\edef containing utf8 umlaut characters

Question

Is it somehow possible to have utf8 characters (e.g. german umlauts) inside \edef ? The following example document behaves fine if the umlaut is given directly as argument like \convertTest{Müller}. The resulting output is "Mueller" after conversion. If "Müller" is given via a \edef like in \convertTest{\authorname}, the resulting output is not as expected ("M.ller"), although the message on \newcommand entry (before convert) is the same.

\documentclass{article}

\usepackage[ngerman]{babel} 

\usepackage[utf8]{inputenc}  
\usepackage[T1]{fontenc}  
\usepackage{stringstrings}


\newcommand{\convertTest}[1]{  
\message {before convert: #1}  
\convertword[q]{#1}{}{ue}  
\message {after convert: \thestring}  
\thestring \ 
}

\begin{document}  
\convertTest{Müller}  
\edef\authorname{Müller}
\convertTest{\authorname}  
\end{document}

Answer 1

Martin is right when he says : an alternative might be to protect ...

You can to do this with

\documentclass{article}
\usepackage[utf8]{inputenc}  
\usepackage[T1]{fontenc}
\usepackage[ngerman]{babel} 
\usepackage{stringstrings}


\newcommand{\convertTest}[1]{  
\message {before convert: #1}  
\convertword[q]{#1}{}{ue}  
\message {after convert: \thestring}  
\thestring \ 
}

\begin{document}  
\convertTest{Müller}
\makeatletter  
 \protected@edef\authorname{Müller}
 \makeatother   
 \authorname
\convertTest{\authorname}  
\end{document} 

You need to use an engine with etex but pfd(e)tex is very common actually

Alternativ : you can put the definition in your preamble :

\makeatletter   
\newcommand{\defauthorname}[1]{ 
 \protected@edef\authorname{#1}
 }
\makeatother 

and you can use this with

\defauthorname{Müller}  
 \authorname  

Answer 2

The umlauts are usually defined as active characters, e.g. like ~. They are basically single-character macros which do not need the backslash in front of them.

Now, \edef expands all macros, and therefore also all active characters, until the resulting tokens cannot be expanded any further. This is whats causing the problem: after \edef the ü isn't a ü any longer:

\show ü
> ü=macro:
->\UTFviii@two@octets ü.

I ran into the same issue when adding group names to my svn-multi package, which should be used inside \csname ... \endcsname, which also expands like \edef (but causes an error if the result is not text-only).

Solution: You could redefine the ü locally prior to the \edef: If all what you want to do is to turn the ü into ue then do this:

\def ü{ue}
\edef\authorname{Müller}
\message{\authorname}%   Prints Mueller!

An alternative might be to protect the ü, e.g. make it robust. I tried the \robustify macro from etoolbox but it didn't work. But this might be an issue with my local encodings.

You might also try to use XeTeX, which has native unicode support. But I personally don't have any experience with it.

Answer 3

\documentclass{article}
\usepackage[T1]{fontenc}  
\usepackage[utf8]{inputenc}  
\usepackage[ngerman]{babel} 
\usepackage{xstring}

\newcommand\convertTest[1]{\StrSubstitute{#1}{ü}{ue}}

\begin{document}  
\convertTest{Müller}  
\edef\authorname{Müller}
\convertTest{\authorname}  
\end{document}