Prior to this change, we've used `componentDef.id` as a key in a Map that acts as a cache to avoid re-creating injector instances for standalone components. In v16, the logic that generates the id has changed from an auto-incremental to a generation based on metadata. If multiple components have similar metadata, their ids might overlap.
This commit updates the logic to stop using `componentDef.id` as a key and instead, use the `componentDef` itself. This would ensure that we always have a correct instance of an injector associated with a standalone component instance.
Resolves#50724.
PR Close#50954
Exposes the function used to transform an input on `ComponentFactory.inputs` and `ComponentMirror.inputs`. We'll need this to support input transforms in `elements`.
PR Close#50713
Components are implied to be self-referencing, but if they explicitly set themselves in the `imports` array, they would throw an error because we weren't filtering them out.
Fixes#50525.
PR Close#50559
According to the HTML specification most attributes are defined as strings, however some can be interpreted as different types like booleans or numbers. [In the HTML standard](https://html.spec.whatwg.org/multipage/common-microsyntaxes.html#boolean-attributes), boolean attributes are considered `true` if they are present on a DOM node and `false` if they are omitted. Common examples of boolean attributes are `disabled` on interactive elements like `<button>` or `checked` on `<input type="checkbox">`. Another example of an attribute that is defined as a string, but interpreted as a different type is the `value` attribute of `<input type="number">` which logs a warning and ignores the value if it can't be parsed as a number.
Historically, authoring Angular inputs that match the native behavior in a type-safe way has been difficult for developers, because Angular interprets all static attributes as strings. While some recent TypeScript versions made this easier by allowing setters and getters to have different types, supporting this pattern still requires a lot of boilerplate and additional properties to be declared. For example, currently developers have to write something like this to have a `disabled` input that behaves like the native one:
```typescript
import {Directive, Input} from '@angular/core';
@Directive({selector: 'mat-checkbox'})
export class MatCheckbox {
@Input()
get disabled() {
return this._disabled;
}
set disabled(value: any) {
this._disabled = typeof value === 'boolean' ? value : (value != null && value !== 'false');
}
private _disabled = false;
}
```
This feature aims to address the issue by introducing a `transform` property on inputs. If an input has a `transform` function, any values set through the template will be passed through the function before being assigned to the directive instance. The example from above can be rewritten to the following:
```typescript
import {Directive, Input, booleanAttribute} from '@angular/core';
@Directive({selector: 'mat-checkbox'})
export class MatCheckbox {
@Input({transform: booleanAttribute}) disabled: boolean = false;
}
```
These changes also add the `booleanAttribute` and `numberAttribute` utilities to `@angular/core` since they're common enough to be useful for most projects.
Fixes#8968.
Fixes#14761.
PR Close#50420
This commit updates the `ApplicationRef.isStable` API to account for
pending rendering task. This is needed as once a pending rendering task
is done, new macrotask and microtask could be created which previously caused these not
to be intercepted and thus ignored when doing SSR.
PR Close#50425
When an embedded view injector is present anywhere above a node in the tree, the `Self` flag was effectively ignored. With this change, embedded view injectors are not checked at all when the `Self` flag is present, because resolution should stop at the current node before reaching any embedded view injector(s).
Fixes#49959
PR Close#50270
Fixes that the host directives feature was incorrectly throwing the conflicting alias error when an aliased binding was being exposed under the same alias.
Fixes#48951.
PR Close#50364
Related to #50272 and #18970, this improves the error message of NG100 by including the class name of the component where the error was triggered.
PR Close#50286
This change explicitly resets a reactive consumer before setting inputs
on directive instances. This is to assure that any potential input setters
do _not_ run in the reactive context.
PR Close#49906
When binding an array to `class` like `[class]="['foo', 'bar']"`, the runtime treats it the same as a literal binding with all the values being `true`, e.g. `{foo: true, bar: true}`. While object literals can only have string keys, arrays can have any value which can lead to errors if the array contains non-string values.
These changes add some logic to stringify the keys and ignore invalid ones.
Fixes#48473.
PR Close#49924
This fix assures that templates functions executed in the creation mode
are run outside of the reactive context. This avoids the situation where
signal reads in a directive constructor (executed as part of the creation
mode) would mark the host component as dirty.
Fixes#49871
PR Close#49883
Currently, the `ViewRef.rootNodes` output is missing anchor (comment) nodes for inner `ViewContainerRef`s,
when an achor node was created for that instance of a `ViewContainerRef` (which happens in all cases except
when an <ng-container> was used as a host for a view container).
This issue affects hydration logic, which relies on the number of root nodes within a view to properly determine
segments in DOM that belong to a particular view.
