Re: xelatex, revtex4-2 vs article, CID mapping destroyed after gs-based concatenation

    * between the original (pre-concatenation) PDF (left = revtex4-2,
    right=article) the text representation/generation has changed:
    ...
    Note that both pdfs contain the same fonts in the same encodings

Even with the (returning) gs bug, this difference seems like it means
there is something different happening with revtex, since gs is not
involved pre-concatenation. This seems bizarre.
    
    Has anyone an idea where to search for what `revtex4-2.cls` is doing
    wrongly?

I searched in revtex4-2.dtx (texmf-dist/source/latex/revtex/revtex4-2.dtx)
for things like pdf, cmap, etc., but saw nothing relevant :(.
Possibly it's doing something unusual with the fonts (too many hits to
look at), but it didn't seem likely.

Afaik, the pdftex primitive that's used for this is \pdffontattr (not
used in revtex); there's also various \pdf...unicode primitives, but
they're not used in revtex either, not surprisingly.

Maybe a diff of the \loggingall output could show something relevant?
Apart from the irrelevancies it will definitely show.

The only other approach I can think of is to run pdftex under gdb up to
the point where the output starts to differ and see why it's doing
different things. Not fun :(. -k