Improve juliacall async documentation #678
Replies: 2 comments 2 replies
-
Oh I need to add big notices to the docs saying that any support for multithreading is experimental and subject to change. One thing that I do know doesn't work very well is calling Julia from a Python thread that isn't the main thread. I think that as soon as Julia enters the task scheduler, it will never return to work on the Python thread, so that thread hangs. I think this is what's happening for you - so this is basically unsupported usage. I guess it would be handy to include known cases where it doesn't work, such as this, in the FAQ. |
Beta Was this translation helpful? Give feedback.
-
I'm mostly concerned about concurrency, not necessarily multithreading. So, are you saying that Julia can't be used with any backend that leverages python's asyncio / FastAPI? (that's many applications). Strangely, I was able to get something to partly work, but I'm now a bit concerned with the potential fragility. |
Beta Was this translation helpful? Give feedback.
Uh oh!
There was an error while loading. Please reload this page.
-
Hello Julia folks,
I recently to use Julia in a python project involving async function definitions, and came across some behavior / requirements that I think could be better documented. For example, it seems that precompiling methods may be needed (or perhaps this is related to GC during compiling) in
async
functions calls. Here is an example that hangs unless functions are precompiled:Beta Was this translation helpful? Give feedback.
All reactions