With beamer, "Missing { inserted." (regression)

[Zdenek Wagner]

Well, you should read carefully The TeXbook and understand how TeX
reads the input, what is a token, when and how different types of
expansion occur (in the mouth or in the stomach) and how the math
lists are built. In addition, you must know the definitions of control
sequences used in the fileds of math atoms. If you do not know them
well, it is better to be at the safer side.

Zdeněk Wagner
http://ttsm.icpf.cas.cz/team/wagner.shtml

pá 18. 2. 2022 v 4:27 odesílatel Vincent Lefevre <vincent at vinc17.net> napsal:
>
> On 2022-02-17 19:17:38 +0100, Enrico Gregorio wrote:
> > The syntax $e_\textnormal{m}$ has never been supported and
> > if it works is just by accident. The proper syntax is to
> > brace _any_ subscript (or superscript), but there’s usually
> > no problem when there is just a _single_ token.
>
> In practice, people just see whether this works, without necessarily
> thinking of other possibilities. Then the real issue is the change of
> behavior by new versions.
>
> I'm wondering... I can see, for instance:
>
> texmf-dist/tex/latex/dvgloss/dvgloss.sty:\makeglshortcut[{[$_\textrm{\footnotesize#1}$\thinspace}
> texmf-dist/tex/latex/nwejm/nwejmart.cls:    $_\mathsf{\arabic*}$
> texmf-dist/tex/latex/unifith/unifith.cls:  \providecommand*{\rb}[1]{\ensuremath{_\mathrm{#1}}}
>
> Isn't that similar?
>
> > However, try $A_\notin$ and you’ll learn that $A_{\notin}$
> > is much better.
>
> This is a bit unfortunate. But I suppose that if this has never
> worked, people would have added braces after seeing a failure.
> This is not a big issue compared to incompatibilities introduced
> in new TeX Live versions, potentially breaking many files.
>
> --
> Vincent Lefèvre <vincent at vinc17.net> - Web: <https://www.vinc17.net/>
> 100% accessible validated (X)HTML - Blog: <https://www.vinc17.net/blog/>
> Work: CR INRIA - computer arithmetic / AriC project (LIP, ENS-Lyon)