The original intent was for typing generation tools like svelte2tsx to use the virtual constructor instead, but it does more harm than good. Most notably, the TS error message when props are wrong becomes completely useless. It's better for tooling to silence the deprecation message and switch over to a new virtual constructor once we remove the deprecated one.
#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>
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>
* chore: make DOM operations lazyily init
* cleanup types
* cleanup types
* cleanup types
* Update packages/svelte/src/internal/client/operations.js
Co-authored-by: Simon H <5968653+dummdidumm@users.noreply.github.com>
* single line annotations
* remove unnecessary coercion
* group statements by type
---------
Co-authored-by: Simon H <5968653+dummdidumm@users.noreply.github.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>
no two-way binding because setting it involves a `DataTransfer` workaround, so it's not really officially supported that way - if you need that, you shouldn't use that binding probably. This matches the behavior in Svelte 4.
Co-authored-by: Rich Harris <rich.harris@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
* lets see if this works
* fix versions
* sigh
* debugging ci is sooo fun
* oh wow
* fix stuff, changelog, add back readme
* appease prettier
* format stuff
related to issue #9088
it doesn't solve the main problem of dependencies getting invalidated whenever value of a variable gets changed.
but it fixes the behavior difference between the code with and without comments
fixes#9092
---------
Co-authored-by: gtmnayan <50981692+gtm-nayan@users.noreply.github.com>
Co-authored-by: Simon H <5968653+dummdidumm@users.noreply.github.com>
fixes#9185.
I narrowed down the issue to the bug surfacing when we use object properties to update style attributes and directives. This fix removes the size check (because a single object will be of size 1 but can affect n attributes/directives via its properties).
In addition, the order of the OR is switched as the earlier condition has some reactive assignments which are not run in the current order when style_changed_var is truthy.
---------
Co-authored-by: github-actions[bot] <github-actions[bot]@users.noreply.github.com>
Co-authored-by: Simon H <5968653+dummdidumm@users.noreply.github.com>
Make IntrinsicElements extends SvelteHTMLElements
so it can be extend with declare module
converting to module so it can extend without needing another type alias
---------
Co-authored-by: Simon H <5968653+dummdidumm@users.noreply.github.com>
Fixes#2943
The issue linked above invalidates a variable that is not defined in <script> but instead is defined as a let: variable, which does not make sense to invalidate, therefore, this PR exists.
In dev mode, Svelte creates a setter to throw an error noting that you can't set that readonly prop, which resulted in the accessor getting applied twice to the custom element wrapper, causing an error
fixes#8971
This should help everyone who has special needs and use cases around custom elements. Since Svelte components are wrapped and only run on connectedCallback, it makes sense to expose the custom element class for modification before that.
- fixes#8954 / closes#8955 - use extend to attach the function manually and save possible values to a prop
- closes#8473 / closes#4168 - use extend to set the proper static attribute and then call attachInternals in the constructor
- closes#8472 - use extend to attach anything custom you need
- closes#3091 - pass `this` to a prop of your choice and use it inside your component
- add some doc for #8987
When the HTML that is coming from raw html is invalid, the browser reshuffles things. Bail in that case to try to repair more often. Should help with https://github.com/withastro/astro/issues/7557
html tags that could be optimized to use innerHTML in mount ignored any hydration code, which leads to everything getting unmounted and mounted again. This takes the non-optimized path for hydration, too.
fixes https://github.com/sveltejs/kit/issues/10245
fixes#8860
This contains a small but unfortunately unavoidable breaking change: If you used `never` to type that the second parameter of `createEventDispatcher` shouldn't be set or that the action accepts no parameters (which the docs recommended for a short time), then you need to change that to `null` and `undefined` respectively
---------
Co-authored-by: github-actions[bot] <github-actions[bot]@users.noreply.github.com>
Co-authored-by: Simon H <5968653+dummdidumm@users.noreply.github.com>
We need to use a different method for getting the meta info because `locate` is used to help construct the source map that references the preprocessed Svelte file. If we would now add source maps to that `locate` function it would go the the original source directly which means skipping potentially intermediate source maps which we would need in other situations. Sadly we can't map the character offset because for that we would need to the original source contents which we don't have in this context.
fixes#8360closes#8362
Can be opt out by setting discloseVersion to false
---------
Co-authored-by: Simon H <5968653+dummdidumm@users.noreply.github.com>
Co-authored-by: Ben McCann <322311+benmccann@users.noreply.github.com>
technically a breaking change because someone with a mutation observer could rely on the order of operations
---------
Co-authored-by: Simon H <5968653+dummdidumm@users.noreply.github.com>
also add some legacy import paths which should be changed in usage sites ASAP
---------
Co-authored-by: Simon H <5968653+dummdidumm@users.noreply.github.com>
Co-authored-by: Simon Holthausen <simon.holthausen@vercel.com>
Simon Holthausen
github-actions[bot]
Simon H
Dominic Gannaway
Yuichiro Yamashita
Rich Harris
Tee Ming
Paolo Ricciuti
gtmnayan
Kelvin Soh
Ben McCann
Sina Salahshour
vhladko
Teo
paulreece
Lyu, Wei-Da
hackape
Nguyen Tran
L