CSS View Transitions Module Level 2

Editor’s Draft,

More details about this document
This version:
https://drafts.csswg.org/css-view-transitions-2/
Issue Tracking:
CSSWG Issues Repository
Inline In Spec
Editors:
Noam Rosenthal (Google)
Khushal Sagar (Google)
Vladimir Levin (Google)
Tab Atkins-Bittner (Google)
Suggest an Edit for this Spec:
GitHub Editor

Abstract

This module defines how the View Transition API works with cross-document navigations.

CSS is a language for describing the rendering of structured documents (such as HTML and XML) on screen, on paper, etc.

Status of this document

This is a public copy of the editors’ draft. It is provided for discussion only and may change at any moment. Its publication here does not imply endorsement of its contents by W3C. Don’t cite this document other than as work in progress.

Please send feedback by filing issues in GitHub (preferred), including the spec code “css-view-transitions-2” in the title, like this: “[css-view-transitions-2] …summary of comment…”. All issues and comments are archived. Alternately, feedback can be sent to the (archived) public mailing list www-style@w3.org.

This document is governed by the 03 November 2023 W3C Process Document.

1. Introduction

This section is non-normative.

View Transitions, as specified in [css-view-transitions-1], is a feature that allows developers to create animated transitions between visual states of the document.

Level 2 extends that specification, by adding the necessary API and lifecycle to enable transitions across a same-origin cross-document navigation, as well as a few additions that make it easier to author pages with richer view transitions.

Level 2 defines the following:

2. Cross-document view transitions

2.1. Overview

This section is non-normative.

2.1.1. Activation

With same-document view transitions, the author starts a view transition using JavaScript, by calling startViewTransition. In cross-document view transition, what triggers a view transition is a navigation between two documents, as long as the following conditions are met:

See the lifecycle section for more details.

2.1.2. Waiting for the new state to stabilize

In same-document view transitions, the author can indicate when the new state of the transition is in a stable state by using the callback passed to startViewTransition. Since cross-document view transitions are declarative, there is no such explicit promise. Instead, the user agent relies on the render-blocking mechanism to decide when the document has reached a stable state. In this way, the author can use the the blocking attribute, to delay the transition until:

Note: overusing the render-blocking mechanism could make it so that the old state remains frozen for a long time, resulting in a jarring user experience. To avoid this, it’s advised to ensure that the render-blocking elements are available in a timely manner.

In this example, the last frame of the old document will be shown, and the animation will be delayed, until all the following conditions are met:

<!DOCTYPE html>
<html>
  <head>
    < !-- This will be render-blocking by default -->
    <link rel="stylesheet" href="style.css">

    < !-- Since this script fixes up the layout, marking it as render blocking will
        ensure it's run before the view transition is activated -->
    <script async href="fixup.js" blocking="render"></script>

    < !-- Wait until the main-article element is seen and fully parsed before
        activating the transition -->
    <link rel="expect" href="#main-article" blocking="render">
  </head>
  <body>
    <header>...</header>
    <main>
      <article id="main-article">...</article>
    </main>
    <article id="secondary-article">...</article>
  </body>
</html>

2.1.3. Customization

The ViewTransition object enables customizing the transition in script. Same-document view transitions use a single ViewTransition object returned from the startViewTransition call for the entire lifecycle. Cross-document view transitions have two ViewTransition objects, one in the old document and one in the new document.
2.1.3.1. Handling the view transition in the old document

The pageswap event is fired at the last moment before a document is about to be unloaded and swapped by another document. It can be used to find out whether a view transition is about to take place, customize it using types, make last minute changes to the captured elements, or skip it if necessary. The PageSwapEvent interface has a viewTransition object, which would be non-null when the navigation is eligible to a view transition, and a activation object, providing handy information about the navigation, like the URL after redirects. The transition’s finished promise can be used for cleaning up after the transition, in case the document is later restored from BFCache.

2.1.3.2. Handling the view transition in the new document

The pagereveal event is fired right before presenting the first frame of the new document. It can be used to find out if the view transition is still valid, by querying the viewTransition attribute. Similar to a same-document view transition, the author can now select different types, make last minute changes to the captured elements, wait for the transition to be ready in order to animate it, or skip it altogether.

2.1.4. Lifecycle

This section is non-normative.

