On Wed, 5 Aug 2026 at 08:50, Patrice Dumas <[email protected]> wrote:
> On Sat, Jul 25, 2026 at 08:46:56PM +0100, David Carlisle wrote:
> > the pdftitle and alt text in includegraphics in the end have to be a PDF
> > encoded string with no markup, so
> > over the years latex has got better at "cleaning up text" basically now
> any
> > unicode text should work, the only characters that you need to quote are
> %
> > { }
>
> Seems like _ needs to be quoted too. Maybe general text protection
> needs to be done, so also protecting $#, or even ~.
>
_ does not need to be quoted as you are making a string not typesetting
\DocumentMetadata{tagging=on}
\documentclass{article}
\usepackage{graphicx}
\begin{document}
\includegraphics[alt={some_underscore_text worth 50$}]{example-image}
\end{document}
runs without error giving the expected alt text on the figure
<Figure xmlns="http://iso.org/pdf2/ssn";
id="ID.08"
alt="some_underscore_text worth 50$"
xmlns:Layout="http://iso.org/pdf/ssn/Layout";
Layout:BBox="{ 148.71233, 427.19861, 468.71155, 667.19801 }"
>
although if the string may be used in multiple contexts so might be
typeset as well as being used as a pdf string
and you want to be safe it's OK to quote them
\includegraphics[alt={some\_underscore\_text worth 50\$}]{example-image}
produces the same output here.
> --
> Pat
>
