<div dir="ltr"><div>Hi Dominik,<br><br></div>can you try the same text with another font? I remember a similar problem with csfonts a few years ago. For explanation, if you have to dra "a" glyph in the font, you start by moving from the reference point to the beginning of the outline. Now you want to draw "á". You have "a" in the font, you have the acute accent. A naive idea is to draw the accent first, then move back to the reference point and then draw "a". Doing this you have two move operations in a sequence, one from the end of the accent outline to the reference font, another from the reference point to the beginning of the "a" outline. this is not allowed. Some rasterizes can cope with it and the result looks as expected, some rasterizers do funny things (they usually delete one of the moves. It might be your case, so probably the font contains a sequence of two moves for the dotbelow accent.<br></div><div class="gmail_extra"><br clear="all"><div><div class="gmail_signature">Zdeněk Wagner<br><a href="http://ttsm.icpf.cas.cz/team/wagner.shtml" target="_blank">http://ttsm.icpf.cas.cz/team/wagner.shtml</a><br><a href="http://icebearsoft.euweb.cz" target="_blank">http://icebearsoft.euweb.cz</a></div></div>
<br><div class="gmail_quote">2016-01-19 7:09 GMT+01:00 Arthur Reutenauer <span dir="ltr"><<a href="mailto:arthur.reutenauer@normalesup.org" target="_blank">arthur.reutenauer@normalesup.org</a>></span>:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"> Answering for Dominik, as that is easy to check by inspecting the PDF<br>
file (the original one, obviously):<br>
<span class=""><br>
> I assume you have already considered: are the fonts embedded in the PDF?<br>
<br>
</span> Yes.<br>
<span class=""><br>
> Did he enter the characters as precomposed combinations or by using<br>
> combining marks? First option more likely, I imagine.<br>
<br>
</span> Irrelevant, but you're right, it's the first option.<br>
<span class=""><br>
> This is a real shot in the dark, but here goes:<br>
> Is it possible that, at some point in the process, the precomposed<br>
> characters were decomposed and then put back together in a way that affected<br>
> the output?<br>
<br>
</span> Possible in general, but that's not what happened here. The<br>
problematic characters are coded in the PDF as single glyphs.<br>
<br>
Best,<br>
<br>
Arthur<br>
<div class="HOEnZb"><div class="h5"><br>
<br>
--------------------------------------------------<br>
Subscriptions, Archive, and List information, etc.:<br>
<a href="http://tug.org/mailman/listinfo/xetex" rel="noreferrer" target="_blank">http://tug.org/mailman/listinfo/xetex</a><br>
</div></div></blockquote></div><br></div>