A successful cross-document view transition goes through the following phases:

  1. In the old Document:

    1. The user initiates a navigation, by clicking a link, submitting a form, pressing the browser’s back button, etc.

      Note: some navigations do not trigger a view-transition, e.g. typing a new address in the URL bar.

    2. When the new Document is ready to be activated, the pageswap event is fired.

    3. If the navigation is same origin, has no cross-origin redirects, and the old Document has opted in to cross-document view transitions, the event’s viewTransition attribute would be a ViewTransition object.

    4. The author can now customize the transition, e.g. by mutating its types, or skip it altogether.

    5. If the ViewTransition is not skipped, the state of the old document is captured.

    6. The navigation proceeds: the old Document is unloaded, and the new Document is now active.

  2. Then, in the new Document:

    1. When the new Document is ready for its first rendering opportunity, an event named pagereveal is fired on the new Document, with a viewTransition attribute.

    2. This ViewTransition's updateCallbackDone promise is already resolved, and its captured elements are populated from the old Document.

    3. This is another opportunity for the author customize the transition, e.g. by mutating its types, or skip it altogether.

    4. The state of the new document is captured as the "new" state of the transition.

    5. From this point forward, the transition continues in a similar fashion to a same-document transition, as per activate view transition.

2.2. Examples

To generate the same cross-fade as in the first example CSS View Transitions 1 § 1.6 Examples, but across documents, we don’t need JavaScript.

Instead, we opt in to triggering view-transitions on navigation in both page 1 and page 2:

// in both documents:
@view-transition {
  navigation: auto;
}

A link from page 1 to or from page 2 would generate a crossfade transition for example 1. To achieve the effect examples 2, 3 & 4, simply put the CSS for the pseudo-elements in both documents.

Note that the @view-transition rule can be used together with media queries. For example, this would only perform the transition when the screen width is greater than:
@view-transition {
  navigation: auto;
}

@media (max-width: 600px) {
  navigation: none;
}
To achieve the effect in example 5, we have to do several things:

In both pages:

@view-transition {
  navigation: auto;
}

In the old page:

addEventListener('click', event => {
  sessionStorage.setItem("lastClickX", event.clientX);
  sessionStorage.setItem("lastClickY", event.clientY);
});

In the new page:

// This would run both on initial load and on reactivation from BFCache.
addEventListener("pagereveal", async event => {
  if (!event.viewTransition)
    return;

  const x = sessionStorage.getItem("lastClickX") ?? innerWidth / 2;
  const y = sessionStorage.getItem("lastClickY") ?? innerHeight / 2;

  const endRadius = Math.hypot(
    Math.max(x, innerWidth - x),
    Math.max(y, innerHeight - y)
  );

  await event.viewTransition.ready;

  // Animate the new document's view
  document.documentElement.animate(
    {
      clipPath: [
        `circle(0 at ${x}px ${y}px)`,
        `circle(${endRadius}px at ${x}px ${y}px)`,
      ],
    },
    {
      duration: 500,
      easing: 'ease-in',
      pseudoElement: '::view-transition-new(root)'
    }
  );
})
To choose which elements are captured based on properties of the navigation, and whether certain images are loaded:

In the old page:

window.addEventListener("pageswap", event => {
  // For example, the page was hidden, or the navigation is cross-document.
  if (!event.viewTransition)
    return;

  // If you don't want view transition for back/forward navigations...
  if (event.activation.navigationType === "traverse") {
    event.viewTransition.skipTransition();
  }

  const newURL = new URL(event.activation.entry.url);
  if (newURL.pathname === "/details" && thumbnail.complete) {
    thumbnail.classList.add("transition-to-hero");

    // This will cleanup the state if the page is restored from BFCache.
    event.viewTransition.finished.then(() => {
      thumbnail.classList.remove("transition-to-hero");
    });
  }

});

In the new page:

window.addEventListener("pagereveal", event => {
  // For example, the page was hidden, the navigation is cross-document, or the transition was skipped in the old document.
  if (!event.viewTransition)
    return;

  const oldURL = new URL(navigation.activation.from.url);
  if (newURL.pathname === "/list") {
    event.viewTransition.types.add("coming-from-list");

    // This will show the thumbnail until the view transition is finished.
    if (!hero.complete) {
      setToThumbnail(hero);
      event.viewTransition.finished.then(() => {
        setToFullResolution(hero);
      })
    }
  }
});

2.3. Opting in to cross-document view transitions

2.3.1. The @view-transition rule

The @view-transition rule is used by a document to indicate that cross-document navigations should setup and activate a ViewTransition.

