How to rigidly chop long lines in plain TeX

Question

I am using plain TeX and want to typeset an appendix that contains several hundred URLs. When a URL goes past the right margin, I would like TeX to stop printing (i.e. not carry into the margin, not print a warning black rectangle) and then pick up the rest of the URL on a new (non-indented) line where it left off.

Something like this, where 1, 2, 3 are akin to footnote numbers:

1: http://www.aaa.com/rem
ainder 2: http://www.bbb.
com/remainder_again  3: h
ttp://www.ccc.com/ending

Is this possible?

I am using a Python script to generate the .tex files, so, if worst comes to worst, I can have Python estimate the content of each line. Unfortunately, I am not using a mono-spaced font, so that could turn pretty ugly.

Thanks in advance.

Answer 1

Since you're generating the TeX code with Python, one option is to separate every character with \hskip0pt. For example:

\def\a{a\hskip0pt}
\def\b{\a\a\a\a\a\a\a\a\a\a}
\def\c{\b\b\b\b\b\b\b\b\b\b}
\def\d{\c\c\c\c\c\c\c\c\c\c}

\d

Answer 2

The LaTeX url package can be used with plain TeX using miniltx (or you could copy the minimal code, of course). For example

\input miniltx %
\input url.sty %
\urlstyle{rm}
\def\UrlBreaks{\do\/\do\a\do\b\do\c\do\d\do\e\do\f\do\g\do\h\do\i\do\j\do\k\do\l\do\m\do\n\do\o\do\p\do\q\do\r\do\s\do\t\do\u\do\v\do\w\do\x\do\y\do\z\do\A\do\B\do\C\do\D\do\E\do\F\do\G\do\H\do\I\do\J\do\K\do\L\do\M\do\N\do\O\do\P\do\Q\do\R\do\S\do\T\do\U\do\V\do\W\do\X\do\Y\do\Z}
\hsize 4.4cm %
\noindent
1.~\url{http://www.aaa.com/remainder}
2.~\url{http://www.bbb.com/remainder_again}
3.~\url{http://www.ccc.com/ending}
\bye

will give you breaking at any character using the method suggested in the url manual and Forcing linebreaks in \url. Things are dependent on line-width: with very few 'normal' spaces in the example above avoiding overfull boxes is hard, and the equivalent of LaTeX's \sloppy may also be required, e.g.

\tolerance 9999 %
\emergencystretch 3em %

I'm assuming we have to allow breaks everywhere: that is normally not such a great plan if it can be avoided.

To understand what is happening, notice that url (ab)uses math mode to allow breaking 'anywhere', making any breakable URL character into a mathbin character inside the URL.

Answer 3

The chapter 14 "How TEX Breaks Paragraphs into Lines" in TeXbook explains the hyphenation is just TeX inserts \discretionary{\char\hyphenchar\font}{}{} at every allowed word break when \hyphenchar is a number from 0-255, so set it to -1 disables hyphenation, and now we want to "manually" insert \discretionary{}{}{} between every characters. This seems scary, but it is easy with plain TeX macros, not even the token register is need.

\def\ub@{\discretionary{}{}{}}
\def\appe@#1#2\eol{\def\g{#2}%
\ifx\g\empty #1\else\ub@#1\appe@#2\eol\fi}
\def\protol@#1://#2\eol{#1:\kern-1.5pt/\kern-2.5pt/\appe@#2\eol}
\def\url#1{\begingroup\count0=\the\hyphenchar\font%
\hyphenchar\font=-1%
\protol@#1\eol
\hyphenchar\font=\count0\endgroup}

The work horse is the \appe@ macro, which takes the first character in the string as #1 and the rest of the string until token \eol as #2, and loops to insert \ub@ between them until #2 becomes empty. Note the use of % to prevent extra whitespace occurs in the result. The \protol@ macro is just something I use to prevent breaking at the protocol part of the url (https:// for example), and make the :// a bit more compact. And \begingroup does not reset \hyphenchar so remember to set it back.

[1] American Cornhole League. ``American Cornhole League Rules and
Regulations,'' November 13, 2019. Available:
\url{https://mysqlvm.blob.core.windows.net/acl-docs/ACL-Rules-Regs-2019-2020.pdf}.
[Accessed: April 17, 2024]\par
[2] American Cornhole Organization. ``Cornhole Rules,'' [online document].
Available:
\url{https://americancornhole.com/downloads/ACO-Season16/ACO-Cornhole-Rules-Sheet1.pdf}.
[Accessed: April 17, 2024]\par

If the long url spans the whole line, it is a good idea to add extra glue around / so TeX can stretch the line a little bit to make it fit. Note that the glue itself would permit TeX to line break at it.

\def\appe@#1#2\eol{\def\g{#2}%
\ifx\g\empty #1\else\ifx#1/\hskip-.5pt plus 1pt/\hskip-1pt plus 1pt%
\else\discretionary{}{#1}{#1}\fi\appe@#2\eol\fi}

Answer 4

We can create a macro which inserts \hskip0pt plus1pt after each printed character. This space is breakable and slightly stretchable, so the paragraph justification is kept at both margins and the long url's are breakable at arbitrary place.

\newcount\urlnum
\def\url#1{\global\advance\urlnum by1 $^{\the\urlnum}$%
   \bgroup\tt \urlA#1\end}
\def\urlA#1{\ifx#1\end\egroup\else
   \string#1\hskip0pt plus1pt\relax\expandafter\urlA\fi}

\hsize=7cm
\noindent Test: 
\url{http://www.aaa.com/remainder}
\url{http://www.bbb.com/remainder_again}
\url{http://www.ccc.com/ending}

\bye