Initialize `zoneSymbolEventNames` and `patches` with `Object.create(null)` instead of `{}`.
This is a hardening change rather than a fix for an exploitable vulnerability. Calling `addEventListener('__proto__', fn)` is not directly attacker-controlled; its presence already implies an application bug. However, if such a call does occur, the current implementation can behave unexpectedly depending on the environment.
For `zoneSymbolEventNames`, accessing `zoneSymbolEventNames['__proto__']` on a plain object invokes the inherited `__proto__` accessor and returns `Object.prototype`, which is truthy. This causes `prepareEventNames()` to be skipped, leaving `symbolEventName` undefined and eventually leading to a runtime error when `window['undefined'] = []` is executed.
In Node.js environments running with `--disable-proto=throw`, the assignment:
```ts id="z8n4qm"
zoneSymbolEventNames['__proto__'] = {};
```
throws immediately because it triggers the disabled `__proto__` setter.
The `patches` registry has a similar issue. A `__proto__` key passed to `__load_patch()` bypasses the duplicate-patch check and reaches:
```ts id="f3v7kx"
patches['__proto__'] = fn(...);
```
which invokes the `__proto__` setter and changes the prototype of the `patches` object.
Using `Object.create(null)` removes the inherited `__proto__` accessor entirely, causing these keys to behave like ordinary properties rather than interacting with JavaScript's prototype machinery.
As part of this change, `patches.hasOwnProperty(name)` is also updated to:
```ts id="n2c8wp"
Object.prototype.hasOwnProperty.call(patches, name)
```
since null-prototype objects do not inherit `hasOwnProperty`.
`model()` signals are special because they combine a signal input with a writable output through an internal `OutputEmitterRef`. During upgrade, `setupOutputs()` subscribes to that emitter to keep Angular → AngularJS two-way binding working.
The issue was that `updateInput()` could overwrite the signal property directly when `isSignal` was `false` (which happens in JIT mode and when `unsafelyOverwriteSignalInputs` is enabled). Once that happened, the original `OutputEmitterRef` was lost, so the two-way binding stopped working.
The fix detects `model()` signals at runtime by checking for both `[SIGNAL]` and a writable `.set()` method, which distinguishes them from read-only `input()` signals. When those traits are present, updates are always applied through `applyValueToInputSignal()` instead of replacing the property directly, regardless of the `unsafelyOverwriteSignalInputs` setting.
Fixes#60599
Without this change, components that use RouterLinkActive in multiple
contexts (e.g. both a navigation menu and body content) are forced to
branch the template for every conditional input:
@if (activeClass) {
<a [routerLink]="href" [routerLinkActive]="activeClass"
[routerLinkActiveOptions]="activeOptions"
[ariaCurrentWhenActive]="ariaCurrent">
<ng-content />
</a>
} @else {
<a [routerLink]="href"><ng-content /></a>
}
Every additional input multiplies the branching, and each @if/@else
injects unwanted comment nodes into the DOM. There is no way to
conditionally attach a directive in Angular templates, making imperative
TypeScript instantiation the only alternative.
Accepting null/undefined collapses this to a single template branch:
<a [routerLink]="href"
[routerLinkActive]="activeClass"
[routerLinkActiveOptions]="activeOptions"
[ariaCurrentWhenActive]="ariaCurrent">
<ng-content />
</a>
When activeClass is undefined (e.g. in content areas), the directive
stays mounted but applies no CSS classes. When it is a string (e.g. in
the navigation), normal active-class behavior applies — no branching, no
extra DOM nodes, no TypeScript workarounds.
- `routerLinkActive`: null/undefined now sets an empty class list.
- `routerLinkActiveOptions`: null and undefined are treated differently:
- undefined → falls back to the default subset match ("not set")
- null → explicit opt-out, link is never considered active
Closes#66233
Preserve explicit referrer policy when the service worker reconstructs asset requests for cache-busted and redirected asset fetches.
For example, an application can load a script or image with referrerPolicy: 'same-origin' or 'origin' to limit referrer data. Dropping that policy can expose more of the current URL to that resource host.
Preserve referrer metadata when the service worker reconstructs asset requests for cache-busted and redirected asset fetches.
For example, an attacker with access to asset host logs could receive a reset token embedded in a page URL if the reconstructed request falls back to default referrer behavior instead of carrying referrer: ''.
`parseHostBindings` throws plain `Error`s for malformed host bindings
(e.g. a property binding with a non-static value, as can happen while
editing in the language service). These were uncaught during directive
analysis, crashing the compiler and the Angular Language Service.
Wrap the call and surface the error as a `FatalDiagnosticError` so it
becomes a diagnostic and analysis can complete normally.
