new odt-exporter: problems with formatting in captions

[martin.guertler@gmx.de (Martin Gürtler)]

Hi,

I just converted a 150 page document I used to convert to odt with the old
exporter. Following problems occurred:

1. table with #+CAPTION
   If the caption contains formatting (~, ==, _, ^), the exporter
   produces an invalid content.xml.  

   minimal example:

---8<------------------------------------------------------------
#+OPTIONS:   num:t toc:nil \n:nil @:t ::t |:t ^:{} -:t f:t *:t <:nil

* First paragraph
#+CAPTION: A_{1}^{2}
| head 1    | head2    |
|-----------+----------|
| content 1 | content2 |
---8<------------------------------------------------------------

The problem can be seen in the following xml fragment:

---8<------------------------------------------------------------
<table:table table:style-name="OrgTable" 
 table:name="A&lt;text:span 
 text:style-name="OrgSubscript"&gt;1&lt;/text:span&gt;&lt;text:span 
 text:style-name="OrgSuperscript"&gt;2&lt;/text:span&gt;">
---8<------------------------------------------------------------

The table:name attribute ends prematurely at the "OrgSubscript". This
could probably be mended by using single quotes for these attributes
inside attributes.

2. again #+CAPTION

   if the caption contains a string with /escaped/ underlines (please
   note the #+OPTION ^:t)
   
---8<------------------------------------------------------------
#+OPTIONS:   num:t toc:nil \n:nil @:t ::t |:t ^:t -:t f:t *:t <:nil

* First paragraph
#+CAPTION: YYYYMMDD\_hhmmss.csv. 
| head 1    | head2    |
|-----------+----------|
| content 1 | content2 |
---8<------------------------------------------------------------

    The escaped _ is not detected; instead, a literal \ and subscript
    are detected, and teh document is invalid as above. This escaping
    seems not to be mentioned in the documentation, probably because it
    is no longer valid. Might be of interest to people who want to
    convert old documents.

3. the \ref syntax seems not to work anymore; substituting by \ref{text}
   by [[text]] worked.       

4. Empty lines between #+NAME and the respective table are no longer tolerated

remarks: I like the new smart quotes. More comfortable than putting in
the utf characters via C-8-RET (which is handy, though, for simple
formulae).

Regards

Martin

-- 
Dr. Martin Gürtler 
Phone: +49-341-2254303
Mobile: +49-1578-4469069