If a library is compiling with Angular v16.1.0, the library will break
for users that are still on Angular v16.0.x. This happens because the
`DirectiveDeclaration` or `ComponentDeclaration` types are not expecting
an extra field for `signals` metadata. This field was only added to the
generic types in `16.1.0`- so compilations fail with errors like this:
```
Error: node_modules/@angular/material/icon/index.d.ts:204:18 -
error TS2707: Generic type 'ɵɵComponentDeclaration' requires between 7 and 9 type arguments.
```
To fix this, we quickly roll back the code for inserting this metadata
field. That way, libraries remain compatible with all v16.x framework
versions.
We continue to include the `signals` metadata if `signals: true` is set.
This is not public API anyway right now- so cannot happen- but imagine
we expose some signal APIs in e.g. 16.2.x, then we'd need this metadata
and can reasonably expect signal-component library users to use a more
recent framework core version.
PR Close#50714
The component exposed by the fixture is not important thus marked as `unknown`.
Exposing the fixture of the `RouterTestingHarness` allows to use the methods & properties of that fixture and makes it compatible with testing libraries relying on `ComponentFixture`
Fixes#48855
PR Close#50280
This commit adds the `signals: boolean` property to the internal
directive/component metadata. This does not add it to the public API
yet, as the feature has no internal support other than compiler
detection.
PR Close#49981
Adds ability for `RouterOutlet` to bind `Router` information to the routed
component's inputs. This commit also exposes some helpers for
implementers of custom outlets to do their own input binding if desired.
Resolves#18967
PR Close#49633
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
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
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
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
BREAKING CHANGE:
`ComponentFactoryResolver` has been removed from Router APIs.
Component factories are not required to create an instance of a component
dynamically. Passing a factory resolver via resolver argument is no longer needed
and code can instead use `ViewContainerRef.createComponent` without the
factory resolver.
PR Close#49239
The `Router` already has the `getCurrentNavigation()` helper function to
expose information about the ongoing navigation.
`lastSuccessfulNavigation` is actually already exposed in the
`Navigation` object. This commit only _slightly_ extends the current API
surface by providing access to the `lastSuccessfulNavigation` object
outside of an ongoing navigation.
fixes#45685
PR Close#49235
This commit introduces helper functions to easily convert `Injectable`s with
functions compatible with `Route` guards to the corresponding guard functions.
These functions will serve to aid in migrating off of the now deprecated
class-based guards, but also provide an easy avenue to still defining
guards as `Injectable` classes if that is desired.
PR Close#48709
41e2a68e30 added a `type` property on all router events, and added all type of events to the `Event` union type, but forgot to remove `RouterEvent`.
This removes the benefit of the `type` field, because it is not possible to write:
```
this.router.events.pipe(filter((event: Event): event is NavigationEnd => event.type === EventType.NavigationEnd)).subscribe(/*...*/);
```
as `RouterEvent` does not have a `type` field (hence TS complains).
This commit fixes the issue by removing `RouterEvent` from the union type.
BREAKING CHANGE: The `RouterEvent` type is no longer present in the `Event` union type representing all router event types. If you have code using something like `filter((e: Event): e is RouterEvent => e instanceof RouterEvent)`, you'll need to update it to `filter((e: Event|RouterEvent): e is RouterEvent => e instanceof RouterEvent)`.
PR Close#46061
The Router scroller only listens for NavigationEnd events. However, the
default behavior of the Router is to ignore navigations to the same URL.
This breaks the anchor scrolling when clicking on an anchor whose
fragment is already in the URL.
fixes#29099
BREAKING CHANGE: The `Scroll` event's `routerEvent` property may also be
a `NavigationSkipped` event. Previously, it was only a `NavigationEnd`
event.
PR Close#48025
In order to test components and services which depend on router navigations, such as `ActivatedRoute` instances, tests currently need to provide a fair bit of boilerplate _or_ they can set up a stub for `ActivatedRoute` and list it in the `providers` to override it in `TestBed`. This approach of stubbing the `ActivatedRoute` creates a situation that can easily cause the test to break. The stub often only mocks out the dependencies that the component/service _currently_ needs. This dependencies might change over time and break the test in an unexpected way. In addition, it is difficult to get the structure of `ActivatedRoute` exactly correct.
This change will allow unit tests to quickly set up routes, trigger real navigations in the Router, and get instances of component's to test along with real instances of `ActivatedRoute`. This all comes without needing to know that the component depends on `ActivatedRoute` at all. This becomes more important when considering that a component may be refactored in the future to use `@Input` rather than access data on the `ActivatedRoute` instance (see #18967). Tests which mock out `ActivatedRoute` would all break, but those which use `navigateForTest` would continue to work without needing any updates.
resolves#15779resolves#48608
PR Close#48552
Class and `InjectionToken`-based guards and resolvers are not as configurable,
are less re-usable, require more boilerplate, cannot be defined inline with the route,
and require more in-depth knowledge of Angular features (`Injectable`/providers).
In short, they're less powerful and more cumbersome.
