Bruno and Ihor,

Tested the `bba-ngnu-pending-contents' branch locally — the
callback-based async example (`my-use-callbacks' engine, elisp source
block with `:nasync yes') runs end-to-end.  `org-babel-execute-src-block'
returns a reglock, the region locks, callbacks fire, the `#+RESULTS:'
table is populated correctly, and the reglock is released.  Smoke-level
only; haven't exercised the threads, futur, or dynamic-block paths yet.

One thing I want to flag: there's no integration code between
`org-pending' and `emacs-jupyter' yet on either side — `ob-jupyter.el'
has its own async-overlay path (`jupyter-org-pending-async-results').
The validation you suggested against jupyter would mean writing a
bridge (a small `jupyter-org-pending-schedule' wrapper using
`:execute-with').  I'm happy to take that on, but I'd rather get FSF
copyright assignment in flight first so any patches can land.  I'll
start that this week.

Pete


> Peter Tittmann <[email protected]> writes:
>
>> Quick sync on where things stand on my end: I rebased your
>> pending-contents branch onto current org-mode HEAD a few days
>> ago. Two merge conflicts in ob-core.el — both resolved by taking
>> your async execution path. The rebased branch is at
>> gitlab.com/petewt/org-pending on the rebase-pending-contents branch
>> if you want to compare against your updated version. Happy to
>> reconcile if our changes diverge.
>>
>> On emacs-jupyter — that would be a great validation
>> case. Nathaniel's async architecture in jupyter-org-client.el is the
>> immediate consumer for org-pending, and I have a working test
>> environment with [scimax](<https://github.com/jkitchin/scimax>). Once
>> you've updated your branch this weekend, we could coordinate on
>> testing.
>
> Peter, note that Bruno has created a new branch.
> I think we are close to done with org-pending.el itself. Certainly done
> API-wise. The next step is ob-core/ob-comint design. For emacs-jupyter,
> we should either design ob-core and/or ob-comint to support all the
> things you need or you need to use org-pending directly. Your input will
> be appreciated.


Reply via email to