Re: [O] BUG: example/src blocks and latex export

[Eric Schulte]

Nick Dokos <address@hidden> writes:

> Consider the following org file:
>
>
> * foo
>
> Verbiage to begin the paragraph
> #+begin_src shell
> get-config.py var section [section ...]
> #+end_src
> and verbiage to end the same paragraph.
>
> * bar
>
> Verbiage to begin the paragraph
> #+begin_example
> get-config.py var section [section ...]
> #+end_example
> and verbiage to end the same paragraph.
>
>
> When exported to latex with current git (Org-mode version 7.7 
> (release_7.7.120.g2edd.dirty)),
> I get:
>
> Verbiage to begin the paragraph
>
> \begin{verbatim}
> get-config.py var section [section ...]
> \end{verbatim}
>
>
>
> and verbiage to end the same paragraph.
> \section{bar}
> \label{sec-2}
>
>
> Verbiage to begin the paragraph
>
> \begin{verbatim}
> get-config.py var section [section ...]
> \end{verbatim}
>
>
> and verbiage to end the same paragraph.
>
> so both instances of "verbiage to end the same paragraph" actually end
> up being in a different paragraph, with three empty lines after a
> source block and two empty lines after an example block, where none
> existed before.  LaTeX indents the newly created paragraph and it
> looks ugly. Of course, just a single empty line is enough to do
> the damage, but the fact that there is more than one and that there
> are different numbers, indicates multiple places where a gratuitous
> newline is inserted.
>

If you want results embedded into a paragraph, then I would suggest
using an inline code block.  Regular code blocks are "blocks" in that
they will break paragraphs and are treated (in my mind) more like
figures than inline elements.

>
> I get sane behavior with the attached patch, but I'm wondering if it
> breaks other backends, so if somebody is willing to test, I'd appreciate
> it (and of course, I'll test as well). For the time being at least, this
> is a trial balloon, not a real patch.
>

Thanks for the patch, I would not want this behavior on by default
because I think the results in the Org-mode file are less visually
pleasing than when results are padded with spaces.  Perhaps this
behavior could be controlled through a new header argument which would
default to the current behavior.

> ISTR this issue coming up on the list recently: did I imagine it?

A similar issue regarding the insertion of spaces through repeated
evaluation and removal of results of a code block has been discussed
recently.  I'm pushing up a small fix for that issue now.

Cheers -- Eric

-- 
Eric Schulte
http://cs.unm.edu/~eschulte/