In addition, continued support increases the API surface which in turn increases
bundle size, code complexity, the learning curve and API surface to teach,
maintenance cost, and cognitive load (needing to grok several different types
of information in a single place).
Lastly, supporting only the `CanXFn` types for guards and `ResolveFn` type
for resolvers in the `Route` interface will enable better code
completion and integration with TypeScript. For example, when writing an
inline functional resolver today, the function is typed as `any` and
does not provide completions for the `ResolveFn` parameters. By
restricting the type to only `ResolveFn`, in the example below
TypeScript would be able to correctly identify the `route` parameter as
`ActivatedRouteSnapshot` and when authoring the inline route, the
language service would be able to autocomplete the function parameters.
```
const userRoute: Route = {
path: 'user/:id',
resolve: {
"user": (route) => inject(UserService).getUser(route.params['id']);
}
};
```
Importantly, this deprecation only affects the support for class and
`InjectionToken` guards at the `Route` definition. `Injectable` classes
and `InjectionToken` providers are _not_ being deprecated in the general
sense. Functional guards are robust enough to even support the existing
class-based guards through a transform:
```
function mapToCanMatch(providers: Array<Type<{canMatch: CanMatchFn}>>): CanMatchFn[] {
return providers.map(provider => (...params) => inject(provider).canMatch(...params));
}
const route = {
path: 'admin',
canMatch: mapToCanMatch([AdminGuard]),
};
```
With regards to tests, because of the ability to map `Injectable`
classes to guard functions as outlined above, nothing _needs_ to change
if projects prefer testing guards the way they do today. Functional
guards can also be written in a way that's either testable with
`runInContext` or by passing mock implementations of dependencies.
For example:
```
export function myGuardWithMockableDeps(
dep1 = inject(MyService),
dep2 = inject(MyService2),
dep3 = inject(MyService3),
) { }
const route = {
path: 'admin',
canActivate: [() => myGuardWithMockableDeps()]
}
// test file
const guardResultWithMockDeps = myGuardWithMockableDeps(mockService1, mockService2, mockService3);
const guardResultWithRealDeps = TestBed.inject(EnvironmentInjector).runInContext(myGuardWithMockableDeps);
```
DEPRECATED: Class and `InjectionToken` guards and resolvers are
deprecated. Instead, write guards as plain JavaScript functions and
inject dependencies with `inject` from `@angular/core`.
PR Close#47924
`errorHandler` and `malformedUriErrorHandler` were already deprecated in
the Router class public API. This change updates the deprecation to
extend to `RouterModule.forRoot` configuration options.
PR Close#48551
`withNavigationErrorHandler` is a close replacement for the
`RouterModule.forRoot.errorHandler` / `Router.errorHandler`.
It provides a quick, short way for users to define a function to handle
`NavigationError` events.
PR Close#48551
The argument against this feature:
Firstly, this isn't a navigation strategy that's very commonly desired.
Secondly, the hash strategy is a location configuration in `@angular/common`
and you can use it outside of the Router or with any custom router implementations.
So while the net effect of using the router is that it updates the URL which is
controlled by the strategy, this feature provider doesn't necessarily need to
exist in the Router. The location strategy is an application-wide configuration
and affects anything that imports `Location`. Of course, the Router does this
in a few places but plenty of other things might as well.
The argument for this feature:
* Discoverability. While `LocationStrategy` is technically in `@angular/common`, the most
common use-case _is_ when using it with the Router.
* Precedence in the `RouterModule.forRoot([], {useHash: true})`
* Precedence in other routing libraries (`createWebHashHistory` in VueJS
and `HashRouter` in React, for example)
* The implementation of `withHashLocation` is much more clear than
`useHash` was. You can look at the function and see that all it's
doing is adding the `HashLocationStrategy` to the providers list.
resolves#48295 / #47986
PR Close#48301
The Router constructor and `setupRouter` factory mainly exist as a
legacy configuration. Since the Router's creation, the style in Angular
has evolved quite a bit. This commit eliminates and cleans up some
unnecessary comlicated code paths related to the router
constructor/factory.
Note that there are edits to the `setupTestingRouter` that could be seen
as breaking. However, it is not for several reasons:
1. The function is documented as a factory function. If used as
documented, the parameters should match what's available in DI
2. The function is totally unused by the Router itself and is not used
in g3 either. I believe it was made publicApi by an error when
updating documentation annotations long ago.
PR Close#48215
This option has already been removed from availability externally. Internally, the last use has been removed
for the type on Route.loadChildren. More testing will be necessary to verify nothing relies on the other
code paths.
PR Close#48292
The events observable is part of the lifecycle of a navigation. This should
be contained in the NavigationTransitions class rather than the Router.
This Subject is re-exported by the router as an observable to maintain
the public API
PR Close#48202
The events observable is part of the lifecycle of a navigation. This should
be contained in the NavigationTransitions class rather than the Router.
