Conversation
🦋 Changeset detectedLatest commit: b1a6b8c 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 |
|
preview: https://svelte-dev-git-preview-svelte-14977-svelte.vercel.app/ this is an automated message |
|
|
Brought this up to speed with Not really sure how best to resolve that so for now I'm not going to try. |
Yeah i planned to get back to this after the writable deriveds PR but wanted to check if that's the road we want to go first. I can work on this if we want. |
|
any news regarding this @paoloricciuti ? |
|
going to close this in favour of #17674 |
I tried to bring #14977 up-to-date but it's slipped too far. Also, I wanted to try a slightly different approach. In this PR, deriveds are memoized if they're created during render — in other words if you have something like this... ```svelte <script> let thing = $derived(expensivelyComputeThing()); </script> ``` ...`thing` will only be computed once. This seems correct since the inputs should never change during render. For deriveds created _outside_ render, we re-run the derived each time it is accessed, which fixes #14954. This way, there's still _some_ overhead compared to how deriveds work in the browser (where they only recompute when their dependencies have changed), but only in the rare places where it is necessary. There is one wrinkle: writable deriveds. On `main` these are just regular old variables, which means they can be written to during render. This PR currently preserves that behaviour, but I'm not sure it's desirable. It prevents the values of non-render-bound deriveds from ever updating, and makes no sense in the context of render-bound deriveds since they shouldn't be changing during render _anyway_. So my preference would be to disallow writes to deriveds on the server, but I'm not sure if we would need to consider that a breaking change. Draft because of that question, and also because I think we might be able to tidy up some stuff around class fields. - [x] figure out if we can delete some existing code around derived class fields - [x] figure out what to do about writable deriveds - [x] add a test ### Before submitting the PR, please make sure you do the following - [x] It's really useful if your PR references an issue where it is discussed ahead of time. In many cases, features are absent for a reason. For large changes, please create an RFC: https://github.com/sveltejs/rfcs - [x] Prefix your PR title with `feat:`, `fix:`, `chore:`, or `docs:`. - [x] This message body should clearly illustrate what problems it solves. - [x] Ideally, include a test that fails without this PR but passes with it. - [x] If this PR changes code within `packages/svelte/src`, add a changeset (`npx changeset`). ### Tests and linting - [x] Run the tests with `pnpm test` and lint the project with `pnpm lint`
Closes #14954
I took a shot with the idea of thunkify every derived on the server to make them "work" in case of reassignment. As pointed out by @Rich-Harris this have some downside:
But I think regardless of the solution we will go with we will probably need functions in some way so this PR can be a good start.
I got 90% there in 10 minutes and then faced a big problem with destructured deriveds. Since we want every destructured identifier to be a function the only way i thought of is to do something like this.
it's a bit more generated code but it works fine.
Again, this is a proposal, don't know if this the actual direction we want to go but even if it's not it could be a good base.
Before submitting the PR, please make sure you do the following
feat:,fix:,chore:, ordocs:.packages/svelte/src, add a changeset (npx changeset).Tests and linting
pnpm testand lint the project withpnpm lint