Sublabels not shown within \psfrag{}{}

[lcnittl]

I came across a problem using \cmpd+{} to substitute sublabels inside a \psfrag command. Interestingly, this is no problem using \replacecmpd{main.sub}, however, using \cmpd+{main.sub} outputs 2 questionmarks as the sublabel.

The whole thing works fine for mainlabels.

\documentclass{article}

\usepackage{graphicx}
\usepackage{auto-pst-pdf}
\usepackage{chemnum}

\begin{document}
  Using chiral alcohol \cmpd+{cmpd1.1} to get chiral ether \cmpd+{cmpd1.2}.
  
  \begin{figure}
    \centering
    \cmpd*{cmpd1.1}
    \psfrag{TMP1}[Bc][Bc]{Chiral molecule \cmpd+{cmpd1.1}}
    \replacecmpd[tag=TMP2]{cmpd1.2}
    \includegraphics[scale=1]{scheme-tmp.ps}
    \caption{Working here (\cmpd+{cmpd1.1}), however, not in psfrag. The second one (\cmpd+{cmpd1.2}) is working in both.}
  \end{figure}
\end{document}

Would you have an idea why this is happening, or how to solve this?

[lcnittl]

Seems that a missing runs=2 option for auto-pst-pdf packages caused this behaviour.

[cgnieder]

There is a subtle difference between both use cases:

\replacecmpd{cmpd1.1} is basically the same as \cmpd*{cmpd1.1}\psfrag{TMP1}{\cmpd{cmpd1.1}}. You are using \cmpd*{cmpd1.1}\psfrag{TMP1}{\cmpd+{cmpd1.1}}.

It seems this slight difference means that two auxiliary compilations are needed instead of one.

BTW: with \cmpd*{cmpd1.1}\replacecmpd+{cmpd1.1} the sublabel wouldn't show, either.

This is a bug, but I haven't found the cause, yet. And since you found an easy fix for the time being I'll put this on hold.

[markellos918]

I have a similar problem. I've been trying to tweak xr to recognise \cmpd labels (see here) – by the way, this could be a good feature to implement in the future. All is fine until I try to use \submaincmpdplain or \cmpd+, for example in section titles. With the first command it doesn't print the sub label at all, with the second it prints ??. This only happens with labels imported from a different document, which I think is equivalent to predefine them in the preamble. Can you please look into this?