<font face="georgia,serif"><span class="Apple-style-span" style="font-family: arial; ">2010/9/22 Ulrike Fischer <span dir="ltr"><<a href="mailto:news3@nililand.de">news3@nililand.de</a>></span></span><br></font><div class="gmail_quote">
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;"><div class="im"><br>
</div>xunicode has changed the definition of \implementTIPAtextx. It is<br>
now (in 0.95)<br>
<br>
\expandafter\ifx\csname scantokens\endcsname\relax<br>
\def\implementTIPAtextx#1{#1\egroup}<br>
\else<br>
\def\implementTIPAtextx#1{\scantokens{#1}\egroup}<br>
\fi<br>
<br>
Reverting to the old definition from 0.91<br>
<br>
\usepackage{xltxtra}<br>
\def\implementTIPAtextx#1{#1\egroup}<br>
<br>
removes the trailing space.<br>
<br>
So it is a bug in xunicode.<div><div class="h5"><a href="http://tug.org/mailman/listinfo/xetex" target="_blank"></a></div></div></blockquote><div><br></div><div>Thank you, Ulrike. Then I think we should ask Ross to have a look at this: if he used \scantokens (when defined), there must be a reason! ;-)</div>
<div><br></div></div><blockquote class="webkit-indent-blockquote" style="margin: 0 0 0 40px; border: none; padding: 0px;"><div class="gmail_quote"><div> Paolo</div></div></blockquote>