Resolves#49849.
PR Close#49867
It might happen that the lifecycle scope represented by DestroyRef becomes
invalid before an onDestroy hook is registered (ex. injector or component
instance got destroyed). In such cases registration of the onDestroy hooks
should no longer be possible.
Fixes#49658
PR Close#49804
Angular doesn't support IE anymore. We can remove the workarounds related to IE.
Some workarounds are keep because of the support of domino but the comments related to IE are removed.
PR Close#49763
This commits updates the render to able to handle the slight differences between platform-server and platform-browser.
This is needed to eventually be able to remove `ServerRendererFactory2` and `EmulatedEncapsulationServerRenderer2` from platform-server.
PR Close#49630
Setting the `nonce` attribute using the property is not supported by Domino. This change update the usage to use `setAttribute` and also add a test to verify that the `nonce` is set when it should.
PR Close#49624
Previously the signals implementation maintained separate interfaces for
`Producer` and `Consumer` nodes, with implementers choosing to implement
one or both interfaces. Operations defined against those interfaces were
exposed as importable functions to be called with the object implementing
the relevant contract as the first argument.
This commit refactors the implementation to merge both abstractions into a
single `ReactiveNode` base class, which represents both producers and
consumers in the graph. Implementers choose to interact with a subset of the
`ReactiveNode` API depending on their role in the graph. Operations are now
available as protected methods on the base class, instead of separate
functions.
PR Close#49529
This commit adds an assertion function to the public API, which allows
authors of functions which rely on `inject` to validate that they're being
called with the right context. This mostly produces a nicer error message
than calling `inject()` and relying on Angular's default error message for
that.
PR Close#49529
This commit implements a simple tracker of the pending tasks during initial rendering. The class allows adding and removing tasks from the set. The class also exposes a promise that gets resolved once the last task is removed.
This tracker is needed to keep track of ongoing processes like Router navigation (and potentially HTTP requests) and acts as a signaling mechanism to SSR and hydration that the application is in the "stable" state and a serialization can be performed.
This class would also act as a future replacement for the `ApplicationRef.isStable` for zoneless applications.
PR Close#49576
Currently we are unsafely unquoting CSS values which in some cases causes valid values to become invalid and invalid values to become valid.
Example:
```html
<div style="width:"1px;""></div>
```
In the above case, `width` has an invalid value of `"1px"`, however the compiler will transform it to `1px` which makes it valid.
On the other hand, in the below case
```html
<div style="content:"foo""></div>
```
`content` has a valid value of `"foo"`, but since the compiler unwraps it to `foo` it becomes invalid. For correctness, we should not remove quotes.
```js
const div = document.createElement('div');
div.style.width='"1px"';
div.style.content='foo';
div.style.width; // ''
div.style.content; // ''
div.style.width='1px';
div.style.content='"foo"';
div.style.width; // '1px'
div.style.content; // '"foo"'
```
More information about values can be found https://www.w3.org/TR/CSS21/syndata.html#value-def-identifier
PR Close#49460
This commit updates the `LView` in Angular to be a `Consumer` of
signals. If a signal is read when executing a template, it marks the
view dirty. In addition, if a signal is read when executing host
bindings, it also marks views dirty.
One interesting thing about signal reads in host bindings
is that they perform a bit better than what we can do with today's
APIs. In order to re-execute host bindings for an `OnPush` component that
might have changed, you would probably inject `ChangeDetectorRef` and call
`markForCheck`. This will mark the _current component_ and parents
dirty. However, host bindings are executed as part of refreshing the
_parent_ so there is really no need to re-execute the current component
if the only thing that changed is the host bindings. When a signal is
read in host bindings, it marks the parent dirty and not the component
that defined the host binding.
Additionally, this commit avoids allocating a full consumer for each
`LView` by re-using a consumer until template execution results in a
signal read. At this point, we assign that consumer to the `LView` and
create a new consumer to "tentatively" use for the future `LView`
template executions.
Co-authored-by: Dylan Hunn <github@dylanhunn.com>
PR Close#49153
`entryComponents` have been deprecated since version 9, because with Ivy they weren't necessary. These changes remove any remaining references.
BREAKING CHANGE:
* `entryComponents` has been deleted from the `@NgModule` and `@Component` public APIs. Any usages can be removed since they weren't doing anyting.
* `ANALYZE_FOR_ENTRY_COMPONENTS` injection token has been deleted. Any references can be removed.
PR Close#49484
This change makes it possible to remove a previously registered destroy
callback - to do so it is enough to call the unregistration function
returned from the onDestroy method call.
