> On Aug 7, 2020, at 8:39 AM, William McCoy <wdm8...@gmail.com> wrote:
> 
> This use of :prologue appeared to me to be very useful.  But for some reason 
> when I try it out it does not work for me.  I just get a message that the 
> code block produced no output and that 'np' is not defined.  Just to check, 
> when I put the import statements directly within my code block it works fine.
> 
> I am running:  Org mode version 9.3.7 (9.3.7-16-g521d7f-elpa
> 
> Any idea what I'm doing wrong?
> 
> 

It is sometimes useful to use C-c C-v C-i to see what header args org has 
detected for a source block. Misspelled words sometimes wreak havoc and 
invisible characters can cause real pain.


Also, it helps to use C-c C-v C-v to to see the expanded code block. When I do 
this with Kens' ECM, I get

import numpy as np; import os
print(np.__version__)

in the preview buffer.

HTH,

Chuck


> On 8/6/20 2:12 PM, Ken Mankoff wrote:
>> Actual example:
>> 
>> 
>> * Prologue test
>> :PROPERTIES:
>> :header-args:python+: :prologue "import numpy as np; import os"
>> :END:
>> 
>> #+BEGIN_SRC python :results output
>> print(np.__version__)
>> #+END_SRC
>> 
>> #+RESULTS:
>> : 1.18.4
>> 
>> 
>> 
>> 
>> On Wed, Aug 5, 2020 at 3:03 PM Ken Mankoff <mank...@gmail.com> wrote:
>> What about using :pre or :prologue and setting it at the header or document 
>> level?
>> 
>> Please excuse brevity. Sent from tiny pocket computer with 
>> non-haptic-feedback keyboard.
>> 
>> On Wed, Aug 5, 2020, 14:22 George Mauer <gma...@gmail.com> wrote:
>> Use case:
>> 
>> I'm using ob-racket but this would apply just as well to a few other 
>> workflows I have with python or js.
>> 
>> I would like to write a helper function in a src block and then 
>> automatically have access to it in other src blocks further down the 
>> document. I don't really want a stateful session (nor does ob-racket support 
>> sessions) so I essentially want the equivalent of automatically including it 
>> everywhere so I don't have to type it out all the time (and have it screw up 
>> syntax coloring/indentation).
>> 
>> Is this currently possible? Does anyone have any ideas for how to extend 
>> things so it is?
> 



Reply via email to