#16977 had one slight regression which might contribute to #16990: The batch from earlier was restored, but that doesn't make sense in this situations since this has nothing to do with our new async logic of batches suspending until pending work is done. As a result you could end up with a batch being created, and then the restore then instead reverting to an earlier batch that was already done, which means a ghost-batch ends up in the set of batches, subsequently triggering time traveling when it shouldn't. This may help with #16990
No test because basically impossible to do so
The previous check was flawed because EFFECT_RAN would be set by the time it is checked, since a promise in a parent component will cause a delay of the inner component being instantiated. Instead we have a new field on the component context checking if the component was already popped (if se we are indeed too late). Don't love it to have a field just for this but I don't see another way to reliably check it.
Fixes#16629
* chore: run boundary async effects in the context of the current batch
* WIP
* reinstate kludge
* fix test
* WIP
* WIP
* WIP
* remove kludge
* restore batch_values after commit
* make private
* tidy up
* fix tests
* update test
* reset #dirty_effects and #maybe_dirty_effects
* add test
* WIP
* add test, fix block resolution
* bring async-effect-after-await test from defer-effects-in-pending-boundary branch
* avoid reawakening committed batches
* changeset
* cheat
* better API
* regenerate
* slightly better approach
* lint
* revert this whatever it is
* add test
* Update feature description for fork API
* error if missing experimental flag
* rename inspect effects to eager effects, run them in prod
* regenerate
* Apply suggestions from code review
Co-authored-by: Simon H <5968653+dummdidumm@users.noreply.github.com>
* tidy up
* add some minimal prose. probably don't need to go super deep here as it's not really meant for non-framework authors
* bit more detail
* add a fork_timing error, regenerate
* unused
* add note
* add fork_discarded error
* require users to discard forks
* add docs
* regenerate
* tweak docs
* fix leak
* fix
* preload on focusin as well
* missed a spot
* reduce nesting
---------
Co-authored-by: Simon H <5968653+dummdidumm@users.noreply.github.com>
* fix: coordinate mount of snippets with await expressions
* try this
* deduplicate
---------
Co-authored-by: Simon H <5968653+dummdidumm@users.noreply.github.com>
The "is this a transparent effect we can ignore" logic for determining whether or not to play a local transition didn't account for pruned block effects. This fix marks the child branch as transparent if the block effect was, and during traversal then checks if the branch is the child of a block - if not that means it's the child of a pruned block effect.
fixes#16826
* perf: skip repeatedly traversing the same derived
We introduce a new flag which marks a derived during the mark_reaction phase and lift it during the execution phase.
* fix
This introduces a new flag which marks a derived as visited during the mark_reaction phase and lifts it during the dirty-check/execution phase (both are necessary because not all deriveds are necessarily executed, but they're always checked).
We could've used a Set on mark_reactions which would've been simpler to reason about, alas it was performing consistenly worse on the kairo_mux_unowned / kairo_mux_owned benchmarks (it kinda makes sense generally, creating and filling a set is more expensive than juggling flags).
Closes#16658, which showcases are particularly gnarly signal graph where many deriveds point to the same deriveds, where we would be traversing the same reactions over and over if we didn't do this.
I also added a similar measure (this time a set/map is unavoidable/makes more sense) for mark_effects, hopefully this helps for #16990
at present we only call batch.increment() when something async happens if we're not inside a pending boundary, but that's incorrect — it means that a batch is committed before everything resolves. When work inside a pending boundary does resolve, the batch becomes a zombie.
At the same time, we don't handle effects inside pending boundaries correctly. They should be deferred until the boundary (and all its parents) are ready.
This PR attempts to fix that — during traversal, when we exit a pending boundary, any effects that were collected get deferred until the next flush. We also distinguish between batch.#pending (any ongoing async work) and batch.#blocking_pending (any async work that should prevent effects outside pending boundaries from being flushed).
* fix: make `$inspect` logs come from the callsite
* default to showing stack trace
* reuse id
* DRY out, improve server-side inspect
* update docs
* ugh
* fix the dang tests
* ugh windows is there no punch bowl you won't poop in?
