Hi Aaron,
Aaron Ecay wrote:
> 2013ko apirilak 18an, Sebastien Vauban-ek idatzi zuen:
>> Bastien wrote:
>>> I applied this patch, thanks a lot. Please see the small changes
>>> I made to the ChangeLog entry for next commit messages:
>>>
>>> http://orgmode.org/cgit.cgi/org-mode.git/commit/?id=25869e
>>
>> How is that suppose to cooperate with ":eval never-export" (which avoids all
>> the evaluations during export)?
>
> I could be wrong, but I don’t seem to see a convenient way to set this
> header arg for both inline source blocks and #+call lines. These both
> have no way of storing their results in a buffer, so in order to have
> them be present in the export they must be evaluated every time.
>
> On the other hand, it is not always desirable to call code blocks on
> every export. They could be very time-consuming to compute. So, a way
> to pick out the non-results-storing blocks is needed.
>
>> I'm not sure to understand why this change is implemented as a variable which
>> we would have to set up in our .emacs or bind in the file, instead of an
>> header argument which we could set wherever we want (system-wide,
>> language-specific, buffer-wide, subtree-wide, or code
>> block-specific)...
>
> What form would this take? Would we have a value of :eval which is
> never-export-unless-inline?
As we're talking of evaluation, and as there is already an ":eval" header
argument (with lots of options), I think it'd better to try to add your
requirement as a new option, yes, instead of having complex rules for knowing
what happens depending on every option of ":eval" times 2 (with or without
your new value).
I'm not sure what ":eval never-export" currently does with inline blocks, but
(depending on that) I'd rather add the new value "never-export-except-inline"
(or "never-export-nor-inline").
Note -- The option names are not correct, just to carry what I mean...
That way, you can impose your new behavior wherever you want, as well in parts
of your Org document.
Best regards,
Seb
--
Sebastien Vauban