The @view-transition rule has the following syntax:

@view-transition {
  <declaration-list>
}

The @view-transition rule accepts the navigation and types descriptors.

Note: as per default behavior, the @view-transition rule can be nested inside a conditional group rule such as @media or @supports.

When the @view-transition rule changes for Document document, the user agent must update the opt-in state for outbound transitions given document.

Note: this needs to be cached in the boolean because the result needs to be read in parallel, when navigating.

2.3.2. The navigation descriptor

Name: navigation
For: @view-transition
Value: auto | none
Initial: none

The 'navigation' descriptor opts in to automatically starting a view transition when performing a navigation of a certain type. Must be present on both the old and new document.

none

There will be no transition.

auto

The transition will be enabled if the navigation is same-origin, without cross-origin redirects, and whose NavigationType is

Note: Navigations excluded from auto are for example, navigating via the URL address bar or clicking a bookmark, as well as any form of user or script initiated reload.

2.3.3. Accessing the @view-transition rule using CSSOM

The CSSRule interface is extended as follows:

partial interface CSSRule {
  const unsigned short VIEW_TRANSITION_RULE = 15;
};

The CSSViewTransitionRule represents a @view-transition rule.

enum ViewTransitionNavigation { "auto", "none" };

[Exposed=Window]
interface CSSViewTransitionRule : CSSRule {
  readonly attribute ViewTransitionNavigation navigation;
  [SameObject] readonly attribute FrozenArray<CSSOMString> types;
};

3. Selective view transitions

3.1. Overview

This section is non-normative.

For simple pages, with a single view transition, setting the view-transition-name property on participating elements should be sufficient. However, in more complex scenarios, the author might want to declare various view transitions, and only run one of them simultaneously. For example, sliding the whole page when clicking on the navigation bar, and sorting a list when one if its items is dragged.

To make sure each view transition only captures what it needs to, and different transitions don’t interfere with each other, this spec introduces the concept of active types, alongside the :active-view-transition and :active-view-transition-type() pseudo-classes.

:active-view-transition matches the document element when it has an active view transition, and :active-view-transition-type() matches the document element if the types in the selectors match the active view transition's active types.

The ViewTransition's active types are populated in one of the following ways:

  1. Passed as part of the arguments to startViewTransition(callbackOptions)

  2. Mutated at any time, using the transition’s types

  3. Declared for a cross-document view transition, using the types descriptor.

3.2. Examples

For example, the developer might start a transition in the following manner:
document.startViewTransition({update: updateTheDOMSomehow, types: ["slide-in", "reverse"]});

This will activate any of the following selectors:

:root:active-view-transition-type(slide-in) {}
:root:active-view-transition-type(reverse) {}
:root:active-view-transition-type(slide-in, reverse) {}
:root:active-view-transition-type(slide-in, something-else) {}
:root:active-view-transition {}

While starting a transition without providing transition types, would only activate ':active-view-transition'':

document.startViewTransition(updateTheDOMSomehow);
// or
document.startViewTransition({update: updateTheDOMSomehow});
/* This would be active */
:root { }
:root:active-view-transition {}

/* This would not be active */
:root:active-view-transition-type(slide-in) {}
:root:active-view-transition-type(any-type-at-all-except-star) {}

3.3. Selecting based on the active view transition

3.3.1. The :active-view-transition pseudo-class

The :active-view-transition pseudo-class applies to the root element of the document, if it has an active view transition.

The specificity of an :active-view-transition is one pseudo-class selector.

An :active-view-transition pseudo-class matches the document element when its node document has an non-null active view transition.

3.3.2. The :active-view-transition-type() pseudo-class

The :active-view-transition-type() pseudo-class applies to the root element of the document, if it has a matching active view transition. It has the following syntax definition:

:active-view-transition-type(<custom-ident>#)

The specificity of an :active-view-transition-type() is one pseudo-class selector.

An :active-view-transition-type() pseudo-class matches the document element when its node document has an non-null active view transition, whose active types contains at least one of the <custom-ident> arguments.

3.4. Changing the types of an ongoing view transition

The ViewTransition interface is extended as follows:

[Exposed=Window]
interface ViewTransitionTypeSet {
  setlike<DOMString>;
};

[Exposed=Window]
partial interface ViewTransition {
  attribute ViewTransitionTypeSet types;
};

The ViewTransitionTypeSet object represents a set of strings, without special semantics.