Fixes#69106
A custom FormValueControl only participates in debounce('blur') if it emits
the touch output on the native blur event. This was undocumented, and the
touch name reads like a focus event, so users wired it to (focus) and
blur-based debouncing silently did nothing.
Add a dedicated guide section with a working example, link the debounce API
reference to it, and clarify the touch JSDoc that it must fire on blur, not
focus.
Fixes#69370
Adds a 'Chaining resources' section to the resource guide covering:
- Basic usage of chain() to depend one resource on another
- Status propagation for all ResourceStatus values (idle, loading,
reloading, error, resolved, local)
- Chaining vs reading .value() directly, shown as an avoid example
- Guidance on passing the chained value directly as params
Also adds an @see link from ResourceParamsContext to the new section.
Closes#69329
Adds an error reference page for NG05102 (UNSUPPORTED_EVENT_TARGET) explaining
what triggers it and how to fix it. Also marks the error code as negative (-5102)
so that in dev mode the error message automatically links to the new guide page
on angular.dev/errors, consistent with other documented runtime errors.
Skip HttpTransferCache serialization for HTTP responses that contain a
Set-Cookie header.
Cookie-setting responses commonly represent session-specific,
user-specific, or security-sensitive state. Serializing their bodies into
SSR TransferState can embed sensitive data into the generated HTML, where
it may be reused during hydration or replayed by a shared cache/CDN.
Fixes a minor issue where the `preload` function in `injectAsync` might cause an uncaught promise error.
I also fixed that in `onIdle` we were passing the wrong function into `assertInInjectionContext`.
Adds a new error reference page for NG05101 (NO_PLUGIN_FOR_EVENT),
which is thrown when no registered EventManagerPlugin supports the
event name passed to addEventListener. The page covers the two common
causes: a typo in the event binding and a missing plugin provider.
After #68581 graduated the Signal Forms APIs to public API and #68654 removed
the experimental warnings from the Signal Forms documentation in adev, the
package READMEs still framed the API as experimental.
Update `packages/forms/signals/PACKAGE.md`: drop the experimental title and
intro, remove the now-shipped entries from "Not yet supported" (interop with
reactive/template forms and strongly-typed binding to UI controls), and remove
the remaining experimental and exploratory wording from the FAQ.
Update `packages/forms/PACKAGE.md`: it listed only two ways to build forms
(reactive and template-driven). Add signal forms as the third.
Fixes#68724
Previously, the modelContext truthiness check could be bypassed via DOM
clobbering (e.g. `<form id="modelContext">`), causing a truthy HTMLElement
to pass the guard and then throw when `registerTool` was called on it.
Replace the truthiness check with a duck-type check that asserts
`registerTool` is a function, rejecting both absent and clobbered values.
Add handling in navigation for creating and destroying injectors scoped
to `ActivatedRoute` life.
The code for creating the injectors is certainly more complicated
than it _could_ be since there's no actual feature built around this yet.
Keeps as much implementation code tree-shakeable as possible:
Raw size: +764 bytes
Gzipped size: +182 bytes
This commit addresses a scaling issue in the signal dependency graph where
the detection of duplicate dependency links would perform a linear scan across
all consumer links of all producers. The linear scan is replaced with a version
comparison of the dependency edge against the current epoch; if they are
equal the existing dependency edge is known to be valid in this epoch. This means
that the link won't be eligible for removal and therefore doesn't have to be
recreated.
`COMMENT_DISALLOWED` is matched globally, so overlapping delimiter
sequences are skipped: `<!-->` only escapes the leading `<!--` and
leaves a live `-->` that can close a programmatically created comment
node early. Drop the `^` anchors so a standalone `>`/`->` is escaped
wherever it appears, which neutralizes the trailing delimiter left
behind by an earlier match.
Adds support for `@Input` transform functions in isolated declarations mode (`emitDeclarationOnly: true`), allowing components and directives to specify `transform` functions without triggering fatal compiler errors.
Synthesizes the `ngAcceptInputType_` write type syntactically:
- For referenced functions (`transform: booleanAttribute`), emits `Parameters<typeof booleanAttribute>[0]`, relying on downstream template type checking to resolve the type.
- For inline functions (`transform: (v: string) => boolean`), extracts `parameters[0].type` directly from the local TypeScript AST.
Removes restrictions around using external references and local directives in `hostDirectives` under isolated declarations mode (`emitDeclarationOnly: true`).
By wrapping the host directive reference in a `WrappedNodeExpr`, TypeScript's declaration emitter seamlessly emits `typeof hostReference.node`, preserving existing imports or local identifiers exactly as authored. Also adds support for translating `PropertyAccessExpression` inside `WrappedNodeExpr` into `QualifiedName` for `.d.ts` emission, ensuring namespace imports (`import * as n from './dir'`) are preserved correctly.