* argh
* how about this
* alright finally
Fixes#16850, fixes#16775, fixes#16795, fixes#16982#16631 introduced a bug that results in the effects within guards being evaluated before the guards themselves. This fix makes sure to iterate the block effects in the correct order (top down)
---------
Co-authored-by: Simon H <5968653+dummdidumm@users.noreply.github.com>
Co-authored-by: Simon Holthausen <simon.holthausen@vercel.com>
* WIP implement `$effect.pending(...)`
* feat: `$state.eager(value)` (#16926)
* runtime-first approach
* revert these
* type safety, lint
* fix: better input cursor restoration for `bind:value` (#16925)
If cursor was at end and new input is longer, move cursor to new end
No test because not possible to reproduce using our test setup.
Follow-up to #14649, helps with #16577
* Version Packages (#16920)
Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
* docs: await no longer need pending (#16900)
* docs: link to custom renderer issue in Svelte Native discussion (#16896)
* fix code block (#16937)
Updated code block syntax from Svelte to JavaScript for clarity.
* fix: unset context on stale promises (#16935)
* fix: unset context on stale promises
When a stale promise is rejected in `async_derived`, and the promise eventually resolves, `d.resolve` will be noop and `d.promise.then(handler, ...)` will never run. That in turns means any restored context (via `(await save(..))()`) will never be unset. We have to handle this case and unset the context to prevent errors such as false-positive state mutation errors
* fix: unset context on stale promises (slightly different approach) (#16936)
* slightly different approach to #16935
* move unset_context call
* get rid of logs
---------
Co-authored-by: Rich Harris <rich.harris@vercel.com>
* fix: svg `radialGradient` `fr` attribute missing in types (#16943)
* fix(svg radialGradient): fr attribute missing in types
* chore: add changeset
* Version Packages (#16940)
* Version Packages
* Update packages/svelte/CHANGELOG.md
---------
Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
Co-authored-by: Rich Harris <rich.harris@vercel.com>
* chore: simplify `batch.apply()` (#16945)
* chore: simplify `batch.apply()`
* belt and braces
* note to self
* unused
* fix: don't rerun async effects unnecessarily (#16944)
Since #16866, when an async effect runs multiple times, we rebase older batches and rerun those effects. This can have unintended consequences: In a case where an async effect only depends on a single source, and that single source was updated in a later batch, we know that we don't need to / should not rerun the older batch.
This PR makes it so: We collect all the sources of older batches that are not part of the current batch that just committed, and then only mark those async effects as dirty which depend on one of those other sources. Fixes the bug I noticed while working on #16935
* fix: ensure map iteration order is correct (#16947)
quick follow-up to #16944
Resetting a map entry does not change its position in the map when iterating. We need to make sure that reset makes that batch jump "to the front" for the "reject all stale batches" logic below. Edge case for which I can't come up with a test case but it _is_ a possibility.
* feat: add `createContext` utility for type-safe context (#16948)
* feat: add `createContext` utility for type-safe context
* regenerate
* Version Packages (#16946)
Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
* chore: Remove annoying sync-async warning (#16949)
* fix
* use `$state.eager(value)` instead of `$effect.pending(value)`
---------
Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
Co-authored-by: Hyunbin Seo <47051820+hyunbinseo@users.noreply.github.com>
Co-authored-by: Ben McCann <322311+benmccann@users.noreply.github.com>
Co-authored-by: Rich Harris <rich.harris@vercel.com>
Co-authored-by: Hannes Rüger <hannes@hannesrueger.de>
Co-authored-by: Elliott Johnson <sejohnson@torchcloudconsulting.com>
* decouple from boundaries
* use queue_micro_task
* add test
* fix
* changeset
* revert
* tidy up
* update docs
* Update packages/svelte/src/internal/client/reactivity/batch.js
Co-authored-by: Simon H <5968653+dummdidumm@users.noreply.github.com>
* minor tweak
---------
Co-authored-by: Simon H <5968653+dummdidumm@users.noreply.github.com>
Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
Co-authored-by: Hyunbin Seo <47051820+hyunbinseo@users.noreply.github.com>
Co-authored-by: Ben McCann <322311+benmccann@users.noreply.github.com>
Co-authored-by: Hannes Rüger <hannes@hannesrueger.de>
Co-authored-by: Elliott Johnson <sejohnson@torchcloudconsulting.com>
quick follow-up to #16944
Resetting a map entry does not change its position in the map when iterating. We need to make sure that reset makes that batch jump "to the front" for the "reject all stale batches" logic below. Edge case for which I can't come up with a test case but it _is_ a possibility.
