<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=utf-8">
</head>
<body text="#000000" bgcolor="#FFFFFF">
<div class="moz-cite-prefix">Zdenek Wagner wrote:<br>
</div>
<blockquote type="cite" cite="mid:CAC1phyY=iZPQO1xaJ+nV2ESe6je7b9kprzNKWDKYv_xj7fEjDQ@mail.gmail.com">
<br>
<blockquote type="cite">
<pre wrap="">
I can now confirm that XeTeX by itself does not misplace the final vowel in my sample text — the problem is seemingly introduced by the use of TikZ's "style characters={ཕེབས་པར་དགའ་བསུ་ཞུ།} with {character command=\Tibetan}"
</pre>
</blockquote>
<pre wrap="">
It depends what exactly you do in TikZ. Especially when typesetting
along a curved path TikZ can insert some code between codepoints and
then the shaping engine does not see the correct groups of characters.
It might help to enclose a consonant with its vowel into braces so
that TikZ considers it one token and does not insert anything in
between.
</pre>
</blockquote>
<br>
Yes, I am experimenting with this as I write — I already use grouping for (e.g.,) Nepali & Arabic :<br>
<br>
<blockquote type="cite">style characters={{स्वा}{ग}{त}{म्}} with {character command=\Nepali},
<p style=" margin-top:0px; margin-bottom:0px; margin-left:0px;
margin-right:0px; -qt-block-indent:0; text-indent:0px;">
style characters={{أهلاً}{و}{سهلاً}} with {character command=\Arabic},</p>
</blockquote>
<br>
but am hampered in trying to do the same with Tibetan due to TeXworks inconsistent treatment of <delete> and <cursor-left/right>. With <delete>, I can break
<font size="+3">ཞུ།</font> down into four component elements, but my cursor keys treat it as inseparable ...<br>
<br>
** Phil.<br>
<br>
</body>
</html>