[Merged by Bors] - Harden slot notifier against clock drift #3519
Conversation
michaelsproul
added
ready-for-review
There was a problem hiding this comment.
Nice! This changes the behaviour slightly regarding degenerate cases, if the function takes longer than one slot to execute then we'll skip slots. Previously we'd form a queue and process them all at once.
I think this new behaviour is better since with the current queuing approach we'd end up producing multiple events for the same slot. That seems to have little value, if any.
Happy to merge!
paulhauner
added
ready-for-merge
Yeah I think this might have been what @ethDreamer observed a few weeks ago! |
|
Penta-batch! bors r+ |
|
There are conflicts in this batch, but I'll let bors figure it out. bors r+ |
|
Already running a review |
## Issue Addressed Partly resolves #3518 ## Proposed Changes Change the slot notifier to use `duration_to_next_slot` rather than an interval timer. This makes it robust against underlying clock changes.
|
Build failed (retrying...): |
|
bors r+ |
|
Already running a review |
## Issue Addressed Partly resolves #3518 ## Proposed Changes Change the slot notifier to use `duration_to_next_slot` rather than an interval timer. This makes it robust against underlying clock changes.
bors
bot
changed the title
## Issue Addressed Partly resolves sigp#3518 ## Proposed Changes Change the slot notifier to use `duration_to_next_slot` rather than an interval timer. This makes it robust against underlying clock changes.
Issue Addressed
Partly resolves #3518
Proposed Changes
Change the slot notifier to use
duration_to_next_slotrather than an interval timer. This makes it robust against underlying clock changes.