On 23/07/17 09:52, Nicolas Goaziou wrote:
> Hello,
>
> Neil Jerram <neil@ossau.homelinux.net> writes:
>
>> I don't think that Scheme output results are correctly handled. Please
>> consider:
>>
>> #+BEGIN_SRC scheme :results output
>> (display "a") (newline)
>> (display "b") (newline)
>> (display "c") (newline)
>> '(a b c)
>> #+END_SRC
>>
>>
>> #+RESULTS: : "a\nb\nc\n"
>>
>> As compared with the Elisp equivalent:
>>
>> #+BEGIN_SRC elisp :results output
>> (princ "a") (terpri)
>> (princ "b") (terpri)
>> (princ "c") (terpri)
>> '(a b c)
>> #+END_SRC
>>
>> #+RESULTS:
>> : a
>> : b
>> : c
>>
>> I have a possible fix for this - on another computer, so not to hand
>> right now - but thought it would be check first whether you agree with
>> me that the Scheme results should be more like the Elisp ones.
> That sounds like a good idea. Could you send your fix as a proper patch
> with a commit message?
>
> Thank you!
>
> Regards,

With apologies for the slow follow up, attached is a patch that works 
well for me.

I think a concern with it would be whether the Geiser retort object and 
functions have always existed, and whether we can rely on them 
continuing to exist.  But - at least for me - every Scheme evaluation 
was previously giving "An error occurred", because of an unexpected 
extra "Mark set" message, so I am pretty sure that my patch is an 
improvement.

I guess (following any revisions) you will also need assignment papers.  
I believe I have these on file for Guile already (since about 1990), and 
it's possible that those are worded so as to cover Emacs as well.  Would 
you mind checking, as I've forgotten how to do that myself?

Regards - Neil