- use interface instead of type (interface is less forgiving)
- rename SourceSignal to Source, etc
- make Derived extend Source, rather than deriveds sharing a type with effects, since they have much more in common with each other
- have a Value type which is a union of Source and Derived, and a Reaction type which is a union of Derived and Effect
- avoid using the Signal type (which is a union of all three) unless necessary
* move some code
* split computations.js into deriveds.js and effects.js
* move reactivity types into separate .d.ts file
* move some signal code
---------
Co-authored-by: Rich Harris <rich.harris@vercel.com>
* fix: treat snippets like normal components when inferring namespace
* n
* simplify
* better desc
* slight adjustment
* feedback
Co-authored-by: Simon H <5968653+dummdidumm@users.noreply.github.com>
* feedback
Co-authored-by: Simon H <5968653+dummdidumm@users.noreply.github.com>
* skip html tag
* test
* changeset name
* cleanup
---------
Co-authored-by: Simon H <5968653+dummdidumm@users.noreply.github.com>
Checking for expected or unexpected props in 2024 is the job TypeScript, and runtime validation has resulted in false positives previously - therefore leave this out in Svelte 5.
closes#10672
- widen ownership when getContext is called, part of #10649
- widen ownership when assigning one proxy to another
- skip first 4 stack trace entries and bail if first after that is not a module, hinting at a mutation encapsulated in a .svelte.js file; part of #10649
* feat: improve ssr html mismatch validation
* update types
* Update packages/svelte/src/internal/server/index.js
Co-authored-by: Simon H <5968653+dummdidumm@users.noreply.github.com>
* Update packages/svelte/src/compiler/validate-options.js
Co-authored-by: Simon H <5968653+dummdidumm@users.noreply.github.com>
* feedback
---------
Co-authored-by: Simon H <5968653+dummdidumm@users.noreply.github.com>
It's a warning because even when typing it out and knowing what to do you'll always be in a state where the validation kicks in, and it would be too distracting to always see a compiler error during that short time frame.
closes#10374
People could've done bind:this and called instance methods on the instance - a rare case, but not impossible. This shims $set and $on when in legacy compat mode. $destroy is never shimmed because you shouldn't manually destroy a component, ever, and there's no way to make that work in the new world.
closes#10420
* fix: correctly handle proxied signal writes before reads
* managed
* Update packages/svelte/src/internal/client/proxy.js
Co-authored-by: Simon H <5968653+dummdidumm@users.noreply.github.com>
---------
Co-authored-by: Simon H <5968653+dummdidumm@users.noreply.github.com>
fixes#10511
We used the animation frame dance previously where the `$effect` timing was slightly different and did not wait on its children before rerunning. With that changed now everything false into place nicely.
---------
Co-authored-by: Rich Harris <rich.harris@vercel.com>
Co-authored-by: Simon Holthausen <simon.holthausen@vercel.com>
The previous `bind_this` implementation had flaws around timing and didn't work correctly when the binding wasn't `$state` in runes mode.
This PR reimplements `bind_this` by aligning it more with how Svelte 4 bindings work, namely reacting to each block context variables updates properly and introducing a mechanism to get the previous binding destination using said variables.
* fix: improve $inspect handling of derived objects
* Update packages/svelte/src/internal/client/runtime.js
Co-authored-by: Simon H <5968653+dummdidumm@users.noreply.github.com>
---------
Co-authored-by: Simon H <5968653+dummdidumm@users.noreply.github.com>
* move effect code
* move source code
* derived
* move more stuff
* simplify
* init is a daft name for something that runs repeatedly
* remove unneeded effect variant
* chore: reshuffle part2 (#10588)
* extract store
* move prop/init to render
* introduce constants file and move all constants there to reduce cyclic dependencies
* move reactive_import to render
* move bubble_event into render, move (raf) task handing into its own file
* expect-error did hide that one
---------
Co-authored-by: Rich Harris <rich.harris@vercel.com>
Co-authored-by: Simon H <5968653+dummdidumm@users.noreply.github.com>
* fix: remove memory leak
* changeset
* mutable_source is only used in non-runes mode, no need to check here
* simplify
* oops
---------
Co-authored-by: Rich Harris <rich.harris@vercel.com>
* fix: remove readonly validation
The readonly dev time validation results in false-negative object equality checks: The original object is proxified and thus comparisons could be between the unproxified and proxified version, which will always return false. Fixes#10372
There's also the problem that an object could be passed into a component but then passed upwards into shared state. At that point, it should no longer be readonly, but it's not possible to statically analyze these points. Fixes#10372
Lastly, the each block logic mutates an internal array and that also throws errors with the readonly logic. Fixes#10037
* reinstate tests
* track ownership of state and mutations
* working?
