Ihor Radchenko <yantar92@posteo.net> writes:

> I have one small nipick comment on the patch:
>
>>    (org-babel-comint-async-register
>>     session (current-buffer)
>>     "ob_comint_async_python_\\(start\\|end\\|file\\)_\\(.+\\)"
>> -   'org-babel-chomp 'org-babel-python-async-value-callback)
>> +   'org-babel-chomp 'org-babel-python-async-value-callback
>> +   t)
>
> Rather than `t', I'd use something more descriptive like 'disable-prompt-filtering.

Is the attached patch what you had in mind? If so I will squash it with
the previous patch.

You can also see both commits together in my branch here:
https://github.com/jackkamm/org-mode/tree/2024-ob-r-async-prompt-bugfix2

> Looks reasonable in general.

Thanks, look forward to applying it.

But first -- note that the current patch is on top of bugfix. I had
mentioned this in my original email but want to double check if it's
OK. In particular, I'm not sure if it's acceptable for bugfix branch
anymore, now that I'm changing the function signature of
`org-babel-comint-async-register' (albeit in a backward-compatible way).

I had originally proposed bugfix since I use R's table() function a lot
(as in my original example), and it was causing me problems when I
belatedly upgraded my work machine to Org 9.7.  But if you prefer, I can
rebase onto main.

Another possibility would be to add a hardcoded check on bugfix to skip
the prompt filtering if the major-mode is R or Python; then on main,
revert the hard-coded check, and update the signature of
`org-babel-comint-async-register' to set it properly with a
variable. But not sure it's worth the hassle -- I might just switch to
using main branch on my work machine at that point.