Let's compile the following code with {xe|lua}atex:

\documentclass{standalone}% or article, doesn't matter
\setmathfont{TeX Gyre Termes Math}%% or TeX Gyre Pagella Math, doesn't matter

Upon viewing the result of compilation


(as PDF) in evince, we select the single letter, copy it and paste it into some UTF8-capable text editor such as gedit or emacs. What we get is N, not ℕ. But http://www.gust.org.pl/projects/e-foundry/tg-math/gust_e-foundry-math_fonts-presentation.pdf does mention ℕ as a letterlike symbol, so, I'd expect to get ℕ! Why does it happen? How to get ℕ in the PDF text layer?

EDIT: Reported to evince as https://gitlab.gnome.org/GNOME/evince/issues/1085 . Also cf. https://gitlab.freedesktop.org/poppler/poppler/issues/724 . Let's keep the question open until we are 100% sure that the PDF file produced complies with a PDF specification.

  • 1
    When I run pdftotext on the resulting PDF to extract the actual textual representation I get . – Henri Menke Feb 19 at 0:52
  • 2
    @user0 Yes, it's a viewer issue. I just verified it using Evince, Okular, and Foxit. Evince and Okular only copy N, whereas Foxit correctly copies . – Henri Menke Feb 19 at 0:55
  • 6
    Voting to close my own question: it turned out to be not TeX-related. – user49915 Feb 19 at 0:57
  • 1
    @user0 I also had a look at the uncompressed PDF and found the encoded glyph U+2115 which is DOUBLE-STRUCK CAPITAL N. – Henri Menke Feb 19 at 0:58
  • 1
    @HenriMenke Thx! Moreover, I found no N in the uncompressed steam that could possibly confuse evince. – user49915 Feb 19 at 1:03

Your Answer

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

Browse other questions tagged or ask your own question.