Note: a ViewTransitionTypeSet can contain strings that are invalid for :active-view-transition-type, e.g. strings that are not a <custom-ident>.

The types getter steps are to return this's active types.

3.5. Activating the transition type for cross-document view transitions

The types descriptor

Name: types
For: @view-transition
Value: none | <custom-ident>+
Initial: none

The 'types' descriptor sets the active types for the transition when capturing or performing the transition, equivalent to calling startViewTransition(callbackOptions) with that types.

Note: the types descriptor only applies to the Document in which it is defined. The author is responsible for using their chosen set of types in both documents.

4. Sharing styles between view transition pseudo-elements

4.1. Overview

This section is non-normative.

When styling multiple elements in the DOM in a similar way, it is common to use the class attribute: setting a name that’s shared across multiple elements, and then using the class selector to declare the shared style.

The view transition pseudo-elements (e.g. view-transition-group()) are not defined in the DOM, but rather by using the view-transition-name property. For that purpose, the view-transition-class' CSS property provides view transitions with the equivalent of HTML classes. When an element with a view-transition-name also has a view-transition-class value, that class would be selectable by the pseudo-elements, as per the examples.

4.2. Examples

This example creates a transition with each box participating under its own name, while applying a 1-second duration to the animation of all the boxes:
<div class="box" id="red-box"></div>
<div class="box" id="green-box"></div>
<div class="box" id="yellow-box"></div>
div.box {
  view-transition-class: any-box;
  width: 100px;
  height: 100px;
}
#red-box {
  view-transition-name: red-box;
  background: red;
}
#green-box {
  view-transition-name: green-box;
  background: green;
}
#yellow-box {
  view-transition-name: yellow-box;
  background: yellow;
}

/* The following style would apply to all the boxes, thanks to 'view-transition-class' */
::view-transition-group(*.any-box) {
  animation-duration: 1s;
}

4.3. The view-transition-class property

Name: view-transition-class
Value: none | <custom-ident>+
Initial: none
Applies to: all elements
Inherited: no
Percentages: n/a
Computed value: as specified
Canonical order: per grammar
Animation type: discrete

The view-transition-class can be used to apply the same style rule to multiple named view transition pseudo-elements which may have a different view-transition-name. While view-transition-name is used to match between the element in the old state with its corresponding element in the new state, view-transition-class is used only to apply styles using the view transitionpseudo-elements (::view-transition-group(), ::view-transition-image-pair(), ::view-transition-old(), ::view-transition-new()).

Note that view-transition-class by itself doesn’t mark an element for capturing, it is only used as an additional way to style an element that already has a view-transition-name.

none

No class would apply to the named view transition pseudo-elements generated for this element.

<custom-ident>+

All of the specified <custom-ident> values (apart from none) are applied when used in named view transition pseudo-element selectors. none is an invalid <custom-ident> for view-transition-class, even when combined with another <custom-ident>.

Note: If the same view-transition-name is specified for an element both in the old and new states of the transition, only the view-transition-class values from the new state apply. This also applies for cross-document view transitions: classes from the old document would only apply if their corresponding view-transition-name was not specified in the new document.

4.4. Additions to named view transition pseudo-elements

The named view transition pseudo-elements (view-transition-group(), view-transition-image-pair(), view-transition-old(), view-transition-new()) are extended to support the following syntax:

::view-transition-group(<pt-name-and-class-selector>)
::view-transition-image-pair(<pt-name-and-class-selector>)
::view-transition-old(<pt-name-and-class-selector>)
::view-transition-new(<pt-name-and-class-selector>)

where <pt-name-selector> works as previously defined, and <pt-name-and-class-selector> has the following syntax definition:

<pt-name-and-class-selector> = <pt-name-selector> <pt-class-selector>? | <pt-class-selector>
<pt-class-selector> = ['.' <custom-ident>]+

When interpreting the above grammar, white space is forbidden:

A named view transition pseudo-element selector which has one or more <custom-ident> values in its <pt-class-selector> would only match an element if the class list value in named elements for the pseudo-element’s view-transition-name contains all of those values.

The specificity of a named view transition pseudo-element selector with either:

is equivalent to a type selector.

The specificity of a named view transition pseudo-element selector with a * argument and with an empty <pt-class-selector> is zero.

