feat: use a WeakMap instead of a symbol to link proxies to values #12868
Conversation
🦋 Changeset detectedLatest commit: 8f5a2b6 The changes in this PR will be included in the next version bump. This PR includes changesets to release 1 package
Not sure what this means? Click here to learn what changesets are. Click here if you're a maintainer who wants to add another changeset to this PR |
| @@ -10,7 +10,7 @@ import { queue_micro_task } from '../../task.js'; | |||
| */ | |||
| function is_bound_this(bound_value, element_or_component) { | |||
| // Find the original target if the value is proxied. | |||
| var proxy_target = bound_value && bound_value[STATE_SYMBOL]?.t; | |||
| var proxy_target = bound_value && proxies.get(bound_value)?.t; | |||
There was a problem hiding this comment.
A bit unrelated: Will this logic even work at all now? If it gets the underlying value, which IIUC is the original object, that one will likely not be the current value because of the change
|
Some benchmarks comparing a large table benchmark running locally on some core operations: Main:
This PR:
So whilst there is a performance loss, it doesn't seem like it's a huge regression. However, it is still a regression and likely one we need to consider vs alternative approaches. The other approach is to dynamically create the proxy object in a closure and thus have access to the metadata that way – which might be faster than a WeakMap, but I have reservations on that. |
Since #12867 doesn't really work, I figured we could try a
WeakMapapproach. It's possible that the performance implications will be a dealbreaker.Before submitting the PR, please make sure you do the following
feat:,fix:,chore:, ordocs:.Tests and linting
pnpm testand lint the project withpnpm lint