Understanding how priority escalation works
Understanding how priority escalation works êŽë š
Updated for Xcode 15
Every task can be created with a specific priority level, or it can inherit a priority from somewhere else. But in two specific circumstances, Swift will raise the priority of a task so itâs able to complete faster.
This always happens because of some specific action from us:
- If higher-priority task A starts waiting for the result of lower-priority task B, task B will have its priority elevated to the same priority as task A.
- If lower-priority task A has started running on an actor, and higher-priority task B has been enqueued on that same actor, task A will have its priority elevated to match task B.
In both cases, Swift is trying to ensure the higher priority task gets the quality of service it needs to run quickly â if something very important can only complete when something less important is also complete, then the less important task becomes very important.
For the most part, this isnât something we need to worry about in our code â think of it as a bonus feature provided automatically by Swiftâs tasks.
However, there is one place where priority escalation might surprise you, and itâs worth at least being aware of it: in our first situation, where a high-priority task uses await on a low-priority task, using Task.currentPriority will report the escalated priority rather than the original priority. So, you might create a task with a low priority, but when you query it a minute later it might have moved up to be a high priority.
At the time of writing, this is the only real âgotchaâ moment with task escalation. The other situation â if you queue a high-priority task on the same actor where a low-priority task is already running â will also involve priority escalation, but wonât change the value of currentPriority. This means your task will run a little faster and it might not be obvious why, but honestly itâs unlikely youâll even notice this.
Similar solutionsâŠ