Repository navigation
Handle Channel force-close with pending blocked monitors blocking another channel #2168
Description
Activity
Also #2167 (comment)
Re: #2167 (comment) we should be good as of #2167, I pushed the requested comment in #2391.
Gonna go ahead and close this as completed in #2167 - while its possible there's some edge cases missed, the tests in #2169 show at least the basic cases here were fixed though I'm sill worried a bit about the completion-notification pipeline which I want to rewrite in 117 as #2389.
Metadata
Metadata
Assignees
Labels
No labels
Now (post-#2167 and its followup) that a monitor can block another channel from making progress we need to make sure that a channel force-closing while blocking another monitor (implying it could leave the other monitor blocked if we restart without completing the required monitors which are gone since the channel was deleted) doen't hang us.