Re: [O] Put result output in different type of code block than original

[Sebastien Vauban]

Hi Burton,

Burton Samograd wrote:
> "Sebastien Vauban" <address@hidden> writes:
>> Hi Rainer and Niels,
>>
>> Rainer M Krug wrote:
>>> On Mon, Oct 10, 2011 at 5:44 PM, Niels Giesen <address@hidden>wrote:
>>>> Say I've got a sh code block invoking curl to some json api, is it possible
>>>> some way to specify that the format of the output when :results output code
>>>> is in a json code block and *not* a sh code block?
>>
>> FYI, this is a need I also have, for example AWK outputting SQL code. There
>> has been some threads about this, but nothing is implemented so far in this
>> direction.
>>
>>>> #+begin_src sh :results output code :exports both
>>>> curl
>>>> https://our-service.org/getstuff?username=dirk\&password=catsbrithday\&fmt=json
>>>> #+end_src
>>>
>>> Does
>>>
>>> #+begin_src sh :results output :exports both
>>> echo #+BEGIN_SRC json
>>> curl
>>> https://our-service.org/getstuff?username=dirk\&password=catsbrithday\&fmt=json
>>> echo #+END_SRC
>>> #+end_src
>>>
>>> help?
>>
>> I like very much the idea of your workaround. Great thought!
>>
>> Though, nope, it does not help the OP, because the echo lines don't output
>> anything here, as # is the symbol to begin a comment (in shell scripts). So,
>> the first line simply outputs a blank line...
>
> Could you escape the comment character by using \#?

Yes, that does work as well (alternative to Eric's quoting method):

#+begin_src sh :results output raw :exports both
echo \#+BEGIN_SRC json
echo curl
echo 
https://our-service.org/getstuff?username=dirk\&password=catsbrithday\&fmt=json
echo \#+END_SRC
#+end_src

but my point was intended to be more general: this workaround may be difficult
to use for differnet languages to marry (my own need, in the past, was for AWK
outputting SQL + another I've forgotten about right now).

Anyway, this currently can be made to work, as shown.

Best regards,
  Seb

PS- I was expecting that multiple evaluations of the code block would lead to
    multiple results blocks, because of the raw specifier. It's not the case.
    Is is perfectly normal and explainable, or a corner case which favors --
    in this case -- this workaround?

-- 
Sebastien Vauban