5. Extending document.startViewTransition()

dictionary StartViewTransitionOptions {
  UpdateCallback? update = null;
  sequence<DOMString>? types = null;
};

partial interface Document {

  ViewTransition startViewTransition(optional (UpdateCallback or StartViewTransitionOptions) callbackOptions = {});
};
The method steps for startViewTransition(callbackOptions) are as follows:
  1. Let updateCallback be null.

  2. If callbackOptions is an an UpdateCallback, set updateCallback to callbackOptions.

  3. Otherwise, if callbackOptions is a StartViewTransitionOptions, then set updateCallback to callbackOptions’s update.

  4. If this’s active view transition is not null and its outbound post-capture steps is not null, then:

    1. Let preSkippedTransition be a new ViewTransition in this’s relevant realm whose update callback is updateCallback.

      Note: The preSkippedTransition’s types are ignored here because the transition is never activated.

    2. Skip preSkippedTransition with an "InvalidStateError" DOMException.

    3. Return preSkippedTransition.

    Note: This ensures that a same-document transition that started after firing pageswap is skipped.

  5. Let viewTransition be the result of running the method steps for startViewTransition(updateCallback) given updateCallback.

  6. If callbackOptions is a StartViewTransitionOptions, set viewTransition’s active types to a clone of types as a set.

  7. Return viewTransition.

6. Algorithms

6.1. Data structures

6.1.1. Additions to Document

A Document additionaly has:
inbound view transition params

a view transition params, or null. Initially null.

can initiate outbound view transition

a boolean. Initially false.

Note: this value can be read in parallel while navigating.

6.1.2. Additions to ViewTransition

A ViewTransition additionally has:
active types

A ViewTransitionTypeSet, initially empty.

outbound post-capture steps

Null or a set of steps, initially null.

6.1.3. Serializable view transition params

A view transition params is a struct whose purpose is to serialize view transition information across documents. It has the following items:
named elements

a map, whose keys are strings and whose values are captured elements.

initial snapshot containing block size

a tuple of two numbers (width and height).

6.1.4. Captured elements extension

The captured element struct should contain these fields, in addition to the existing ones:
class list

a list of strings, initially empty.

6.2. Resolving the @view-transition rule

To resolve @view-transition rule for a Document document:

Note: this is called in both the old and new document.

  1. If document’s visibility state is "hidden", then return "skip transition".

  2. Let matchingRule be the last @view-transition rule in document.

  3. If matchingRule is not found, then return "skip transition".

  4. If matchingRule’s navigation descriptor’s computed value is none, then return "skip transition".

  5. Assert: matchingRule’s navigation descriptor’s computed value is auto.

  6. Let typesDescriptor be matchingRule’s types descriptor.

  7. If typesDescriptor’s computed value is none, then return a set « ».

  8. Return a set of strings corresponding to typesDescriptor’s computed value.

6.3. Setting up the view transition in the old Document

6.3.1. Check eligibility for outbound cross-document view transition

To check if a navigation can trigger a cross-document view-transition? given a Document oldDocument, a Document newDocument, a NavigationType navigationType, and a boolean isBrowserUINavigation:

Note: this is called during navigation, potentially in parallel.

  1. If the user agent decides to display an implementation-defined navigation experience, e.g. a gesture-based transition for a back navigation, the user agent may ignore the author-defined view transition. If that is the case, return false.

  2. If oldDocument’s can initiate outbound view transition is false, then return false.

  3. If navigationType is reload, then return false.

  4. If oldDocument’s origin is not same origin as newDocument’s origin, then return false.

  5. If newDocument was created via cross-origin redirects, then return false.

  6. If navigationType is traverse, then return true.

  7. If isBrowserUINavigation is true, then return false.

  8. Return true.

6.3.2. Setup the outbound transition when ready to swap pages

To setup cross-document view-transition given a Document oldDocument, a Document newDocument, and proceedWithNavigation, which is an algorithm accepting nothing:

Note: This is called from the HTML spec.

  1. Assert: These steps are running as part of a task queued on oldDocument.

  2. If oldDocument’s can initiate outbound view transition is false, then return null.

  3. Let transitionTypesFromRule be the result of resolving the @view-transition rule for oldDocument.

  4. Assert: transitionTypesFromRule is not "skip transition".

    Note: We don’t know yet if newDocument has opted in, as it might not be parsed yet. We check the opt-in for newDocument when we fire the pagereveal event.

