Hi Rainer,
really cool. I am very short on time right now, so I can't promise I'll
get around to test it soon.
Just one comment:
Rainer M Krug <rai...@krugs.de> writes:
> On 01/20/14, 19:38 , Eric Schulte wrote:
>>>
>>> I will check it a little bit longer and see that I can display
>>> the error message in the dummy graphic. Should have it by
>>> tomorrow.
>>>
>>
>> Sounds good. Alternately, maybe you could get the R process to
>> fail and print an error to STDERR so that the export will stop
>> immediately and inform the user of the error.
>
> I think it is better to catch the error and return with the
> computations in the next block, because
>
> 1) a code block which produces a graph is most often an end point,
> i.e. further blocks should (as I see it) not rely on the output of
> this block
If all (your) documents follow that rule, that blocks producing
graphical output are not needed later on, than it might be cool to have
an [draft]-like option in Org that by default replaces all images (not
just failing ones) with a cheap-to-export graphic. For some of my
documents that is a real time saver.
I can mimic that now by passing a variable 'draft' to the code blocks
and deal with it internally. But that means a lot of manual coding.
>
> 2) if the export continues, one can fix more then one error at a time,
> so being more efficient.
>
> 3) If the export continues, it makes it possible to use this to create
> placeholder graphs:
>
> #+begin_src R :file TheFantasticGraph.pdf :results graphics
> stop("This is a placeholder for a new fantastic graph"
> #+end_src
>
> Where TheFantasticGraph will display the message "This is a
> placeholder for a new fantastic graph".
>
> I have now added the error to the normal R output in addition to a
> graph, but the export is not canceled.
>
>
>>
>>>
>>> I'll send you the patch then.
>>>
>>
>> Sounds great, thanks.
>
> OK - here it is attached (my first patch to org :-) ) - let me know if
> it is OK.
>
Really nice addition! Thanks already (without testing the patch)!
[ ... ]
Regards,
Andreas
