# How can I line-break a long string of numeric digits and hyphens?

I have a cls file that contains a definition for source code:

\lstnewenvironment{code}[1][]
{
\lstset{
basicstyle=\ttfamily\footnotesize,
breaklines=true,
frame=lines,
extendedchars=true,
captionpos=b,
caption=#1
}
}
{
}


I have a problem with code listings and line breaking but only on lines containing strings longer than the line width that consist of digits and hyphens only. Other long strings break fine.

The specific example of a long string that highlighted this problem to me was trying to display a "SID" value in a narrow column. An example of one of these is shown below:

# wbinfo --name-to-sid myuser
S-1-5-21-4099219672-1275272411-291422405-1104 SID_USER (1)


The string "1-5-21-4099219672-1275272411-291422405-1104" is too wide to fit in a narrow column but it is not broken up. As a further (extremely long line) example:

12345-word-67890-word-09876-word-54321-word-12345-word-67890-word-09876-word-54321-word-12345-word-67890-word-09876-word-54321-word-12345-word-67890-word-09876-word-54321-word-12345-word-67890-word-12345-67890-09876-54321-12345-67890-09876-54321-12345-67890-09876-54321-word-54321-word-12345-word-67890-word-09876-word-54321-word-12345-word-67890-word-09876-word-54321-word-12345-word-67890-word-09876-word-54321


gets split where necessary between the letter d of word and the hyphen that follows it. The subsrring consisting of only numbers and hypens is not split regardless of its length.

So, to summarise, line breaking works BUT ONLY when the line contains non-numeric characters.

What I want is for it to break on hyphens (as it does) but to do it always and not only when a hyphen is preceded by a letter.

• listings provides prebreak and postbreak keys (default is empty {}). They might be of better use than fiddling with \discretionary and literate. – Werner Dec 7 '13 at 20:39
• On looking into this further I find that line breaking works fine without the literate (in fact, it is better because the lines don't get quashed vertically). What I have realized is that the problem only occurs in lines consisting of only numbers and dashes. I shall update my question with some examples. – starfry Dec 9 '13 at 12:43
• For anyone reading this and wondering about the mentions of literate, I originally thought that was the cause of my problems but it has nothing to do with it. I have updated the question text to reflect this and help avoid misguided answers. – starfry Dec 10 '13 at 20:42

You can break this using literate turning the - into a discretionary:

\lstset{literate={-}{{-\allowbreak}}{1} }


or allowing breaks at -, 0 and 1

\lstset{literate={-}{{-\allowbreak}}{1}
{0}{{0\allowbreak}}{1}
{1}{{1\allowbreak}}{1} }


You can allow breaks for other numbers if you so wish by extending the pattern. The format is

{character sequence}{{replacement}}{length}


Note the extra brackets around replacement. Multiple such sequences can be separated by blank space or a newline for readability.

Code for first example

\documentclass[twocolumn]{article}

\usepackage{listings}

\lstnewenvironment{code}[1][]
{
\lstset{
basicstyle=\ttfamily\footnotesize,
breaklines=true,
frame=lines,
extendedchars=true,
captionpos=b,
caption=#1,
literate={-}{{-\allowbreak}}{1}
}
}
{
}

\begin{document}

\begin{code}
# wbinfo --name-to-sid myuser
S-1-5-21-4099219672-1275272411-291422405-1104 SID_USER (1)
\end{code}

\end{document}


Code for second example

\documentclass[twocolumn]{article}

\usepackage{listings}

\lstnewenvironment{code}[1][]
{
\lstset{
basicstyle=\ttfamily\footnotesize,
breaklines=true,
frame=lines,
extendedchars=true,
captionpos=b,
caption=#1,
literate={-}{{-\allowbreak}}{1}
{0}{{0\allowbreak}}{1}
{1}{{1\allowbreak}}{1}
}
}
{
}

\begin{document}

\begin{code}
# wbinfo --name-to-sid myuser
S-1-5-21-4099219672-1275272411-291422405-1104 SID_USER (1)
\end{code}

\end{document}