* remove old changeset
* tidy
* error
* simplify
* fix
* fix
* fix
* tidy
* make it a warning
* rename test
* remove unused test
* update tests
* slap ts-expect-error everywhere, because its too finicky otherwise
* oops
* oh no the hall monitor is here
* only call add_owner in dev
* only owners can transfer ownership
* simplify
* fixes
* tidy up
* fix type error
* while we're at it
* rename file
* rename functions
* add some comments
* move ownership checking logic
* ugh eslint
* more detailed message
* only add filename in dev
* comment
* update tests
* move more code
* undo change to sourcemap tests
* allow proxy to have multiple owners
* use SignalDebug
* i was doing this all wrong
* tidy up
* implement inheritance
* fix
* tidy up
* update filename stuff
* changeset
---------
Co-authored-by: Simon Holthausen <simon.holthausen@vercel.com>
Co-authored-by: Rich Harris <rich.harris@vercel.com>
Adjusts the escaping mechanism done for server compilation. For template literals it's now only applied when explicitly told, which is the case for generated literals from the html template. Fixes a bug where a template literal string inside the `@html` tag was wrongfully escaped (https://github.com/sveltejs/svelte/issues/10359#issuecomment-1949678228)
In Svelte 4, scripts inside `@html` were not executed when it was created client-side. This is because `innerHTML = ..` which was used under the hood does not execute scripts due to security reasons. This adjusts the code so the same is true for Svelte 5.
Ensure update methods of actions and reactive statements work with fine-grained `$state` by deep-reading them. This is necessary because mutations to `$state` objects don't notify listeners to only the object as a whole, it only notifies the listeners of the property that changed.
fixes#10460
fixes https://github.com/simeydotme/svelte-range-slider-pips/issues/130
* chore: ignore test directories left over from other branches
* tighten up
* one comment will do
---------
Co-authored-by: Rich Harris <rich.harris@vercel.com>
* failing test for :is(...)
* simplify
* trim
* factor out truncate function
* pass stylesheet around
* recurse into :is and :where
* fix types
* fix types
* almost there
* gah so close
* tada
* changeset
* simplify
* feat: nested CSS support (#10491)
* parse nested CSS
* tests
* track parent rules
* some progress
* switch it up
* pruning
* works
* changeset
* lint
* error early on invalid nesting selector
* tidy
* note to self
* fix some specificity stuff
* failing test
* note to self
* fix: correctly scope CSS selectors with descendant combinators (#10502)
* fix traversal, but break some other stuff
* man this is fucken hard
* fixes
* getting closer
* be conservative for now
* tidy
* invert
* invert
* simplify
* switch
* for now
* progress
* i think it works?
* fix
* tidy up
* revert some stuff
* remove some junk
* handle weird cases
* update
* tweak
* shrink
* changeset
---------
Co-authored-by: Rich Harris <rich.harris@vercel.com>
---------
Co-authored-by: Rich Harris <rich.harris@vercel.com>
* Update playgrounds/sandbox/run.js
* changeset
---------
Co-authored-by: Rich Harris <rich.harris@vercel.com>
Add source map merging for preprocessors and get tests passing.