Mark the iframe `credentialless` attribute as security-sensitive so dynamic
bindings are handled consistently with other iframe attributes that affect the
initial navigation, such as `sandbox`, `allow`, `referrerPolicy`, `csp`, and
`fetchPriority`.
Because `credentialless` must be present before the iframe starts loading to
affect the navigation’s credential mode, late dynamic updates can leave the final
DOM looking correct while the initial request was not loaded credentiallessly.
Fixes that when a listener unsubscribes from an `output` within its own callback, it was preventing subsequent listeners from running.
These changes fix the issue by not mutating the array while the emit loop is running, but replacing the listener with `null` and coming back later to remove it.
Fixes#69325.
Only cache locale data loaded from the global locale registry when an actual locale entry is found.
This prevents attacker-controlled missing locale identifiers from being retained indefinitely in SSR when locale lookup falls back to a parent locale or the built-in English locale, avoiding unbounded process memory growth in locale-aware pipes and formatters.
Previously, the compiler disallowed translation of any attribute starting with 'on' for security reasons. This incorrectly disallowed translation of the 'on' attribute itself, which is not an event handler.
This commit introduces `isPossibleEventHandler` to verify that the property name has a length greater than 2 in addition to starting with 'on'. This allows attributes like 'on' to be translated while still correctly disallowing actual event handlers like 'onerror', 'onclick', etc.
`animate.leave` was skipped — the element was removed from the DOM
synchronously instead of running its leave animation — whenever a
sibling instance of the same template entered in a different DOM parent
during the same change-detection tick (e.g. an exclusive-expansion
accordion or nav where opening section B collapses section A).
`leavingNodes` is keyed by `TNode`, which is shared by every instance of
a template. When a node was inserted, `cancelLeavingNodes` force-removed
any tracked leaving node whose DOM parent differed from the entering
node's parent (the `leavingParent !== newParent` branch added to
de-duplicate a dynamic component re-rendered into a fresh overlay pane).
For two distinct live sibling instances that merely share a `TNode`,
"different parent" is the normal situation, so the still-animating
sibling was ripped out.
Track the declaration view of each leaving element alongside it, and
only perform the cross-parent removal when the entering element belongs
to the same declaration view as the leaving one — i.e. the same logical
view re-rendered, the case the branch was written for. Two distinct
instances of a shared template have different declaration views, so
their `animate.leave` is now left to run to completion.
This preserves the dynamic-component/overlay de-duplication (#67032) and
the drag-and-drop node-move rescue (#67361), which are unchanged.
Fixes#69291
Serialize transfer cache request parameters without comma-joining repeated values so distinct HttpClient requests cannot reuse the same cached response.
Initialize zoneSymbolEventNames with Object.create(null) instead of {}.
This is hardening only. addEventListener('__proto__', fn) is not
directly attacker-controllable — its presence in an application is
itself an application bug and a prerequisite for any issue here.
Without this change, if that application bug exists, two unexpected
behaviors follow depending on environment:
Browser: zoneSymbolEventNames['__proto__'] reads the __proto__ getter
and returns Object.prototype (truthy), bypassing prepareEventNames.
symbolEventName resolves to undefined and window['undefined'] = []
throws TypeError.
Node.js + --disable-proto=throw: the assignment
zoneSymbolEventNames['__proto__'] = {} inside prepareEventNames
triggers the disabled __proto__ setter and throws.
Using Object.create(null) removes the __proto__ accessor from the
map so the key is treated as a plain missing property in both cases.
`findAnchorFromDocument` interpolates the raw url fragment into
`[name="${target}"]` for the shadow DOM lookup, so a fragment reachable
through the router when `anchorScrolling` is on can break out of the
attribute selector and make `querySelector` throw or match unrelated
nodes, and it also breaks legitimate anchor names containing a quote.
Wrap the value in `CSS.escape` so it stays a single attribute-value token.
This limit breaks ts-ignore comments when using this for our source->source transform.
Rather than overridding it there, it's just removed here since we don't care about the limit
This test flakes occasionally because it is called in production when a platform is created and unregistered when a platform is destroyed. However, not all tests properly clean up their platforms, meaning we can accidentally leak platforms between tests. If this happens, we end up have an event listener created from the production code path and a second event listener from the test. When the test emits the event, both listeners respond and it causes too many responses which fails the test.
Ideally, all tests would clean up the platforms correctly, but this seems difficult to guarantee for all Angular tests and is likely to break over time. The simplest solution is just destroy any leaked platform before the test starts. It's a bit elegant, but the safest option.
Normalize SVG animation attributeName lookup to also recognize lowercase attributename before allowing dynamic animation value bindings.
Add runtime and platform-server SSR regression coverage for lowercase attributename retargeting.
Matthieu Riegler