[O] Unicode characters export from babel code blocks ?

[BernardH]

Hi, In some literate program, I'd like to generate a data file. I do
it with a shell here-document (is there a better way just to inline
some text to tangle in a file?).  It works when tangling but not when
exporting to the pdf. Because there is a catch : I wanted to use
unicode \25A0 ■ and \25A1 □ instead of '#' and ' '. With
#+LATEX_HEADER: \usepackage{xltxtra} 
#+LATEX_HEADER: \setmainfont{DejaVu Serif} 
I can get the squares in the rgular text,
however, the exported code remains blank where the squares should be
☹.


Any help greatly appreciated !
Best Regards (and thx for the great ord-mode !)

Bernard

PS: excerpt from my babel code below
…
Using =\25A0= aka «black square» (■) and =\25A1= aka «white square» (□) to
represent the wall and empty spaces of the maze.

#+begin_src sh :file maze.txt :exports code :tangle yes
cat - <<END
□□□□□□□□□
□■■□□□■■□
□□□□□□□□□
□□□■□■□□□
□□□■□■□□□
□□□■■■□□□
□□□□□□□□□
□■■□□□■■□
□□□□□□□□□
END
#+end_src
…

the pdf does not containanything