- fixed some issues around the `sources` array where they weren't calculated relative to the input correctly
- adjusted some CSS tests because due to our own CSS parser the AST is less granular, so there are less mappings now. Don't think this is a problem, but worth thinking about
- removed enableSourcemap but only log a warning, the reason this was introduced was to mitigate a bug in Vite which occured when having the source map inlined into the SSR'd output. Since SSR doesn't contain inlined CSS anymore (and if it did, we would omit the source map) the reason for which it was introduced no longer exists
- files without js mapping in it have no source mappings yet (originally added in Svelte 4 for #6092)
* rename $derived.call to $derived.by
* more replacements, plus a compiler error
* regenerate types
---------
Co-authored-by: Rich Harris <rich.harris@vercel.com>
Relax the runtime error to only throw when you're passing a binding with an undefined value. This makes it possible to provide components in a way that can be used more flexibly while keeping the error to guard against the case that we want to avoid: a default value propagation up.
* fix: handle sole empty expression tags
When there's only a single expression tag and its value evaluates to the empty string, special handling is needed to create and insert a text node
fixes#10426
* fix
* need this, too
* Update packages/svelte/src/internal/client/operations.js
---------
Co-authored-by: Rich Harris <richard.a.harris@gmail.com>
Closes#9420.
This PR creates an $effect.pre (before beforeUpdate and an $effect (for afterUpdate) and, inside those, listen for all locally declared signals plus reactive props. This does mean that we need to link the locally declared signals to the component context (the reverse of the current behaviour, wherein we link the component context to locally declared signals).
fixes#10296
Also make sure to use the server export conditions when resolving Svelte imports from inside the server compiler output
---------
Co-authored-by: Rich Harris <rich.harris@vercel.com>
Co-authored-by: Simon Holthausen <simon.holthausen@vercel.com>
- Previously, any each block parents where used as keys for the sub group. That's wrong, it should only be using each blocks whos declarations contribute to the `bind:group` expression. Fixes#10345
- Only the left-most identifier of the expression was used to determine the binding group. This is wrong, all identifiers within the expression need to be taken into account. Fixes#9947
* fix: don't reuse proxies when state symbol refers to stale value
When somebody copies over the state symbol property onto a new object (you can retrieve the symbol by using `Reflect.ownKeys(...)`), it was wrongfully assumed that it always relates to the current value. This PR adds an additional check that this is actually the case.
This also adds a dev time warning when an object is frozen but contains a state property, which hints at a bug in user land.
fixes#10316
* lint
* rename
* remove warning
* update test
* changeset
---------
Co-authored-by: Rich Harris <rich.harris@vercel.com>
Co-authored-by: Rich Harris <richard.a.harris@gmail.com>
* strip typescript assertions before analysis
* add test
* changeset
* move ts handling from transform to before analysis
* format
* types
* remove unwrap_ts_expression
---------
Co-authored-by: Simon Holthausen <simon.holthausen@vercel.com>
We're now on Svelte v5.0.0-next.42 which made one test match the 42 of the next version instead of the number in the code, resulting in that test suddenly failing. This fixes the regex to replace that number with x's instead
* give this another try
* fix: lint
* fix: Forgot to save
* feat: it works boiiii
* look, ok, it did work, i just needed to update the snapshots
* bruh
* changeset
* feat: ok I think the client snippet block finally works
* feat: current tests pass; I'm sure I'm missing stuff for new things
* fix: snapshot
* feat: I think non-destructured rest should work now?
* chore: duplicated computation
* feat: Tests (passing and failing
* feat: it's... alive?
