> I've just tested it on Windows 7 x64, with a fully updated MiKTeX 2.8
(MiKTeX-XeTeX 2.8.3758 (0.9995.1 texlive svn 15079) (MiKTeX 2.8)).
>
> I recall having this exact problem before with Asana Math and (possibly)
Neo Euler, but not Cambria Math.
>
> I removed the earlier installed version of unicode-math, and ran your
test case and it now works fine for me. However, if I install the copy
of Asana-Math.otf you provided locally and replace
> \setmathfont{Asana-Math.otf} with \setmathfont{Asana Math} (i.e. using
the version from my Fonts folder instead) the problem returns and I get
the output you had in your original email. I've also retested this with
Cambria Math and Neo Euler (not using local copies, but from the Fonts
folder), both of which work correctly.
Thank you very much, Michael Lynch and Khaled Hosny, for very helpful
comments, and Will Robertson -- for his posting (on my behalf) which led
to the workaround suggested by Mike.
That was very helpful indeed. I confirm to have Asana-Math.otf both in the
system font directory C:\Windows\Fonts and in my MiKTeX 2.8 tree. (the
latter must have been installed without me ever noticing it when I was
upgrading from MiKTeX 2.7 to 2.8; before, when I had Asana Math only in my
system font directory I wasn't experiencing the described problem).
I didn't have to remove Asana-Math.otf from C:\Windows\Fonts. I just
followed Mike's suggestion and replaced
\setmathfont{Asana Math}
by
\setmathfont{Asana-Math.otf}
in my XeLaTeX source file.
No more off-by-one shift in character mapping. Everything looks as it should.
Many thanks to Will Robertson for his 'unicode-math' package and to
Apostolos Syropoulos for Asana Math.
Mariusz Wodzicki
--------------------------------------------------
Subscriptions, Archive, and List information, etc.:
http://tug.org/mailman/listinfo/xetex
