0

I recently upgraded from texlive-2019 to texlive-2020, and my index command broke.

The relevant code is:

\newcommand{\BasisChange}[2]{[#2\!\leftarrow\!#1]}

A change of basis matrix\index{$\BasisChange{\mathcal A}{\mathcal B}$}.

In texlive-2019, this would produce an entry in my .ind file of

  \item {$[\mathcal  B\tmspace  -\thinmuskip {.1667em}\leftarrow \tmspace  -\thinmuskip {.1667em}\mathcal  A]$}, 
        161

whereas in texlive-2020, it produces

  \item $[\mathcal  B\
    \subitem \leftarrow \
      \subsubitem \mathcal  A]$, 161

It seems that the \! is being interpreted as the start of a subitem.

Is there a way I can get the old behaviour back or "protect" my math command so \! isn't misinterpreted?

1
  • 1
    Clearly it shouldn't behave that way, but my comment is about something else. When using a math expression as an index term, you really should consider providing a sort field. The index sort is really dumb -- with makeindex, plain ascii -- so you want some control over where this entry gets placed. – barbara beeton Jan 17 at 3:22
2

What's happening here is that the macro gets expanded when its contents are written to the .idx file you get an entry of

\indexentry{$\mathcal B\!\leftarrow\!\mathcal A$}

(There are other issues with how you've defined things. Try doing \BasisChange{A}{\mathcal B} for example.)

The exclamation marks get interpretted as being subitems. I think that just doing the basics of changing your index command to

\index{B-A@\BasisChange{A}{B}}

will fix the problem as makeindex shouldn't be parsing for subentries past the @ if I remember correctly.

Edit: Digging a bit deeper this won't work. You could try

\index{B-A@{A}{B}|BasisChange}

but I've not checked that. End edit

Putting \protect before \BasisChange should also work and if you have a lot of these in your document, you could use \DeclareRobustCommand (actually, probably \DeclareRobustCommand* would be even better) in place of \newcommand so you get the \protection automatically.

As for why this is broke with the update, I did a cursory search through the git history and it looks like it might be connected with changes regarding robustness that were part of the 2019 patches to LaTeX, but I didn't see anything obvious.

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.