  5. If oldDocument’s active view transition is not null, then skip oldDocument’s active view transition with an "AbortError" DOMException in oldDocument’s relevant Realm.

    Note: this means that any running transition would be skipped when the document is ready to unload.

  6. Let outboundTransition be a new ViewTransition object in oldDocument’s relevant Realm.

  7. Set outboundTransition’s active types to transitionTypesFromRule.

    Note: the active types are not shared between documents. Manipulating the types in the new document does not affect the types in newDocument, which would be read from the types descriptor once newDocument is revealed.

    Note: the ViewTransition is skipped once the old document is hidden.

  8. Set outboundTransition’s outbound post-capture steps to the following steps given a view transition params-or-null params:

    1. Set newDocument’s inbound view transition params to params.

      Note: The inbound transition is activated after the dispatch of pagereveal to ensure mutations made in this event apply to the captured new state.

    2. Call proceedWithNavigation.

  9. Set oldDocument’s active view transition to outboundTransition.

    Note: The process continues in perform pending transition operations.

  10. The user agent should display the currently displayed frame until either:

    Note: this is to ensure that there are no unintended flashes between displaying the old and new state, to keep the transition smooth.

  11. Return outboundTransition.

6.3.3. Update the opt-in flag to reflect the current state

To update the opt-in state for outbound transitions for a Document document:
  1. If document has been revealed, and the result of resolving the @view-transition rule is not "skip transition", then set document’s can initiate outbound view transition to true.

  2. Otherwise, set document’s can initiate outbound view transition to false.

6.3.4. Proceed with navigation if view transition is skipped

Append the following steps to skip the view transition given a ViewTransition transition:
  1. If transition’s outbound post-capture steps is not null, then run transition’s outbound post-capture steps with null.

Note: This is written in a monkey-patch manner, and will be merged into the algorithm once the L1 spec graduates.

6.3.5. Proceed with cross-document view transition after capturing the old state

Prepend the following step to the Perform pending transition operations algorithm given a Document document:
  1. If document’s active view transition is not null and its outbound post-capture steps is not null, then:

    1. Assert: document’s active view transition's phase is "pending-capture".

    2. Let viewTransitionParams be null;

    3. Set document’s rendering suppression for view transitions to true.

      Though capture the old state appears here as a synchronous step, it is in fact an asynchronous step as rendering an element into an image cannot be done synchronously. This should be more explicit in the L1 spec.

    4. Capture the old state for transition.

    5. If this succeeded, then set viewTransitionParams to a new view transition params whose named elements is a clone of transition’s named elements, and whose initial snapshot containing block size is transition’s initial snapshot containing block size.

    6. Set document’s rendering suppression for view transitions to false.

    7. Call transition’s outbound post-capture steps given viewTransitionParams.

Note: This is written in a monkey-patch manner, and will be merged into the algorithm once the L1 spec graduates.

6.4. Activating the view transition in the new Document

To resolve inbound cross-document view-transition for Document document:
  1. Assert: document is fully active.

  2. Assert: document has been revealed is true.

  3. Update the opt-in state for outbound transitions for document.

  4. Let inboundViewTransitionParams be document’s inbound view transition params.

  5. If inboundViewTransitionParams is null, then return null.

  6. Set document’s inbound view transition params to null.

  7. If document’s active view transition is not null, then return null.

    Note: this means that starting a same-document transition before revealing the document would cancel a pending cross-document transition.

  8. Resolve @view-transition rule for document and let resolvedRule be the result.

  9. If resolvedRule is "skip transition", then return null.

  10. Let transition be a new ViewTransition in document’s relevant Realm, whose named elements is inboundViewTransitionParams’s named elements, and initial snapshot containing block size is inboundViewTransitionParams’s initial snapshot containing block size.

  11. Set document’s active view transition to transition.

  12. Resolve transition’s update callback done promise with undefined.

  13. Set transition’s phase to "update-callback-called".

  14. Set transition’s active types to resolvedRule.

  15. At any given time, the UA may decide to skip the inbound transition, e.g. after an implementation-defined timeout. To do so, the UA should queue a global task on the DOM manipulation task source given document’s relevant global object to perform the following step: If transition’s phase is not "done", then skip the view transition transition with a "TimeoutError" DOMException.

  16. Return transition.

6.5. Capturing the view-transition-class

When capturing the old or new state for an element, perform the following steps given a captured element capture and an element element:
  1. Set capture’s class list to the computed value of element’s view-transition-class.

