bold and italic markup is sometimes lost in index terms

[siegfriedzoetzsche]

When converting

\documentclass{article}
\begin{document}

\textbf{BOLD\index{NORMAL}}

\textbf{BOLD\index{\textbf{BOLD}}}

\textbf{BOLD\index{\textit{ITALIC}}}

\textit{ITALIC\index{NORMAL}}

\textit{ITALIC\index{\textbf{BOLD}}}

\textit{ITALIC\index{\textit{ITALIC}}}

{NORMAL\index{NORMAL}}

{NORMAL\index{\textbf{BOLD}}}

{NORMAL\index{\textit{ITALIC}}}

\end{document}

with LaTeXML version 0.8.2; revision 48de2c4

latexml --nocomments test.tex

I get

[...]
  <para xml:id="p1">
    <p><text font="bold">BOLD<indexmark>
          <indexphrase key="NORMAL">NORMAL</indexphrase>
        </indexmark></text></p>
  </para>
  <para xml:id="p2">
    <p><text font="bold">BOLD<indexmark>
          <indexphrase key="BOLD">BOLD</indexphrase>
        </indexmark></text></p>
  </para>
  <para xml:id="p3">
    <p><text font="bold">BOLD<indexmark>
          <indexphrase key="ITALIC"><text font="italic">ITALIC</text></indexphrase>
        </indexmark></text></p>
  </para>
  <para xml:id="p4">
    <p><text font="italic">ITALIC<indexmark>
          <indexphrase key="NORMAL">NORMAL</indexphrase>
        </indexmark></text></p>
  </para>
  <para xml:id="p5">
    <p><text font="italic">ITALIC<indexmark>
          <indexphrase key="BOLD"><text font="bold">BOLD</text></indexphrase>
        </indexmark></text></p>
  </para>
  <para xml:id="p6">
    <p><text font="italic">ITALIC<indexmark>
          <indexphrase key="ITALIC">ITALIC</indexphrase>
        </indexmark></text></p>
  </para>
  <para xml:id="p7">
    <p>NORMAL
<indexmark>
        <indexphrase key="NORMAL">NORMAL</indexphrase>
      </indexmark></p>
  </para>
  <para xml:id="p8">
    <p>NORMAL
<indexmark>
        <indexphrase key="BOLD"><text font="bold">BOLD</text></indexphrase>
      </indexmark></p>
  </para>
  <para xml:id="p9">
    <p>NORMAL
<indexmark>
        <indexphrase key="ITALIC"><text font="italic">ITALIC</text></indexphrase>
      </indexmark></p>
  </para>
[...]

In <para xml:id="p2"> we have

<indexphrase key="BOLD">BOLD</indexphrase>

while in <para xml:id="p5"> and <para xml:id="p8"> we have

<indexphrase key="BOLD"><text font="bold">BOLD</text></indexphrase>

Also how would one distinguish <para xml:id="p1"> from <para xml:id="p2">?

I am not sure what the expected behaviour would be. Depending on whether text[@font='bold']/indexmark/indexphrase/text() is supposed to inherit the boldness from the text ancestor I would expect

<indexphrase key="NORMAL"><text font="medium">NORMAL</text></indexphrase>

in <para xml:id="p1"> or

<indexphrase key="BOLD"><text font="bold">BOLD</text></indexphrase>

in <para xml:id="p2">.

italic markup shows an analog problem. I have not tested with \emph{}

[brucemiller]

Normally, you want that inheritance (which was hard enough to get right in LaTeXML! :> ). But in the case of metadata like \index which gets pulled out of its context and used elsewhere, it both needs to restart from a default font, and record it's own font standalone. I couldn't quite solve this without creating (yet) another magic attribute _standalone_font which marks elements which need this treatment. I think this should work now, w/o (seemingly) cluttering up the common cases. Thanks for the report!