Reasons for connecting to Jupyter kernel to hang?

[johnW-ret]

I'm trying to add ocaml-jupyter Comms support for variable sharing, and I noticed that the line registering a target for my init script would always cause evaluation to hang. (I was using a custom build to skip the recognized language check here, though it shouldn't matter since the hang occurs here, before the call to CreateCommandEventChannelAsync.)

I thought it was an issue with my code or the library I was using, but now using the production build (1.0.507501+7b66b9f76838c0b391c3ee1827d8bccc30eecc6a), I am getting hanging issues connecting to an Anaconda kernel. Running #!connect jupyter with an init script of something like print(5) shows the code output, but the next cells that uses the kernel clocks forever.

---

While writing this post, I noticed the Anaconda kernel successfully completes the cell with a Kernel added message but hangs on the next cell using the new kernel as described above, while the OCaml kernel hangs on #!connect jupyter cell, running up to the point the Comm target is created, so the behavior between the two is different. However, because both targets have no issues before the Comms target is created and then have hanging issues after, I imagine they might be related.

Variable sharing with Python was working last night, and I haven't had any environment changes since then, so I'm wondering why it isn't working? I've tried both running code from Anaconda prompt and not.. nothing.

Any tips appreciated 🙂

[johnW-ret]

Update:
Python kernel no longer hangs on my current dev build (commit c791f80fc26328349da56af58dfaf4f721917703) but does still hang for release version 1.0.507501. OCaml kernel behavior is still the same and may be unrelated.