>
> fyi: the number changes because a non-char node is checked (so it then<br>
> depends on what is in the 'char' mem slot of that node) ... the checking is<br>
> somewhat different from pdftex because we need to make sure that kerning<br>
> between the nucleus and first script char (direct or in math sub list) is<br>
> done right (advanced otf kerning)<br>
><br>
<br>
<br>
</span>Ah, so do you mean the fix will be not to make this check in the cases<br>
that the node is not a char node?<br>
<span><font color="#888888"><br>
David<br>
</font></span></blockquote></div><br><br clear="all"><div>it should be fixed in trunk.</div><div><br></div>-- <br><div>luigi<br></div>
</div></div>