PR Close#49493
Showing a minimum app to reproduce the bug.
1. Create the app and add angular material.
```
ng new project
cd project
ng add @angular/material
```
1. Overwrite the src/app/app.component.html with minimal content.
```
<button mat-button *ngIf="true"><span *ngIf="true" class="{{'a'}}"></span></button>
```
1. Run the app. The button is not shown because of an exception.
```
main.ts:6
ERROR TypeError: item.toLowerCase is not a function
at isCssClassMatching (core.mjs:8726:35)
at isNodeMatchingSelector (core.mjs:8814:22)
at isNodeMatchingSelectorList (core.mjs:8931:13)
at matchingProjectionSlotIndex (core.mjs:14179:13)
at Module.ɵɵprojectionDef (core.mjs:14222:49)
at MatButton_Template (button.mjs:113:99)
at executeTemplate (core.mjs:10534:9)
at renderView (core.mjs:10356:13)
at renderComponent (core.mjs:11529:5)
at renderChildComponents (core.mjs:10216:9)
```
Because isCssClassMatching() function does not take care if the value is not string, while attrs[] may contain AttributeMarker which is actually numbers, item.toLowerCase() throws the exception.
Just inserted a check if the item is string.
Created a testcase for the original fix. It causes an exception without the fix.
fix(core): add a check code to avoid an exception inside isCssClassMatching
Showing a minimum app to reproduce the bug.
1. Create the app and add angular material.
```
ng new project
cd project
ng add @angular/material
```
1. Add `import { MatButtonModule } from '@angular/material/button'`,
and also MatButtonModule inside @NgModule imports in src/app/app.module.ts to use MatButtonModule.
1. Overwrite the src/app/app.component.html with minimal content.
```
<button mat-button *ngIf="true"><span *ngIf="true" class="{{'a'}}"></span></button>
```
1. Run the app. The button is not shown because of an exception.
```
main.ts:6
ERROR TypeError: item.toLowerCase is not a function
at isCssClassMatching (core.mjs:8726:35)
at isNodeMatchingSelector (core.mjs:8814:22)
at isNodeMatchingSelectorList (core.mjs:8931:13)
at matchingProjectionSlotIndex (core.mjs:14179:13)
at Module.ɵɵprojectionDef (core.mjs:14222:49)
at MatButton_Template (button.mjs:113:99)
at executeTemplate (core.mjs:10534:9)
at renderView (core.mjs:10356:13)
at renderComponent (core.mjs:11529:5)
at renderChildComponents (core.mjs:10216:9)
```
Because isCssClassMatching() function does not take care if the value is not string, while attrs[] may contain AttributeMarker which is actually numbers, item.toLowerCase() throws the exception.
Just inserted a check if the item is string.
PR Close#48888
Adds support for marking a directive input as required. During template type checking, the compiler will verify that all required inputs have been specified and will raise a diagnostic if one or more are missing. Some specifics:
* Inputs are marked as required by passing an object literal with a `required: true` property to the `Input` decorator or into the `inputs` array.
* Required inputs imply that the directive can't work without them. This is why there's a new check that enforces that all required inputs of a host directive are exposed on the host.
* Required input diagnostics are reported through the `OutOfBandDiagnosticRecorder`, rather than generating a new structure in the TCB, because it allows us to provide a better error message.
* Currently required inputs are only supported during AOT compilation, because knowing which bindings are present during JIT can be tricky and may lead to increased bundle sizes.
Fixes#37706.
PR Close#49468
When having a recursive circle of imports on standalone components, `queueTypesFromModulesArrayRecur` triggered a `Maximum call stack size exceeded` error.
This commit fixes this.
Fixes#49469
PR Close#49473
This reverts commit 13dd614cd1.
This breaks a g3 Typescript compilation tests where diagnostics are
expected for a missing input in the component.
PR Close#49467
Angular uses inline styles to insert the styles associated with a component. This violates the strict styles [Content Security Policy](https://web.dev/strict-csp/) which doesn't allow inline styles by default. One way to allow the styles to be applied is to set a `nonce` attribute on them, but because the code for inserting the stylesheets is deep inside the framework, users weren't able to provide it without accessing private APIs.
These changes add a new `CSP_NONCE` injection token that will allow users to provide a nonce, if their app is using CSP. If the token isn't provided, the framework will look for an `ngCspNonce` attribute on the app's root node instead. The latter approach is provided as a convenience for apps that render the `index.html` through a server, e.g. `<app ngCspNonce="{% randomNonceAddedByTheServer %}"></app>`.