* chore: Clean up my messes
* chore: devtime stuff
* fix: fmt
* chore: see if this fixes repl
* chore: make naming more offensive
* fix: Don't throw on missing keys, return undefined as it usually would
* Update packages/svelte/src/compiler/phases/1-parse/state/tag.js
Co-authored-by: Simon H <5968653+dummdidumm@users.noreply.github.com>
* Update packages/svelte/src/compiler/phases/1-parse/state/tag.js
Co-authored-by: Simon H <5968653+dummdidumm@users.noreply.github.com>
* fix: Hopefully default param values now work
* dumb
* types
* feat: Test it
* fix: Turns out javascript parameters are optional
* feat: The Final Solution
* document function
* feat: Better bracket matching, unit tests
* feat: exclude test files from publish
* feat: More unit tests
* feat: Use more efficient parsing for @const
* Update .changeset/curvy-cups-cough.md
Co-authored-by: Simon H <5968653+dummdidumm@users.noreply.github.com>
* Update packages/svelte/package.json
Co-authored-by: Simon H <5968653+dummdidumm@users.noreply.github.com>
* Update packages/svelte/src/compiler/phases/1-parse/utils/bracket.js
Co-authored-by: Simon H <5968653+dummdidumm@users.noreply.github.com>
* fix: changesets
* chore: additional comments
* fix: kill foreach
* fix: foreach again
* feat: Docs
* Revert "fix: kill foreach"
This reverts commit 9a688cc543.
* fix: My own stupidity
* fix: style
* fix - maybe
* Update sites/svelte-5-preview/src/routes/docs/content/01-api/03-snippets.md
* Update tag.js
Co-authored-by: Rich Harris <richard.a.harris@gmail.com>
* Update .changeset/curvy-cups-cough.md
Co-authored-by: Simon H <5968653+dummdidumm@users.noreply.github.com>
* chore: Remove rest params
* Delete .changeset/eighty-rivers-wash.md
* fix: Honestly idk why it was broken but it's fixed now
* fix: var name lol
* fix: typegen
* fix: idk
* fix: It looks like a bunch of unformatted shit came in through main?? idk
* Revert "fix: It looks like a bunch of unformatted shit came in through main?? idk"
This reverts commit ab851d5627.
* fix: format again
* this is getting ridiculous
* Update tag.js
Co-authored-by: Rich Harris <richard.a.harris@gmail.com>
* fix errors
* simplify a bit
* use read_context
* use read_context for const as well
* remove unused code
* unused import
* unused export
* remove spread args. sorry elliott
* tidy up SnippetBlock interface
* fix test
* simplify
* tweak
* revert example, so that it matches the surrounding text
* move PropsWithChildren back to public.d.ts
* update typing docs, so that it flows from previous example
* temporarily revert const parsing changes, to get prettier working again (???)
* oops
---------
Co-authored-by: Simon H <5968653+dummdidumm@users.noreply.github.com>
Co-authored-by: Rich Harris <richard.a.harris@gmail.com>
Co-authored-by: Rich Harris <rich.harris@vercel.com>
Infer namespace from parents where possible, and do a runtime-best-effort where it's not statically known
fixes#9645
---------
Co-authored-by: Simon Holthausen <simon.holthausen@vercel.com>
If you had `on:` directives listening to the same name (through multiple on:click on the same element or indirectly through multiple `<svelte:window>` elements with event listeners of the same name) there was a bug of delegation firing too often. This PR fixes that by tweaking the "should I continue with the given path index" logic.
fixes#10271
Someone could programmatically add a class to an element and Svelte doesn't see it, so having global be part of a modifier is necessary so that Svelte doesn't mark it as unused
fixes#10210
If a list is passed to a component and an item of that list is passed as a slot prop back up, then mutating a property of that item did not result in a rerun. The reason was that derived is using object identity equality, resulting in the value not being updated. To fix it, we use safe-equals in this situations instead.
* fix: simplify event delegation logic
Only delegate event attributes, and don't take into account bindings/actions while determining that. Never delegate `on:` directives. This simplifies the logic and makes it easier to explain, while avoiding any accidental breaking changes when upgrading from Svelte 4 to 5 without changing code.
