Add a way to shield a stream's underlying channel #171
Conversation
Add a ``tractor._portal.StreamReceiveChannel.shield_channel()`` context manager which allows for avoiding the closing of an IPC stream's underlying channel for the purposes of task re-spawning. Sometimes you might want to cancel a task consuming a stream but not tear down the IPC between actors (the default). A common use can might be where the task's "setup" work might need to be redone but you want to keep the established portal / channel in tact despite the task restart. Includes a test.
goodboy
added
enhancement
tractor/_portal.py
Outdated
| @contextmanager | ||
| def shield_channel( | ||
| self | ||
| ) -> typing.AsyncGenerator['StreamReceiveChannel', None]: |
There was a problem hiding this comment.
Lol this type sig is obv wrong; originally made it async, for no good reason..
|
|
||
| async with tractor.open_nursery() as n: | ||
|
|
||
| stream = await(await n.run_in_actor( |
There was a problem hiding this comment.
Btw, what do y'all think about a ActorNursery.stream_from_actor() sugar for this pattern?
The portal is feeling a little awkward (too future-y) in the streaming case, but maybe it's just me?
There was a problem hiding this comment.
In theory you can actually just call Portal.result() before returning from .run_in_actor() but it means each call is then blocking and you'd have to spawn multiple (one-off) actors using tasks. Not sure if this is a problem and maybe will get hashed out with the discussion in #66.
|
I prefer one of these:
|
|
Sticking with |
|
Ok think this is ready finally (after wrangling some tests code along the way). I'll probably merge in the morning if noone pipes up. |
Add a
tractor._portal.StreamReceiveChannel.shield_channel()context manager which allows for avoiding the closing of an IPC stream's underlying channel for the purposes of task re-spawning.Sometimes you might want to cancel a task consuming a stream but not tear down the IPC
between actors (the default). A common use can might be where the task's "setup" work might need to be redone but you want to keep the established portal / channel in tact despite the task restart.
Includes a test.
I'm still not sure on the name here..
Any opinions @salotz , @guilledk?
Note it's used inside a
with stream.<meth name>Short list:
.shield_channel().shielded_channel().shield()boolproperty.shieldyou assign manually?.shield_from_close().channel_kept_open().channel_shielded()Thoughts appreciated.