This PR addresses adding the nonce to framework-generated styles. There will be follow-up PRs that add support for it in critical CSS tags in the CLI, and in Angular Material.
Fixes#6361.
PR Close#49444
Adds support for marking a directive input as required. During template type checking, the compiler will verify that all required inputs have been specified and will raise a diagnostic if one or more are missing. Some specifics:
* Inputs are marked as required by passing an object literal with a `required: true` property to the `Input` decorator or into the `inputs` array.
* Required inputs imply that the directive can't work without them. This is why there's a new check that enforces that all required inputs of a host directive are exposed on the host.
* Required input diagnostics are reported through the `OutOfBandDiagnosticRecorder`, rather than generating a new structure in the TCB, because it allows us to provide a better error message.
* Currently required inputs are only supported during AOT compilation, because knowing which bindings are present during JIT can be tricky and may lead to increased bundle sizes.
Fixes#37706.
PR Close#49453
The mentioned 2 classes have been combined since it is no longer required to have a separate `SharedStylesHost` for SSR. This changes also reduces the memory usage footprint as remove 1 Map that stores the CSS strings.
PR Close#49424
This reverts commit 1a6ca68154.
This breaks tests in google3 which might be depending on private APIs. We
need to update these tests before we can land this PR.
PR Close#49449
Adds support for marking a directive input as required. During template type checking, the compiler will verify that all required inputs have been specified and will raise a diagnostic if one or more are missing. Some specifics:
* Inputs are marked as required by passing an object literal with a `required: true` property to the `Input` decorator or into the `inputs` array.
* Required inputs imply that the directive can't work without them. This is why there's a new check that enforces that all required inputs of a host directive are exposed on the host.
* Required input diagnostics are reported through the `OutOfBandDiagnosticRecorder`, rather than generating a new structure in the TCB, because it allows us to provide a better error message.
* Currently required inputs are only supported during AOT compilation, because knowing which bindings are present during JIT can be tricky and may lead to increased bundle sizes.
Fixes#37706.
PR Close#49304
With the introduction of `EnvironmentInjector`, we added an operation to run
a function with access to `inject` tokens from that injector. This operation
only worked for `EnvironmentInjector`s and not for element/node injectors.
This commit deprecates `EnvironmentInjector.runInContext` in favor of a
standalone API `runInInjectionContext`, which supports any type of injector.
DEPRECATED: `EnvironmentInjector.runInContext` is now deprecated, with
`runInInjectionContext` functioning as a direct replacement:
```typescript
// Previous method version (deprecated):
envInjector.runInContext(fn);
// New standalone function:
runInInjectionContext(envInjector, fn);
```
PR Close#49396
Prior to this change component styles generated on the server where removed prior to the client side component being rendered and attached it's own styles. In some cases this caused flickering. To mitigate this `initialNavigation: enabledBlocking'` was introduced which allowed the remove of server styles to be defer to a latter stage when the application has finished initialization.
This commit changes the need for this, by not removing the server generated component styles and reuse them for client side rendering.
PR Close#48253
DestroyRef represents a concept of lifecycle scope where destroy
callbacks can be registered. Such callbacks are automatically
executed when a given scope ends it lifecycle.
In practice the most common lifecycle scopes would be represented by:
- a component or en embedded view;
- instance of `EnvironnementInjector`.
PR Close#49158
Fixes that the expression converter was producing code that throws a runtime error if a non-null assertion is used as a part of a safe read, write or call.
Fixes#48742.
PR Close#48801
`makeEnvironmentProviders` constructs the wrapped `EnvironmentProviders`
type, which can only be used in environment injectors (not element
injectors). It makes sense that `makeEnvironmentProviders` should be able
to accept existing `EnvironmentProviders`-wrapped providers, since it will
be providing the same guarantee, but the current types do not allow this.
This commit fixes the typings to allow nesting `EnvironmentProviders` and
adds a test to verify that it will work.
PR Close#48720
Allows for self-closing tags to be used for non-native tag names, e.g. `<foo [input]="bar"></foo>` can now be written as `<foo [input]="bar"/>`. Native tag names still have to have closing tags.
Fixes#39525.
PR Close#48535
Since we generate a `.mjs` file as entry-point for jasmine tests,
a couple of issues prevented the transitive dependencies from
bootstrap targets to be brought in (causing resolution errors):
1. The `_files` (previously `_esm2015`) targets are no longer needed,
and they also miss all the information on runfiles.
2. The aspect for computing linker mappings does not respect the
`bootstrap` attribute from the `spec_entrypoint` so we manually
add the extract ESM output targets (this rule works with the aspect
and forwards linker mappings).
PR Close#48521
Matthieu Riegler