Fixes#10165
Related to #9714
* update types
Add recursive check for logic blocks, ignore things such as ConstTags and Comments
closes#10025
---------
Co-authored-by: Simon Holthausen <simon.holthausen@vercel.com>
It's an unnecessary restruction because it can be worked around (hide it behind a getter/setter), already works for bind:x and prevents valid use cases
- the objects could contain getters with reactive values, so we play it safe and assume they're always reactive - fixes#10065
- isolate spreads with call expression similar to how we do it with other effects -fixes #10013
When doing `$inspect({ x, y })`, both `x` and `y` are now unstated if they are signals, compared to before where `unstate` was only called on the top level object, leaving the proxies in place which results in a worse debugging experience.
Also improved typings which makes it easier to find related code paths.
closes#9969
this also fixes the following along the way:
the + in nth-of-type(+xn-b) would be parsed as a combinator.
invalid cases like these are not allowed anymore:
b(+/-)b
-ax
-ax-b
-b
* fix: prevent infinite loops stemming from invalidation method
The logic was flawed: the captured signals where always added to the previous captured no matter what, which meant a) memory leak b) that when another one runs afterwards, it will falsely contain the signals from the previous run
fixes#9788
* fix lint
- Expect the thing that's checked to be wrapped with the proxy already, so that we can just check for the state symbol
- Make error message more descriptive
* use proxy instead of signal in createRoot
* DRY
* remove for now
* lint
* chore: use proxies instead of signals for spread/rest props (#9801)
* use proxies instead of signals for spread/rest
* fix some spread attribute stuff
* remove is_signal calls
* simplify some more
* more
* remove some unnecessary unwrapping
* another
* simplify
* simplify
* simplify
* remove another MaybeSignal
* more
* remove more unwraps
* code-golf, docs
---------
Co-authored-by: Rich Harris <rich.harris@vercel.com>
Co-authored-by: Simon Holthausen <simon.holthausen@vercel.com>
* add missing jsdoc annotation
---------
Co-authored-by: Rich Harris <rich.harris@vercel.com>
Co-authored-by: Simon Holthausen <simon.holthausen@vercel.com>
* WIP
* update tests
* only make readonly in runes mode
* remove this for now
* changeset
* ugh
* add reassignment test
* tweak message
---------
Co-authored-by: Rich Harris <rich.harris@vercel.com>
* feat: add unstate utility function
* Update packages/svelte/src/internal/client/proxy/proxy.js
Co-authored-by: Simon H <5968653+dummdidumm@users.noreply.github.com>
* update docs
* add class support
* oops
* lint
* fix docs
* remove symbol and class support
---------
Co-authored-by: Simon H <5968653+dummdidumm@users.noreply.github.com>
Co-authored-by: Rich Harris <richard.a.harris@gmail.com>
* magic objects
* read length eagerly — triggers reconciliation
* nested magic
* tests
* more tests
* fix array memory leak
* skipped, partially passing array test
* Fix each
revert bad changes
* more 1337
* eliminate closures
* maybe this is unnecessary?
* only create sources for own properties
* more
* rename stuff
* shuffle things around
* emit $.proxy
* remove proxy helper from public API
* only create sources for writable properties
* update test
* get tests passing
* fix
* remove state-not-mutated warning, which is no longer valid
* track reassignments separately from mutations
* update test - effects no longer fire on mutations unnecessarily
* move util into utils file
* move each logic into its own module; breathe sigh of relief
* tweaks
* more tweaks
* improve runtime
* fix mistake
* ensure we proxy when assigning to state
* fix test
* handle frozen
* create sources when reading proxy properties inside deriveds
* only mutate in legacy mode
* add immutable to transform state
* remove unused second argument to derived
* remove unused second argument to source, and runtime immutable option to createRoot (not sure what that was doing there?)
