3

I have an issue when copying and pasting numbers from a PDF generated with pdflatex (pdfTeX 3.14159265-2.6-1.40.19 (TeX Live 2018/W32TeX)) and Acrobat Reader/Acrobat Professional. Also reported at https://github.com/latextemplates/LNCS/issues/22.

MWE:

\documentclass{article}
\usepackage{cfr-lm}

\setlength\parindent{0pt}
\obeylines

\begin{document}
Normal number: 100.
Math mode number: $100$.
\end{document}

Acrobat Professional copy result:

Normal number: 322.
Math mode number: 100.

With oldstyle=false:

\documentclass{article}
\usepackage[rm={oldstyle=false}]{cfr-lm}

\setlength\parindent{0pt}
\obeylines

\begin{document}
Normal number: 100.
Math mode number: $100$.
\end{document}

Result:

Normal number: 􀀀􀀀.
Math mode number: 100.

paste result

With lmodern, everything works as expected:

\documentclass{article}
\usepackage{lmodern}

\setlength\parindent{0pt}
\obeylines

\begin{document}
Normal number: 100.
Math mode number: $100$.
\end{document}

Result:

Normal number: 100.
Math mode number: 100.

Notes

Note that this issue does not appear with SumatraPDF.

This issue also happens with pdf.js in Firefox. Also, using lualatex (LuaTeX, Version 1.07.0 (TeX Live 2018/W32TeX)) does not help.

It also happens with the documentation of cfr-lm. Go to http://texdoc.net/pkg/cfr-lm (with Firefox), copy "Version 1.5 - 2015-02-01". The pasted result is: "Version 3.7 — 4237–24–23"

  • 1
    Why do you insist on using pdftex when we have luatex/xetex and fontspec? I'm sure whatever this package does can also be done with fontspec. – Martin Schröder May 8 '18 at 9:48
  • 1
    I can't reproduce this in Okular either. I wouldn't recommend using this package with LuaTeX: in that case, use fontspec. There's no reason to use cfr-lm in that case. – cfr May 22 '18 at 0:38
6

You can use glyphtounicode:

\documentclass{article}
\usepackage[T1]{fontenc}
\usepackage{cfr-lm}
\input glyphtounicode
\pdfgentounicode=1
\setlength\parindent{0pt}
\obeylines
\pagestyle{empty}
\begin{document}
1234567890 copied: 1234567890

\end{document}
| improve this answer | |
  • This is a case, where cmap does not help. Thank you for pointing it out. – koppor May 8 '18 at 10:52
  • You don't need to load fontenc here. Do you think cfr-lm should load glyphtounicode? (Conditionally, if pdfTeX is used?) – cfr May 22 '18 at 0:40

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.