introduce atom.onInit hook #2901
Conversation
dmaskasky
force-pushed
the
partial-sync-effect-4
branch
from
6f476d6 to
fb31cd5
Compare
dmaskasky
force-pushed
the
partial-sync-effect-4
branch
from
fb31cd5 to
c22a59b
Compare
dmaskasky
force-pushed
the
partial-sync-effect-4
branch
from
c22a59b to
c2daee3
Compare
dmaskasky
force-pushed
the
partial-sync-effect-4
branch
from
c2daee3 to
0beb306
Compare
dmaskasky
force-pushed
the
partial-sync-effect-4
branch
from
0beb306 to
6ed520e
Compare
dmaskasky
force-pushed
the
partial-sync-effect-4
branch
from
6ed520e to
1b0ccae
Compare
| type Batch = Set<() => void>[] & { | ||
| /** High priority functions */ | ||
| H: Set<() => void> | ||
| /** Medium priority functions */ | ||
| M: Set<() => void> | ||
| /** Low priority functions */ | ||
| L: Set<() => void> |
There was a problem hiding this comment.
If we were to introduce atomState.us, I don't think we need queue array. H/M/L would be probably more performant. If not, it still feels more readable. But, if queue array is more performant, I would change my mind.
There was a problem hiding this comment.
Bundle size with queue array is smaller, and the performance is the same. A queue array is nice because it makes flush evaluation agnostic to the shape of batch.
| error?: unknown | ||
| } | ||
|
|
||
| export function atomSyncEffect(effect: Effect): Atom<void> { |
There was a problem hiding this comment.
Can we keep this in effect.test.ts?
There was a problem hiding this comment.
Yes, I just have test for puzzle 3 right now. This PR is not ready to merge until all jotai-effect tests are passing.
There was a problem hiding this comment.
I'm keeping separate for now while I have this PR in draft. I still have to remove debugLabels, labeled atomStates, and console.logs as well.
dmaskasky
force-pushed
the
partial-sync-effect-4
branch
from
131c556 to
3839a86
Compare
dmaskasky
force-pushed
the
partial-sync-effect-4
branch
from
3839a86 to
76439b6
Compare
dmaskasky
force-pushed
the
partial-sync-effect-4
branch
from
76439b6 to
9edf139
Compare
dmaskasky
force-pushed
the
partial-sync-effect-4
branch
from
9edf139 to
080eeb7
Compare
dmaskasky
changed the title
|
Please reopen a new PR with https://github.com/pmndrs/jotai/tree/partial-sync-effect-4 branch. I will add my suggestions there. So, please avoid force pushing remote branches in the jotai repo. |
Related PRs
#2888
#2801
Summary
Uses atomState change and mount hooks to schedule effect and cleanup.
Key Points
atom.unstable_onInitthat fires in place when an atomState is first created in the store.storeandatomStateare passed as params toatom.unstable_onInit.atomState.hthat fires in-place when an atom has mounted or unmounted, and if an atom has unmounted it schedules a cleanup to run with batch priority 0.5. Batch is a param passed toatomState.hfunction.atomState.uthat fires in-place when an atom has updated, and if an atom has updated it schedules an effect to run with batch priority 0.5. Batch is a param passed toatomState.ufunction.Check List
pnpm run prettierfor formatting code and docs