[lltx] announcing lualatex-doc

[Manuel Pégourié-Gonnard]

Hi Ulrike,

Le 08/11/2010 15:29, Ulrike Fischer a écrit :
> 1. Section 2.2.4 Font internals and 5.2 Partially working:
> 
> euenc no longer uses/contains  a modified version of xunicode.
> Fontspec now loads the original version (it defines temporarly
> \XeTeXpicfile to get around the "ifxetex"-test of xunicode).
> 
Arf, I should have know that. Thanks for correcting me.

> 2. Section 5.1. "Just working"
> 
> Soul and listings don't have a problem with utf8-chars broken into
> pieces.

I must admit I don't use them regularly, but Heiko has some wrapper packages
(soulutf8 and listingsutf8) so I assumed there was problems. (I also seem to
remember Denis mentioning problems with listings). I'll check the documentation.

> soul gets the numbers of characters in its argument by "printing"
> the argument with a tt-fonts and dividing the length.

Oh, nice trick.

> As a default
> it use ectt as tt-font and naturally this doesn't give the correct
> result if the argument contains glyphs above the first 256 positions
> (e.g. greek or €). The following will give some suprising result
> with xelatex:
> 
> \documentclass{article}
> \usepackage{soul}
> \usepackage{fontspec}
> \makeatletter 
> % \font\SOUL at tt="LMMono10-Regular" %helps with xetex
> \makeatletter
> \begin{document}
> Euro: \ul{ab€€€} \ul{a€€€b}
> \end{document}
> 
> 
> (It fails with my lualatex (0.60.2) because of a problem discussed
> in july on the luatex list. I don't know if this problem has been
> solved in newer versions).
> 
Ok, thanks for the information.

> listings has a similar problem: It activates and defines output only
> for the first 256 characters. So you can get quite suprising output
> if your input uses chars from outside this range:
> 
> \documentclass{article}
> \usepackage{fontspec}
> \usepackage{listings}
> \makeatletter
> %\lst at CCPut \lst at ProcessLetter {"20AC}\z@
> %\lstset{MoreSelectCharTable=\catcode`\€\active}
> \makeatother
> \begin{document}
> \begin{lstlisting}[columns=flexible]
> a€b
> \end{lstlisting}
> 
> x \lstinline|a€b| y
> \end{document}
> 
> On the whole one should not forget that "256" was a long time a
> limit and that this limit is often burried quite deep in some
> packages.
> 
Right. I remember running some tests a while ago, and I suppose I just forgot to
test with characters higher that 256 (I was already pretty happy that things
worked with some multi-byte characters).

I just corrected the xunicode part, and quickly adapted the rest, but I'll
rework it more thouroughly later, I need to get my own ideas clear on this
subject. I hope the temporary version is acceptable in the meanwhile.

Thanks for your very useful remarks,
Manuel.

> 
> 
>  
>