Since #16866, when an async effect runs multiple times, we rebase older batches and rerun those effects. This can have unintended consequences: In a case where an async effect only depends on a single source, and that single source was updated in a later batch, we know that we don't need to / should not rerun the older batch.
This PR makes it so: We collect all the sources of older batches that are not part of the current batch that just committed, and then only mark those async effects as dirty which depend on one of those other sources. Fixes the bug I noticed while working on #16935
* fix: unset context on stale promises
When a stale promise is rejected in `async_derived`, and the promise eventually resolves, `d.resolve` will be noop and `d.promise.then(handler, ...)` will never run. That in turns means any restored context (via `(await save(..))()`) will never be unset. We have to handle this case and unset the context to prevent errors such as false-positive state mutation errors
* fix: unset context on stale promises (slightly different approach) (#16936)
* slightly different approach to #16935
* move unset_context call
* get rid of logs
---------
Co-authored-by: Rich Harris <rich.harris@vercel.com>
If cursor was at end and new input is longer, move cursor to new end
No test because not possible to reproduce using our test setup.
Follow-up to #14649, helps with #16577
Test for #16912
Also some explanation what the bug was:
1. async batch kicks off
2. outer async work succeeds, still something pending, so doesn't do anything for now
3. something unrelated writes to a signal (in the remote functions case it's the query writing to loading, raw etc), which creates a new batch
4. new batch executes. since there are multiple batches it takes the previous value which means if block is still alive. commits that, since no async work from the perspective of this branch
5. inner async work succeeds. now the batch has zero pending async work so it can flush. But the if block is no longer dirty since it was done by the other batch already -> never undos the other work
#16912 fixes it by still traversing the tree which means the if block deletion is scheduled to commit later, which it then does
We have an each block optimization that omits the comment when the each block is the sole child of an element. This optimization clashes with async which wants to skip ahead to the sibling closing comment.
For now we therefore remove that optimization when the each block is async. In the long run we should instead optimize _all_ cases where _any_ block is the sole child of an element, in both async and sync mode, consistently.
fixes#16905fixes#16907
* fix: update error message to clarify usage with radio inputs
* show an extra bind checked error when a radio type
* typo
* replace 'group' binding with , drive-by fix some other messages
---------
Co-authored-by: Rich Harris <rich.harris@vercel.com>
Race them against each other - in almost all cases requestAnimationFrame will fire first, but e.g. in case the window is not focused or a view transition happens, requestAnimationFrame will be delayed and setTimeout helps us resolve fast enough in that case
Fixes#16429
Fixes https://github.com/sveltejs/kit/issues/14220
* fix: depend on reads of deriveds created within reaction (async mode)
As part of https://github.com/sveltejs/kit/pull/14481 we discovered that deriveds created within reactions and reading from them in that same reaction is actually useful in some cases, as such a use case we couldn't imagine yet in #15564 has appeared.
