Re: [MiKTeX] xetex/xelatex crash

[Ulrike F. <li...@ni...>]

am Sonntag, 6. Januar 2008 um 16:32 schrieb Christian Schenk:

> Christian Schenk wrote:
>> jr wrote:
>>> Anyway, it took some time, but I finally managed to create a minimal=20
>>> test case. Umlauts in a non-unicode file crash XeTeX. If you save the=
=20
>>> following as ANSI, it will crash. If saved as UTF-8, it won't.
>>=20
>>=20
>> FYI: XeTeX requires Unicode encoded input files. If you feed it with=20
>> "garbage", then anything can happen.

> Forgot to mention the \XeTeXinputencoding primitive:
> \XeTeXinputencoding bytes
> This will tell XeTeX that your input is platform-specific. I guess, this
> has other side effects.

The following work fine for a ANSI-file for xelatex and pdflatex:

\RequirePackage{ifxetex}
\ifxetex
\XeTeXinputencoding "latin1"
\fi

\documentclass{article}
\ifxetex
\usepackage{fontspec}
\setmainfont{Georgia}
\else
\usepackage[ansinew]{inputenc}
\fi

\begin{document}
f=FCr uns
\end{document}

But I must admit that I would like it too if xetex would recover a bit
more gracefully if one forgets the correct input encoding line.
Violation access errors are a bit frightening.

--=20
Mit freundlichen Gr=FC=DFen
Ulrike Fischer
mailto:li...@ni...