Not all UTF8 symbols work in lstlistings

Question

I am on Linux and I have this document that I compile using xelatex main.tex:

\documentclass[12pt]{article}

\usepackage{color}
\usepackage{xcolor}
    \definecolor{code_linenumbers_b}{RGB}{215,215,215}
    \definecolor{code_linenumbers}{RGB}{255,255,255}
    \definecolor{code_background}{RGB}{250,250,250}
    \definecolor{code_comments}{RGB}{185,185,185}
    \definecolor{code_borders}{RGB}{86,211,239}

\usepackage{mathspec}
    \setallmonofonts[
        %Path=.,
        Extension=.ttf,
        Scale=0.8,
        UprightFont=*--regular,
        BoldFont=*--bold,
        ItalicFont=*--italic
    ]{liberation-mono}

\usepackage{listings}
    %Global
    \lstset{
        %frame
        frame=l,
        captionpos=b,
        numbersep=3.5mm,
        xleftmargin=0.35cm,
        xrightmargin=0cm,
        framesep=0mm,
        framexleftmargin=2.5mm,
        framerule=0.5mm,
        %colors
        backgroundcolor=\color{code_background},
        basicstyle=\ttfamily\color{black},
        numberstyle=\ttfamily\bfseries\color{code_linenumbers},
        fillcolor=\color{code_linenumbers_b},
        commentstyle=\color{code_comments},
        keywordstyle=\color{black}\bfseries,
        rulecolor=\color{code_borders},
        %code format
        showspaces=false,
        breaklines=true,
        breakatwhitespace=true,
        tabsize=3,
        columns=flexible
    }

\begin{document}
    \lstinputlisting[
        language=C,
        caption=\lstinline{main.c},
        label=r001,
        numbers=left,
        xleftmargin=1.15cm,
        framesep=8mm
    ]{main.c}

\end{document}

and my main.c file looks like this:

#include <stdio.h>

int main(){

    //            ir
    //            ↓
    //        012345678
    //      0     *
    //      1    ***
    //      2   *****
    //      3  *******
    // ir → 4 *********

    return 0;
}

But after compiling (with no errors or warnings) I get the resulting main.pdf where arrows ↓ (U+2193) & → (U+2192) are not properly positioned:

All the *.ttf fonts are inside the same folder than the main.tex and main.c.

ziga@ziga-laptop:mwe$ ls -l
total 468
-rw-r--r-- 1 ziga ziga 120308 May 21 17:30 liberation-mono--bold-italic.ttf
-rw-r--r-- 1 ziga ziga 107620 May 21 17:30 liberation-mono--bold.ttf
-rw-r--r-- 1 ziga ziga 126308 May 21 17:30 liberation-mono--italic.ttf
-rw-r--r-- 1 ziga ziga 110204 May 21 17:29 liberation-mono--regular.ttf
-rw-r--r-- 1 ziga ziga    229 May 21 18:45 main.c
-rw-r--r-- 1 ziga ziga   1479 May 21 18:50 main.tex

and they do have a glyphs for arrows ↓ (U+2193) & → (U+2192) which I double checked by opening all the *.ttf files with Fontforge:

xelatex is an UTF-8 engine and should support UTF-8 encoded multibyte characters... So why don't they render properly?

Thank you! This solved half of my problem. Now arrows are aligned properly but are still not the same color as the rest of the comments. Any Idea on how to solve this as well? — 71GA, May 21 '20 at 20:07

hair-splitter · Answer 1 · 2020-05-21T21:18:06.000

1

I use this method: The 'listings' package and UTF-8

An MWE:

\documentclass{article}
\usepackage{fontspec}
\usepackage{xcolor}
\usepackage{listings}
\makeatletter
\lst@InputCatcodes
\def\lst@DefEC{%
 \lst@CCECUse 
 \lst@ProcessLetter
  ^^^^2192^^^^2193^^00}
\lst@RestoreCatcodes
\makeatother
\lstset{
   basicstyle=\ttfamily,
   keepspaces,
   commentstyle=\color{gray},
   extendedchars
}
\begin{document}
    \lstinputlisting[language=C]{main.c}
\end{document}

edited May 21 '20 at 21:18

answered May 21 '20 at 21:10

hair-splitter

2,741

Is there a fix to make this work for all utf8 characters? I don't want to write this all the time... – 71GA May 21 '20 at 21:17
No. Only U+2193 and U+2192 work in this code. – hair-splitter May 21 '20 at 21:20
1

See the link in the answer. – hair-splitter May 21 '20 at 21:21
I checked the link and tried the last answer there which is to use texcl=true inside \lstset{}. This colored my comments properly, but removed any kind of indentation at the beginning of the line. So if I use this all the comments are mooved completely to the left. – 71GA May 21 '20 at 21:34
The first answer is the relevant. – hair-splitter May 22 '20 at 06:17
Yes. But it is painful to define ^^^^2192... for each UTF-8 character... Don't you think? – 71GA May 22 '20 at 08:07

Not all UTF8 symbols work in lstlistings

1 Answers1