* oops, backwards
* dedicated binding.kind for legacy reactive imports
* avoid using prop_source in more cases. bit hacky, could be tidier, but it works
* distinguish between source and mutable_source
* remove immutable option from mount
* remove unused apparatus around immutable option
* deprecate immutable
* fix
* tweak
* better default value handling
* remove unnecessary exports
* whitespace is free
* remove obsolete symbol from proxy
* cleanup ts
* improve runtime perf by removing version reads in has()
* add missing proxy call
* tune perf of has() more
* ensure we only create sources in effect_active()
* fix proxy of computed fields
* simplify and fix issue with indexed each blocks
* fix compiler errors around exported state
* only create source for state that is reassigned
* temporary fix, we should come up with something better than this
* readonly props
* fix test
* add test for bind:
* make readonly dev-only
* docs
* forbid setPrototypeOf
* lol whoops
---------
Co-authored-by: Rich Harris <rich.harris@vercel.com>
Co-authored-by: Dominic Gannaway <dg@domgan.com>
* fix: handle ts expressions when dealing with runes
related to #9639
* docs, more tests
* simplify
---------
Co-authored-by: Rich Harris <rich.harris@vercel.com>
- doesn't add spread_dynamic_element_attributes when there are no attributes — Unnecessary spread_dynamic_element_attributes call #9646
- skips the child render function altogether if there is nothing to do
If someone has an existing SvelteComponent type definition and using the slot generic to type the default slot, automatically add that slot to the prop type as "children"
#9589 - add comment and space as reusable templates to save a few bytes. We can definitely take this idea further, but this is a base to iterate from.
---------
Co-authored-by: Rich Harris <rich.harris@vercel.com>
Co-authored-by: Simon Holthausen <simon.holthausen@vercel.com>
* failing test
* fix undefined index bug
* Update packages/svelte/src/compiler/phases/1-parse/read/context.js
Co-authored-by: Simon H <5968653+dummdidumm@users.noreply.github.com>
---------
Co-authored-by: Rich Harris <rich.harris@vercel.com>
Co-authored-by: Simon H <5968653+dummdidumm@users.noreply.github.com>
* set expression end to typeAnnotation end where appropriate
* fix
* woops
* lint
---------
Co-authored-by: Rich Harris <rich.harris@vercel.com>
Co-authored-by: Simon H <5968653+dummdidumm@users.noreply.github.com>
* shavings
* dro unused max safe int
* get rid of array_ref and object_ref
* extract STATUS_MASK
* fix test
* revert is_controlled change
* more run_all
Related to sveltejs/language-tools#2204 / sveltejs/language-tools#2039
The Svelte 5 version of #9486 and #9498
---------
Co-authored-by: Rich Harris <richard.a.harris@gmail.com>
Co-authored-by: Rich Harris <rich.harris@vercel.com>
* fix: address bug in before/after update
fix: address bug in before/after update
* Add changeset
* use every instead of filter - more explicit and enables early-exit from the loop
* Update logic and comment
---------
Co-authored-by: Rich Harris <rich.harris@vercel.com>
- add event delegation to spread_attributes
- add event attributes to spread
- don't delegate when bindings/actions on the same element in order to preserve backwards compatibility of ordering
- don't hoist identifiers when one of them is used in an event that is not delegateable
---------
Co-authored-by: Simon Holthausen <simon.holthausen@vercel.com>
I came to the conclusion that when we're making up arbitrary types, we might as well keep the old class. That way:
- one less thing to worry about (language tools and other tooling basically can continue to spit out SvelteComponent )
- we can more clearly mark $set , the constructor etc as being deprecated and no longer functioning unless you use that legacy compatibility mode
- much more ergonomic to type for the user:
- const someInstance: SvelteComponent<..> instead of const someInstance: ReturnType<typeof Component<..>>
- If you're using generics, you can do export class MyComponent<T> extends SvelteComponent<{ prop: T }> {} instead of having to type out the whole function in a way that I'm not even sure how to do with generics
Simon H
Dominic Gannaway
Rich Harris
Ahmad
Daniel Imfeld
gtmnayan
Simon Holthausen
S. Elliott Johnson
Karol
Rogério Luiz Aques de Amorim
Ben McCann
Nguyen Tran
Jeremy Deutsch
Holger Dal Mogensen
Tom
Yuichiro Yamashita
Theodore Brown
Tee Ming
Paolo Ricciuti