fix AbortController re-initialization (electric-db-collection)

[fcannizzaro]

Hi! I debugged what was happening in #418 and finally I discovered the issue's cause.

The AbortController inside the sync config of electric-db-collection after being aborted (by the cleanup fn after the gcTime and 0 active subscribers) was never re-initialized again.

And cause the signal was already aborted, the electric client skips the sync request! (https://github.com/electric-sql/electric/blob/3baf67a11887fea0bd046b06e081a93f16c88173/packages/typescript-client/src/client.ts#L474)

I only moved the AbortController instantiation inside the sync() function instead of the parent scope.

[changeset-bot]

🦋 Changeset detected

Latest commit: 3c47b2a

The changes in this PR will be included in the next version bump.

This PR includes changesets to release 1 package

Name	Type
@tanstack/electric-db-collection	Patch

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

[pkg-pr-new]

More templates

• tanstack-start-db-projects-example
• @tanstack/db-example-react-todo
• @tanstack/db-example-solid-todo

@tanstack/db

npm i https://pkg.pr.new/@tanstack/db@523

@tanstack/db-ivm

npm i https://pkg.pr.new/@tanstack/db-ivm@523

@tanstack/electric-db-collection

npm i https://pkg.pr.new/@tanstack/electric-db-collection@523

@tanstack/query-db-collection

npm i https://pkg.pr.new/@tanstack/query-db-collection@523

@tanstack/react-db

npm i https://pkg.pr.new/@tanstack/react-db@523

@tanstack/solid-db

npm i https://pkg.pr.new/@tanstack/solid-db@523

@tanstack/svelte-db

npm i https://pkg.pr.new/@tanstack/svelte-db@523

@tanstack/trailbase-db-collection

npm i https://pkg.pr.new/@tanstack/trailbase-db-collection@523

@tanstack/vue-db

npm i https://pkg.pr.new/@tanstack/vue-db@523

commit: 3c47b2a

[samwillis]

This is great, really well spotted.

I don't suppose you could try and create a test that reproduces the problem so we can validate this fix and prevent a regression in future?

[KyleAMathews]

Looks great! Could you add a test which reproduces the bug? Call cleanup & then query again to make sure sync starts up again?

[fcannizzaro]

Tomorrow I'll make these changes and write a test for it

[KyleAMathews]

I'm pretty sure I've run into the same bug so thanks for figuring it out!

[KyleAMathews]

A customer is running into this in production so I'm going to add a test to get this out. Thanks again for debugging things and putting up the PR!!

[fcannizzaro]

Yeah if you can write the test since I tried multiple ways to write it but I don't know how to test this case.. cause the mock is done on a global AbortController and ShapeStream but now it changes for every sync() so how it should be done ?

[KyleAMathews]

Yeah, I couldn't get it to work either after trying for a bit — our mocks need modified quite a bit to make a test work. Let's merge and release.

[KyleAMathews]

I deployed this to my app and confirmed it fixed the bug