We think it's ultimately better to rerun on those cases, so we're going to make this change in async mode (that way the behavior doesn't change unless you have enabled the experimental flag)
* fix tests
Fixes#16816Fixes#16811
---------
Co-authored-by: Simon Holthausen <simon.holthausen@vercel.com>
Co-authored-by: Simon H <5968653+dummdidumm@users.noreply.github.com>
* feat: First pass at payload
* first crack
* snapshots
* checkpoint
* fix: cloning
* add test option
* big dumb
* today's hard work; few tests left to fix
* improve
* tests passing no wayyyyy yo
* lots of progress, couple of failing tests around selects
* meh
* solve async tree stuff
* fix select/option stuff
* whoop, tests
* simplify
* feat: hoisting
* fix: `$effect.pending` sends updates to incorrect boundary
* changeset
* stuff from upstream
* feat: first hydrationgaa
* remove docs
* snapshots
* silly fix
* checkpoint
* meh
* ALKASJDFALSKDFJ the test passes
* chore: Update a bunch of tests for hydration markers
* chore: remove snippet and is_async
* naming
* better errors for sync-in-async
* test improvements
* idk man
* merge local branches (#16757)
* use fragment as async hoist boundary
* remove async_hoist_boundary
* only dewaterfall when necessary
* unused
* simplify/fix
* de-waterfall awaits in separate elements
* update snapshots
* remove unnecessary wrapper
* fix
* fix
* remove suspends_without_fallback
---------
Co-authored-by: Rich Harris <rich.harris@vercel.com>
* Update payload.js
Co-authored-by: Rich Harris <rich.harris@vercel.com>
* checkpoint
* got the extra children to go away
* just gonna go ahead and merge this as the review comments take up too much space
* chore: remove hoisted_promises (#16766)
* chore: remove hoisted_promises
* WIP optimise promises
* WIP
* fix <slot> with await in prop
* tweak
* fix type error
* Update packages/svelte/src/compiler/phases/3-transform/server/visitors/SvelteHead.js
* chore: fix hydration treeshaking (#16767)
* chore: fix hydration treeshaking
* fix
* remove await_outside_boundary error (#16762)
* chore: remove unused analysis.boundary (#16763)
* chore: simplify slots (#16765)
* chore: simplify slots
* unused
* Apply suggestions from code review
* chore: remove metadata.pending (#16764)
* Update packages/svelte/src/compiler/phases/3-transform/server/visitors/SnippetBlock.js
* put this back where it was, keep the diff small
* Update packages/svelte/src/compiler/phases/types.d.ts
Co-authored-by: Rich Harris <rich.harris@vercel.com>
* chore: remove analysis.state.title (#16771)
* chore: remove analysis.state.title
* unused
* chore: remove is_async (#16769)
* chore: remove is_async
* unused
* Apply suggestions from code review
Co-authored-by: Rich Harris <rich.harris@vercel.com>
* cleanup
* lint
* clean up payload a bit
* compiler work
* run ssr on sync and async
* prettier
* inline capture
* Update packages/svelte/src/compiler/phases/3-transform/server/visitors/EachBlock.js
* chore: simplify build_template (#16780)
* small tweak to aid greppability
* chore: fix SSR context (#16781)
* at least passing
* cleanup
* fix
* remove push/pop from exports, not needed with payload
* I think this is better but tbh not sure
* async SSR
* qualification
* errors:
* I have lost the plot
* finally
* ugh
* tweak error codes to better align with existing conventions, such as they are
* tweak messages
* remove unused args
* DRY out a bit
* unused
* unused
* unused
* simplify - we can enforce readonly at a type level
* unused
* simplify
* avoid magical accessors
* simplify algorithm
* unused
* unused
* reduce indirection
* TreeState -> SSRState
* mark deprecated methods
* grab this.local from parent directly
* rename render -> fn per conventions (fn indicates 'arbitrary code')
* reduce indirection
* Revert "reduce indirection"
This reverts commit 3ec461baad.
* tweak
* okay works this time
* no way chat, it works
* fix context stuff
* tweak
* make it chainable
* lint
* clean up
* lint
* Update packages/svelte/src/internal/server/types.d.ts
Co-authored-by: Rich Harris <rich.harris@vercel.com>
* sunset html for async
* types
* we use 'deprecated' in other messages
* oops
---------
Co-authored-by: Rich Harris <rich.harris@vercel.com>
#16773 added a test with _expected.html to a wrong suit, so it was ignored, and this allowed to slip in a new bug of printing the falsy hidden attribute as hidden (shortened enabled form).
That fix wasn't released yet, so no changeset.
* chore: generate CSS hash using the filename
* fix all tests but one
* slightly kludgy fix
* try this
* fix
---------
Co-authored-by: Rich Harris <rich.harris@vercel.com>
Consolidates our scheduling system that had diverged into two - queue_micro_task and Batch.ensure-queues and resolves some edge case race conditions related to flushSync along the way.
