Hi Sebastien,
Sebastien Vauban <sva-n...@mygooglest.com> writes:
> Andreas Leha wrote:
>> Sebastien Vauban <sva-n...@mygooglest.com> writes:
>>> Andreas Leha wrote:
>>>>> For me, that's the correct behavior, as inline code blocks are
>>>>> *only expected to be evaluated during export*.
>>>>
>>>> I disagree here.
>>>
>>> Though, this is what Eric Schulte wrote:
>>>
>>> ┌────
>>> │ Currently inline blocks like don't associate themselves with their
>>> │ results, they are only expected to be evaluated on export.
>>> └────
>>>
>>> If things have changed, I'm not aware of it.
>>>
>>>> As limiting the use of inline code to eval-on-export-only renders
>>>> all the org-babel-execute-subtree and related functionality useless.
>>
>> I have been using inline results for reports quite a bit, but not
>> lately. I might be wrong here, but from my memory they used to work
>> (i.e. were replaceable) for a while (at least when 'wrapped') unless
>> they were set to produce 'raw' results (which was a serious limitation
>> and led me to change my workflow).
>>
>> PS: a quick check reveals that they are indeed not replaceable (even
>> when wrapped)
>>
>> Nonetheless, from a literate programming perspective, I think that
>> replaceable (and raw) inline results are definitely desirable.
>> Regardless of the state of their implementation in orgmode right now.
>
> FWIW, I'm not -- yet? -- convinced we should see the results of inline
> code blocks inlined in the paragraph (and I'm not sure either it does
> not cause interpretation problems); but, for sure, I'd love to be able
> to preview the value interactively, at least.
>
>> So, I do not doubt, that you and Nicolas are right with that
>> replaceable inline results are not implemented and are -- from
>> orgmodes perspective -- expected to be evaluated only during export.
>>
>> My message was meant more as a feature request saying that I consider
>> replaceable inline results useful and would like to see them supported
>> by org.
>
> Could you better explain your statement: "Limiting the use of inline
> code to eval-on-export-only renders all the org-babel-execute-subtree
> and related functionality useless"?
>
> I'm not sure to fully understand your use-case. That'd certainly be
> worth explaining why you think it must be changed in the first instance
> if you'd like Eric or Nicolas (or someone else) to change that.
>
Consider this simple toy example. If you press 'C-c C-v s' in the
subtree, the document is 'broken' for export. If you press 'C-c C-v s'
twice or more the document is 'broken' even for display.
--8<---------------cut here---------------start------------->8---
#+PROPERTY: header-args:R :session *testR*
* Simplistic Literate Program
Let us generate some data $X$ with
#+begin_src R :exports code :results none
X <- rnorm(1:1000)
#+end_src
and plot the density of it
#+begin_src R :results graphics :file test.png :exports both
plot(density(X))
#+end_src
#+results:
[[file:test.png]]
This data has mean src_R{mean(X)}.
--8<---------------cut here---------------end--------------->8---
Regards,
Andreas
