Re: Umlauts in LaTeX export

[Nick Dokos <nicholas.dokos@hp.com>]

Jean-Marie Gaillourdet <jmg@gaillourdet.net> wrote:

> Dear Richard,
> 
> Stefan Vollmar <vollmar@nf.mpg.de> writes:
> 
> > Dear Richard,
> >
> > sitting in front of a German keyboard, writing 
> >
> > Gödel
> >
> > seems to be the obvious solution for modern LaTeX and Emacs
> > versions - you could define some shortcut to insert the appropriate
> > Unicode character into your text (as your keyboard probably does not
> > feature a "ö" key), or copy/paste the Umlauts from another Emacs
> > file as necessary. If you do not need it very often, this might be a
> > reasonable alternative.
> 
> Although I am german, I use an american keyboard layout for coding and
> everything else. But there is a nice emacs solution to enter umlauts:
> =C-x RET C-\ german-postfix RET= This enables an input method which
> allows you to enter all german umlauts: ä ü ö Ä Ü Ö and ß. 
> 
> Entering an `a' followed immediately by an `e' generates an ä, followed
> by another `e' it becomes `ae`, similar for ü and ö . `s` followed by
> `z` generates an `ß`. Larger variants are typed by typing two large
> letters.
> 

There are a couple of assumptions here (and in Eric F.'s mail about the
TeX input method as well). One is that the buffer is encoded in UTF-8:
if you use e.g iso-8859-1, you can use whatever input method you want,
but you'll end up with a byte in your file that LaTeX won't like.

The second assumption (which is satisfied by default when an org file is
exported to LaTeX) is that \usepackage[utf8]{inputenc} is used in the
LaTeX file.[fn:1]

Assuming that both of these assumptions are satisfied, this is indeed
the best way to deal with umlauts, accented characters, cedillas and
the like: the buffer *looks* like the LaTeX output. But remember that
there is an encoding there nevertheless.

Nevertheless that does not absolve org from dealing with \" properly. In
fact, it deals with it correctly in a heading but not in the text:


--8<---------------cut here---------------start------------->8---

* G\"odel

G\"odel
--8<---------------cut here---------------end--------------->8---

gives:

--8<---------------cut here---------------start------------->8---
...
\section{G\"odel}
\label{sec-1}


G\''odel

--8<---------------cut here---------------end--------------->8---

However, surrounding the o with braces breaks things in both places.

I think part of the problem is that headings and text go through
different processing: e.g. text goes through org-export-latex-content,
whereas headings don't. So fixing a problem like this in one place is
not enough.

Nick

Footnotes:
[fn:1]  This may or may not be correct if you use omega or xetex or
one of the more recent TeX variants, but I don't know much about them.