* fix: send `$effect.pending` count to the correct boundary
* make boundary.pending private, use boundary.is_pending consistently
* move error to correct place
* we need that error
* update JSDoc
When an async derived already has a previous promise that is still pending, we were not accessing the `then` property of the new promise. If that property access causes signals to be read, that meant that those dependencies were lost and as such the derived wouldn't rerun anymore when it should. The fix is to make sure to always access the thennable.
* fix : remove cursor manipulation for input bindings
Old Fix: Restore input binding selection position (#14649)
Current Fix: Remove unnecessary cursor manipulation as the presence of runes no longer requires special handling.
* fix : add change set to my previous commit
* Revert "fix : add change set to my previous commit"
This reverts commit 6ca8ef3f97.
* fix: revert previous changeset added new to fix lint errors
* chore : resolve lint error to fix pipeline issue
* Revert "fix: revert previous changeset added new to fix lint errors"
This reverts commit 91094949a6.
* fix: input binding to handle code in a synchronous manner
Introduced Promise.resolve to ensure that the 'set' operation completes before the 'get' operation Minimizing update delays.
* Fix: resolve cursor jumps and change sets
* better fix
* test
* changeset
* simplify
* failing test
* gah we can't fix the input in an effect, need to do it here, but after a tick so that changes have been flushed through each blocks
* add explanatory comment
* fix test
* this seems to work?
---------
Co-authored-by: Hariharan Srinivasan <hariharan.srinivasan@kadfire.com>
Co-authored-by: Rich Harris <rich.harris@vercel.com>
Fixes#16591
This PR introduces a check for builtin custom elements (is attribute) inside the set_custom_element_data helper in order to correctly set properties that have a setter.
Implemented by reusing the `async_body` function inside `Fragment.js`. Also removes the ability to reference a `{@const ...}` of an implicit child inside a boundary pending/failed snippet:
- existing duplication of consts can have unintended side effects, e.g. async consts would unexpectedly called multiple times
- what if a const is the reason for the failure of a boundary, but is then referenced in the failed snippet?
- what if an async const is referenced in a pending snippet? deadlock
- inconsistent with how it behaves for components where this already does not work
Implemented via the experimental flag so the behavior change only applies there as this is a breaking change strictly speaking. Also added a compiler error for this.
closes#16462
* fix: properly catch top level await errors
async errors within the template and derived etc are properly handled because they know about the last active effect and invoke the error boundary correctly as a response. This logic was missing for our top level await output.
Fixes#16613
* test
* use helper for async bodies (#16641)
* use helper for async bodies
* unused
* fix
* failing test + fix
---------
Co-authored-by: Simon Holthausen <simon.holthausen@vercel.com>
---------
Co-authored-by: Rich Harris <rich.harris@vercel.com>
* fix: skip effects inside dynamic component that is about to be destroyed
When a dynamic component was updated to a different instance and its props were updated at the same time, effects inside the component were still called with the already-changed props.
The fix is to mark the branch as skipped to never got to those effects.
Fixes#16387
* undo accidental commit
* fix: wrap `abort` in `without_reactive_context`
* only call without_reactive_context when the controller exists
---------
Co-authored-by: Rich Harris <rich.harris@vercel.com>
* Update for TS7
I am testing Typescript 7's JS support, which I've largely rewritten
during the switch to Go. That means Javascript code will need to change
much more than Typescript code. Fortunately, most of the changes are for
the better: Javascript semantics are now nearly identical to Typescript
semantics. It's much stricter and no longer has some persistent bugs
that arose from shady JS handling I wrote years ago.
This PR changes Svelte so that it compiles both with TS5.* and TS7,
which means that occasionally there are duplicative or non-obvious
changes. I'll annotate the interesting changes to explain why I made them.
Because TS7 is quite a way off, I don't know whether you'll want to take
this PR. Most of the changes are for the better, because they're due to
stricter TS-aligned checking. But some are neutral and there is the
previously mentioned duplication in a few places.
