And, in case it is useful, the MathML:
<math xmlns="http://www.w3.org/1998/Math/MathML" display="block">
<mrow>
    <munder class="msub">
      <mrow>
    <munder accentunder="false">
    <mrow>
        <mi>x</mi>
        <mi>x</mi>
        <mi>x</mi>
        <mo class="MathClass-op">…</mo>
        <mi>x</mi>
        <mi>x</mi>
    </mrow>
    <mo>︸</mo>
      </munder>
      </mrow>
      <mrow>
    <mi>n</mi>
    </mrow>
    </munder>
</mrow>
</math>

This is rendered correctly by MathJax and by Firefox native rendering. If I transform that using pandoc to LaTeX, for instance, I get:
\[\underset{n}{\underset{\}\ }{xxx\ldots xx}}\]
This renders in the same unexpected manner as the Word rendering when going the other way.

Thanks,
Emma

On Tuesday, 3 October 2017 10:36:44 UTC+1, Emma Cliffe wrote:

Morning,

I am having trouble getting pandoc to convert underbrace either into Word or out of Word. I wondered if there is either something I am doing wrong, whether this is a known problem (I can't find a report) or if there is a work round for converting into Word.

If I try to convert:
\[
\underbrace{xxx\ldots xx}_n
\]
for instance then what I get in Word (linear format) is:
((xxx…xx)┬⏟)┬n
which does not render as a stretched underbrace.

I would expect to get:
⏟(xxx…xx)┬n
which is typed directly into Word as \underbrace(xxx\ldots xx)_n (then space to trigger). You get the same linear interpretation if you use the mouse to create the structure.

The above is the direction I am interested in. I am actually converting from MathML and see the same behaviour as in the cut down test of md or tex to docx I have explained above.

For completeness I tried to convert in the other direction. Starting from the above structure in Word I end up with:
\[\underset{n}{}\]
which has not only lost the xxx...xx but underbrace has become underset which is not the same thing.

Does anyone know anything more about this?

Many thanks,
Emma