Note: This is written in a monkey-patch manner, and will be merged into the algorithm once the L1 spec graduates.

Privacy Considerations

This specification introduces no new privacy considerations.

Security Considerations

To prevent cross-origin issues, at this point cross-document view transitions can only be enabled for same-origin navigations. As discussed in WICG/view-transitions#200, this still presents two potential threats:

  1. The cross-origin isolated capability in both documents might be different. This can cause a situation where a Document that is cross-origin isolated can read image data from a document that is not cross-origin isolated. This is already mitigated in [[css-view-transitions-1#sec], as the same restriction applies for captured cross-origin iframes.

  2. A same-origin navigation might still occur via a cross-origin redirect, e.g. https://example.com links to https://auth-provider.com/ which redirects back to https://example.com/loggedin.

    This can cause a (minor) situation where the cross-origin party would redirect the user to an unexpected first-party URL, causing an unexpected transition and obfuscating that fact that there was a redirect. To mitigate this, currently view transitions are disabled for navigations if the Document was created via cross-origin redirects. Note that this check doesn’t apply when the Document is being reactivated, as in that case the cross-origin redirect has already taken place.

    Note: this only applies to server-side redirects. A client-side redirect, e.g. using [^meta/http-equiv/refresh^], is equivalent to a new navigation.

  3. This feature exposes more information to CSS, as so far CSS was not aware of anything navigation-related. This can raise concerns around safety 3rd-party CSS. However, as a general rule, 3rd-party stylesheets should come from trusted sources to begin with, as CSS can learn about the document or change it in many ways.

See Issue #8684 and WICG/view-transitions#200 for detailed discussion.

Conformance

Document conventions

Conformance requirements are expressed with a combination of descriptive assertions and RFC 2119 terminology. The key words “MUST”, “MUST NOT”, “REQUIRED”, “SHALL”, “SHALL NOT”, “SHOULD”, “SHOULD NOT”, “RECOMMENDED”, “MAY”, and “OPTIONAL” in the normative parts of this document are to be interpreted as described in RFC 2119. However, for readability, these words do not appear in all uppercase letters in this specification.

All of the text of this specification is normative except sections explicitly marked as non-normative, examples, and notes. [RFC2119]

Examples in this specification are introduced with the words “for example” or are set apart from the normative text with class="example", like this:

This is an example of an informative example.

Informative notes begin with the word “Note” and are set apart from the normative text with class="note", like this:

Note, this is an informative note.

Advisements are normative sections styled to evoke special attention and are set apart from other normative text with <strong class="advisement">, like this: UAs MUST provide an accessible alternative.

Tests

Tests relating to the content of this specification may be documented in “Tests” blocks like this one. Any such block is non-normative.


Conformance classes

Conformance to this specification is defined for three conformance classes:

style sheet
A CSS style sheet.
renderer
A UA that interprets the semantics of a style sheet and renders documents that use them.
authoring tool
A UA that writes a style sheet.

A style sheet is conformant to this specification if all of its statements that use syntax defined in this module are valid according to the generic CSS grammar and the individual grammars of each feature defined in this module.

A renderer is conformant to this specification if, in addition to interpreting the style sheet as defined by the appropriate specifications, it supports all the features defined by this specification by parsing them correctly and rendering the document accordingly. However, the inability of a UA to correctly render a document due to limitations of the device does not make the UA non-conformant. (For example, a UA is not required to render color on a monochrome monitor.)

An authoring tool is conformant to this specification if it writes style sheets that are syntactically correct according to the generic CSS grammar and the individual grammars of each feature in this module, and meet all other conformance requirements of style sheets as described in this module.

Partial implementations

So that authors can exploit the forward-compatible parsing rules to assign fallback values, CSS renderers must treat as invalid (and ignore as appropriate) any at-rules, properties, property values, keywords, and other syntactic constructs for which they have no usable level of support. In particular, user agents must not selectively ignore unsupported component values and honor supported values in a single multi-value property declaration: if any value is considered invalid (as unsupported values must be), CSS requires that the entire declaration be ignored.

Implementations of Unstable and Proprietary Features

To avoid clashes with future stable CSS features, the CSSWG recommends following best practices for the implementation of unstable features and proprietary extensions to CSS.

Non-experimental implementations

Once a specification reaches the Candidate Recommendation stage, non-experimental implementations are possible, and implementors should release an unprefixed implementation of any CR-level feature they can demonstrate to be correctly implemented according to spec.