* add changeset
* revert scribbles mistakenly added
* pnpm run format
* revert mistaken edit
* return to function declarations w/type tag
* add job to ci.yml
* skipLibCheck for now
* no need for a changeset here
---------
Co-authored-by: Rich Harris <rich.harris@vercel.com>
Co-authored-by: Rich Harris <hello@rich-harris.dev>
* fix: emit `snippet_invalid_export` instead of `undefined_export` for exported snippets
* try this
* this should work
* remove some junk (this never did anything)
* add test
---------
Co-authored-by: Rich Harris <rich.harris@vercel.com>
* Add missing hyphen in "server-side rendering"
* Fix incorrect use of "as such"
* regenerate types
---------
Co-authored-by: Chew Tee Ming <chew.tee.ming@nindatech.com>
* Fix double event processing (#16522)
Firefox seems to garbage collect event objects during event propagation if no global reference to the event object is kept.
That discards the __root marker set on the event object to early,
leading to duplicate processing.
* minor tweaks
---------
Co-authored-by: ComputerGuy <63362464+Ocean-OS@users.noreply.github.com>
* use sh instead of bash in source blocks
no bash-specific functionality is used
* regenerate
---------
Co-authored-by: Rich Harris <rich.harris@vercel.com>
* fix: throw on duplicate class field declarations
* doh
* handle assignment in constructor
* fix
* apply suggestion from review
* fix
* fix failing test
mini-cleanup post #16487 - we don't need to do the work of scheduling an effect that's already dirty which means it already scheduled its root effect to run
* don't mark_reactions inside decrement, it can cause infinite loops
* revert mark_reactions changes
* preserve DIRTY/MAYBE_DIRTY status of deferred effects
* changeset
* tweak
* - add state changes resulting from an $effect to a separate new batch
- schedule rerunning effects based on the sources that are dirty, not just rerunning them all blindly (excempting async effects which will have run by that time already)
* test
* better fix
* tests
* this fixes the last test somehow
* fix#16477
* typo
* copy over changeset from #16477
* copy over changeset from #16464
* changeset
* dedupe
* move flushing_sync check inside Batch.ensure
* unused
* flushing_sync -> is_flushing_sync
* remove flush_effects method
* dedupe declaration
* tweak
* tweak
* update comment — it _does_ feel slightly wrong, but no wronger than the rest of this cursed function
---------
Co-authored-by: Simon Holthausen <simon.holthausen@vercel.com>
Increment the version to ensure any dependent deriveds are marked dirty when the subscription is picked up again later. If we didn't do this then the comparison of write versions would determine that the derived has a later version than the subscriber, and it would not be re-run.
Fixes#16311Fixes#15888
The prior logic was wrong because it reexecuted when something was clean, but we want to when it's not. The remaining fix was to also check the reactions: If an effect is destroyed and it was the last reaction of a derived then the derived is set to `MAYBE_DIRTY`. We therefore also need to check if the derived still has anyone listening to it, and only then reexecute it.
Fixes#16363
Since #9070 it's prefered to use the "enhance svelte elements" way of doing things. That's so long ago that we can now remove the other option from the docs, and eventually remove the backwards compatibility that still allows the old way of doing things.
Fixes#16134
* Add a warning when the misuse of `reset` in an `error:boundary` causes an error to be thrown when flushing the boundary content
* Prevent uncaught errors to make test fails when they are expected and are fired during template effects flush
* reset should just be a noop after the first call
* correctly handle errors inside boundary during reset
* handle errors in the correct boundary
---------
Co-authored-by: Rich Harris <rich.harris@vercel.com>
#16268 introduced a slight regression where the state is not reset completely upon compilation. It did reset warnings but not other state, which meant if file A succeeds but file B fails in the parsing state (before the state was reset for real) it would get wrong filename info. This fixes it by setting the filename at the very beginning.
Rich Harris
github-actions[bot]
Simon H
Alejandro Torres
Paolo Ricciuti
David
ComputerGuy
Elliott Johnson
Hannes Rüger
7nik
Yuki Ishii
Jungzl
Jakub Kolesár
LeeWxx
Aaron Ajose
Brittany Harris
dai
Gurpreet Atwal
hariharan
Edoardo Cavazza
Harsh Mandan
shamokit
Ben McCann
Nathan Shively-Sanders
Isak Regnström
Anthony
Elliot Bentley
Laszlo Korte
fkobi
Bladesheng
Hyunbin Seo
Ray Thurn Void
jjones315