Re: [O] Exporting source code blocks as LaTeX figures

[James Harkins]

On Mon, May 20, 2013 at 2:30 PM, Thomas S. Dye <address@hidden> wrote:
>> Then the next problem... I'm going to have some figures that need to
>> span two columns. According to [1], I should use \begin{figure*} ...
>> \end{figure*}. Is there a way to do that using a special block? (First
>> guess, "#+BEGIN_figure*" causes the special block not to be
>> recognized.) Or do I need to write the literal LaTeX code for that?
>
> No, you don't need to write literal LaTeX code. See Section 12.7.4 of
> the manual, which describes setting the :float attribute to
> 'multicolumn'.

That section of the manual seems to describe the behavior for images. I did try:

#+ATTR_LaTeX: :multicolumn :options [htb]
#+BEGIN_figure
...
#+END_figure

... and :multicolumn had no effect on the output LaTeX code (either
with or without the preceding ":" -- I tried both):

\begin{figure}[htb]
\lstset{language={},label=code1,caption={Some code},numbers=none}
\begin{lstlisting}
// SuperCollider code here
.....
\end{lstlisting}
\end{figure}

Nor does this make a difference:

#+ATTR_LaTeX: :float multicolumn :options [htb]
#+BEGIN_figure

And if I omit BEGIN_figure altogether, and try to apply :float
directly to the source code block, then the floating environment
doesn't appear in the LaTeX code at all.

The other reference to multicolumn is for table export, and this isn't
a table either. So I think, as currently designed, :multicolumn simply
doesn't apply.

I'm aware that BEGIN_figure delimits a "special block" which
translates into LaTeX as an arbitrary environment, and not all
environments can be floated safely... but perhaps one approach for the
future would be to consider BEGIN_figure to be an extra-special
"special block" where we know that the properties of LaTeX floating
bodies are valid.

Or relax the restriction so that BEGIN_figure* no longer tries to
treat _ as a subscript marker.

hjh