To establish and maintain the interoperability of CSS across implementations, the CSS Working Group requests that non-experimental CSS renderers submit an implementation report (and, if necessary, the testcases used for that implementation report) to the W3C before releasing an unprefixed implementation of any CSS features. Testcases submitted to W3C are subject to review and correction by the CSS Working Group.

Further information on submitting testcases and implementation reports can be found from on the CSS Working Group’s website at https://www.w3.org/Style/CSS/Test/. Questions should be directed to the public-css-testsuite@w3.org mailing list.

Index

Terms defined by this specification

Terms defined by reference

References

Normative References

[CSS-CASCADE-5]
Elika Etemad; Miriam Suzanne; Tab Atkins Jr.. CSS Cascading and Inheritance Level 5. URL: https://drafts.csswg.org/css-cascade-5/
[CSS-SYNTAX-3]
Tab Atkins Jr.; Simon Sapin. CSS Syntax Module Level 3. URL: https://drafts.csswg.org/css-syntax/
[CSS-VALUES-4]
Tab Atkins Jr.; Elika Etemad. CSS Values and Units Module Level 4. URL: https://drafts.csswg.org/css-values-4/
[CSS-VIEW-TRANSITIONS-1]
Tab Atkins Jr.; Jake Archibald; Khushal Sagar. CSS View Transitions Module Level 1. URL: https://drafts.csswg.org/css-view-transitions-1/
[CSS22]
Bert Bos. Cascading Style Sheets Level 2 Revision 2 (CSS 2.2) Specification. URL: https://drafts.csswg.org/css2/
[CSSOM-1]
Daniel Glazman; Emilio Cobos Álvarez. CSS Object Model (CSSOM). URL: https://drafts.csswg.org/cssom/
[DOM]
Anne van Kesteren. DOM Standard. Living Standard. URL: https://dom.spec.whatwg.org/
[HTML]
Anne van Kesteren; et al. HTML Standard. Living Standard. URL: https://html.spec.whatwg.org/multipage/
[INFRA]
Anne van Kesteren; Domenic Denicola. Infra Standard. Living Standard. URL: https://infra.spec.whatwg.org/
[RFC2119]
S. Bradner. Key words for use in RFCs to Indicate Requirement Levels. March 1997. Best Current Practice. URL: https://datatracker.ietf.org/doc/html/rfc2119
[SELECTORS-3]
Tantek Çelik; et al. Selectors Level 3. URL: https://drafts.csswg.org/selectors-3/
[SELECTORS-4]
Elika Etemad; Tab Atkins Jr.. Selectors Level 4. URL: https://drafts.csswg.org/selectors/
[WEBIDL]
Edgar Chen; Timothy Gu. Web IDL Standard. Living Standard. URL: https://webidl.spec.whatwg.org/

Informative References

[CSS-CONDITIONAL-3]
David Baron; Elika Etemad; Chris Lilley. CSS Conditional Rules Module Level 3. URL: https://drafts.csswg.org/css-conditional-3/

Property Index

Name Value Initial Applies to Inh. %ages Anim­ation type Canonical order Com­puted value
view-transition-class none | <custom-ident>+ none all elements no n/a discrete per grammar as specified

@view-transition Descriptors

Name Value Initial
navigation auto | none none
types none | <custom-ident>+ none

IDL Index

partial interface CSSRule {
  const unsigned short VIEW_TRANSITION_RULE = 15;
};

enum ViewTransitionNavigation { "auto", "none" };

[Exposed=Window]
interface CSSViewTransitionRule : CSSRule {
  readonly attribute ViewTransitionNavigation navigation;
  [SameObject] readonly attribute FrozenArray<CSSOMString> types;
};

[Exposed=Window]
interface ViewTransitionTypeSet {
  setlike<DOMString>;
};

[Exposed=Window]
partial interface ViewTransition {
  attribute ViewTransitionTypeSet types;
};

dictionary StartViewTransitionOptions {
  UpdateCallback? update = null;
  sequence<DOMString>? types = null;
};

partial interface Document {

  ViewTransition startViewTransition(optional (UpdateCallback or StartViewTransitionOptions) callbackOptions = {});
};

Issues Index

Though capture the old state appears here as a synchronous step, it is in fact an asynchronous step as rendering an element into an image cannot be done synchronously. This should be more explicit in the L1 spec.