Prevent runBlocking failure when the Worker it runs on terminates
#3585
Conversation
dkhalanskyjb
changed the title
runBlocking failing when the Worker it runs on terminatesrunBlocking failure when the Worker it runs on terminates
| fun keepAlive() { | ||
| Worker.current.executeAfter(afterMicroseconds = 100_000) { | ||
| if (!isCompleted) { | ||
| keepAlive() |
There was a problem hiding this comment.
I'm not really fond of this change.
It has really weird semantics "let's invoke something on a regular basis to workaround the issue that only reproduces during termination", it interferes with nested runBlockings (producing probably unnecessary tasks) and opens a can of worms for "what happens if 'start' or 'joinBlocking' throw system exception (e.g. OOM)". For now it seems it will create a task that is just stuck forever
There was a problem hiding this comment.
These are all good points, I fixed the code accordingly.
Regarding the weird semantics, I don't see anything weird about it. You could say, "to workaround the issue that only reproduces during termination" about any mechanism that recognizes something canceling or closing, yet such code is common. Here, we're just explicitly making sure the worker knows further work will be expected from it.
There was a problem hiding this comment.
Also, "keepalive" is a common term denoting exactly this approach, though from a different IT area: https://en.wikipedia.org/wiki/Keepalive
Fixes #3578
A competitor to #3584