How do I use certain characters as preferred for line breaks with xurl?

Question

xurl is a great solution.

Here is a quote from the pramble:

Package xurl loads package url by default and defines possible url breaks for all alphanumerical characters and = / . : * - ~ ' "

How do I add underscore = / . : * - ~ ' " and _ as preferred characters to break the following url

https://www.weibull.com/pubs/2011_RAMS_planning_a_reliability_growth_program_utilizing_historical_data.pdf

It should only break words as a last resort.

Minimal code showing the issue:

\documentclass{minimal}
\usepackage{xurl}
\begin{document}
Here is the url \url{https://www.weibull.com/pubs/2011_RAMS_planning_a_reliability_growth_program_utilizing_historical_data.pdf}.
\end{document}

Answer 1

In theory it is possible to allow line breaks after A-Z, a-z, 0-9 in \url in addition to other possible break points, but with a higher penalty (which means LaTeX will prefer other hyphenation points).

I have my doubts that this will have an effect that is significantly different from just loading xurl, which allows breaks everywhere with the same penalty, though. If a URL takes up a significant portion of the line, chances are there is little space left to shrink and enlarge, which means that TeX has little leeway in deciding where exactly to break the line: Either the URL can be broken pretty much where it meets the margin or not.

Anyway, here is how biblatex does it

\documentclass{article}
\usepackage{url}
\usepackage{lipsum}
\mathchardef\UrlNotSoGreatBreakPenalty=800
\usepackage{etoolbox}
\begingroup
\def\do#1{%
  \gappto\UrlSpecials{%
    \do#1{%
      \mathchar`#1
      \penalty\UrlNotSoGreatBreakPenalty}}}
\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
\do\1\do\2\do\3\do\4\do\5\do\6\do\7\do\8\do\9\do\0
\endgroup
\begin{document}
\lipsum[1]
Text \url{https://www.weibull.com/pubs/2011_RAMS_planning_a_reliability_growth_program_utilizing_historical_data.pdf}.
\lipsum[2]
\end{document}

This assigns a penalty of 800 to not so great breaks after A-Z, a-z, 0-9. Breaks after punctuation has penalty 700, big breaks after : a penalty of 500.

---

You could allow some stretchable space as suggested in Ulrike Fischer's answer to 'Uneven' breaks in long URLs (\url) (which has already been shamelessly stolen by biblatex: https://github.com/plk/biblatex/issues/850, https://github.com/plk/biblatex/pull/886). This may make it possible to allow your URLs to break in nicer places at the cost of some more whitespace between characters.

\documentclass{article}
\usepackage{url}
\usepackage{lipsum}
\mathchardef\UrlNotSoGreatBreakPenalty=800
\newmuskip\urlalnumskip
\setlength{\urlalnumskip}{0mu plus 2mu}
\usepackage{etoolbox}
\begingroup
\def\do#1{%
  \gappto\UrlSpecials{%
    \do#1{%
      \mathchar`#1
      \mskip\urlalnumskip
      \penalty\UrlNotSoGreatBreakPenalty}}}
\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
\do\1\do\2\do\3\do\4\do\5\do\6\do\7\do\8\do\9\do\0
\endgroup
\begin{document}
\lipsum[1]
Text \url{https://www.weibull.com/pubs/2011_RAMS_planning_a_reliability_growth_program_utilizing_historical_data.pdf}.
\lipsum[2]
\end{document}

Note how the characters in the first line are spaced out quite a bit to make the nicer break possible.

If you play around with the penalty and the stretchable space you will find that the penalty only has a real effect if the stretchable space can give it enough leeway.

Answer 2

This creates a \url command that allows breaks at specified characters, here given as /, ., and _.

\documentclass{article}
\usepackage[pass,showframe]{geometry}
\catcode`_=12 
\newcommand{\url}[1]{%
  \begingroup
  \ttfamily
  \begingroup\lccode`~=`/\lowercase{\endgroup\def~}{/\penalty0 }%
  \begingroup\lccode`~=`.\lowercase{\endgroup\def~}{.\penalty0 }%
  \begingroup\lccode`~=`_\lowercase{\endgroup\def~}{_\penalty0 }%
  \catcode`/=\active\catcode`.=\active\catcode`_=\active
  \scantokens{#1\noexpand}%
  \endgroup
}
\catcode`_=8 
\begin{document}
Here is the url \url{https://www.weibull.com/pubs/2011_RAMS_planning_a_reliability_growth_program_utilizing_historical_data.pdf}.
\end{document}