Problem with character placement (XeLaTeX, Linux Libertine and Unicode)

Question

The following test document:

\documentclass{article}
\usepackage{fontspec}
\setmainfont{Linux Libertine O}
\begin{document}
aɻ
\end{document}

displays the second character above, instead of after the first, at least on my machine, when compiled with XeLaTeX.

Is this a problem with the font, XeTeX, or what exactly? (Of course, this also fails in my actual document)

$ xelatex -v
XeTeX 3.1415926-2.4-0.9998 (TeX Live 2012/Debian)
kpathsea version 6.1.0
Copyright 2012 SIL International and Jonathan Kew.
There is NO warranty.  Redistribution of this software is
covered by the terms of both the XeTeX copyright and
the Lesser GNU General Public License.
For more information about these matters, see the file
named COPYING and the XeTeX source.
Primary author of XeTeX: Jonathan Kew.
Compiled with ICU version 49.1 [with modifications for XeTeX]
Compiled with zlib version 1.2.7; using 1.2.7
Compiled with FreeType2 version 2.4.10; using 2.4.11
Compiled with fontconfig version 2.10.1; using 2.10.2
Compiled with libpng version 1.2.49; using 1.2.49
Compiled with poppler version 0.20.5

$ uname -a
Linux Rykos 3.8.0-24-generic #35-Ubuntu SMP Fri May 31 14:07:41 UTC 2013  x86_64 x86_64 x86_64 GNU/Linux

Answer 1

Same problem here, but with Arial Unicode it works, so it looks as if the problem is partly due to the font. You can insert a kern to avoid the misplacement: a\kern0ptɻ.

Answer 2

That is a bug in the font, the glyph for ɻ has an anchor point of type mark instead of base glyph, so the rendering engine treats it like a combining mark and places it over glyphs that have the same anchor but with base glyph type. This should be reported to the font developers.

The fact that LuaTeX does not place it above the preceding glyph is a bug in the layout engine used (luaotfload here).

Answer 3

The font advertises correctly the character width, so you can use a zero kern.

\documentclass{report}

\usepackage{fontspec}

\setmainfont{Linux Libertine O}

\begin{document}

aɻ \textit{aɻ} \textbf{aɻ} \textbf{\textit{aɻ}}

\end{document}

Adding a zero width kern fixes the issue at the expense of losing kerning information with the preceding character, but it doesn't seem a big loss.

\documentclass{report}

\usepackage{fontspec}
\usepackage{newunicodechar}

\setmainfont{Linux Libertine O}

% fix the strange behavior
\newunicodechar{ɻ}{\leavevmode\kern0pt ɻ}

\begin{document}

aɻ \textit{aɻ} \textbf{aɻ} \textbf{\textit{aɻ}}

\end{document}

There is no such problem if one uses Libertinus Serif.