This Subject is re-exported by the router as an observable to maintain
the public API
PR Close#48202
As mentioned in #46021, `canMatch` guards can replace `canLoad`. There
are slight differences between the two but the purpose of preventing
user access to feature modules is still achievable. There are several
reasons keeping `CanLoad` around is detrimental to the API surface:
* Lazy loading should not be an architectural feature of an application. It's an
optimization you do for code size. That is, there should not be an architectural
feature in the router to directly specifically control whether to lazy load something or
not based on conditions such as authentication. This slightly
different from the `canMatch` guard: the guard controls whether
you can use the route at all and as a side-effect, whether we download the code.
`CanLoad` only specified whether the code should be downloaded so `canMatch` is
more powerful and more appropriate.
* The naming of `CanLoad` will be potentially misunderstood for the `loadComponent` feature.
Because it applies to `loadChildren`, it feels reasonable to think that it will
also apply to `loadComponent`. This isn’t the case: since we don't need
to load the component until right before activation, we defer the
loading until all guards/resolvers have run.
* Unnecessary API surface bloat where two features (CanMatch and CanLoad) do
essentially the same thing. This affects code size for supporting two
nearly identical features as well as the learning and teaching journey
for them both.
* `CanLoad` guards have the downside of _only_ being run once to prevent
loading child routes. Once that passes and children are loaded, the
guard never runs again. As a result, developers need to always provide
_both_ canLoad and a canActivate in case the answer to the guard flips
back to `false`. This is not the case for `canMatch`, which will run
on every navigation.
DEPRECATED: CanLoad guards in the Router are deprecated. Use CanMatch
instead.
PR Close#48180
The router providers a configurable `onSameUrlNavigation` value that
allows developers to configure whether navigations to the same URL as
the current one should be processed or ignored. However, this only acts
as a default value and there isn't an API for easily overriding this for
a single navigation. Instead, developers are forced to update the value
of the property on the router instance and remember to reset it.
This feature fills a small gap in the Router APIs that enables
developers to accomplish the task of force reloading a bit easier.
Lengthy discussion about this here: #21115
PR Close#48050
This commit extracts the Router's navigation "transitions" to a separate
file. The goal here is twofold:
- Separate the Router's logic into more manageable chunks. Rather than
having to always edit a 1000+ line file, this separates different
pieces into smaller, more focused files
- More importantly, this sets the groundwork for separating the Router
state from the individual navigation transition. That is, it would be
good for the rxjs pipeline to _not_ modify the Router state but only
keep track of its own intermediate stages and emit outwards during
important events. The Router can subscribe to those events and make
updates to its state. This would theoretically allow us to swap those
implementations independently. That is, the Router's state management
can be changed without changing the transition handling and
vice-versa.
PR Close#48092
The Router currently silently ignores navigations for two reasons:
1. By default, same URL navigations are ignored. When this situation is
encountered, the navigation is ignored without any events
2. A `UrlHandlingStrategy` may ignore some URLs. For situations when the
strategy returns `false` for `shouldProcessUrl`, the Router silently
ignores the URL and updates its internal state without running
matching, guards, or resolver logic.
This commit adds new `NavigationSkipped` events for the above two situations.
PR Close#48024
None of the public properties of the `Router` are meant to be writeable.
They should all be configured using other methods, all of which have been
documented.
DEPRECATED: router writable properties
The following strategies are meant to be configured by registering the
application strategy in DI via the `providers` in the root `NgModule` or
`bootstrapApplication`:
* `routeReuseStrategy`
* `titleStrategy`
* `urlHandlingStrategy`
The following options are meant to be configured using the options
available in `RouterModule.forRoot` or `provideRouter`.
* `onSameUrlNavigation`
* `paramsInheritanceStrategy`
* `urlUpdateStrategy`
* `canceledNavigationResolution`
The following options are available in `RouterModule.forRoot` but not
available in `provideRouter`:
* `malformedUriErrorHandler` - This was found to not be used anywhere
internally.
* `errorHandler` - Developers can instead subscribe to `Router.events`
and filter for `NavigationError`.
PR Close#48006
The router has several default strategy implementations that are either
assigned via `new XStrategy()` property assignements or via the factory
function for the `Router` itself. These can be simplified by providing
the default implementation in the strategy `Injectable`. Custom
implementations can still override the defaults by specifying them in
the `providers` list.
Notably, this change also makes the `UrlHandlingStrategy` and
`RouteReuseStrategy` defaults injectable via DI rather than having to
indirectly grab them by going through the `Router` instance and
accessing the properties.
PR Close#48022
The `setupTestingRouter` function is not necessary. The `Router` can be fully
initialized through the DI configuration in the same way as it is in
production. Tests should use `provideRouter` or `RouterTestingModule`.
PR Close#48008
The public API for enabling this option has already been removed. This change
updates the internal Router code to not include any logic to support it.
PR Close#47933
This function is too similar in name to the new `provideRouter` and can
lead to misconfiguring the router. The function itself is just a small
wrapper around `{provide: ROUTES, multi: true, useValue: routes}`. It is
rarely necessary to provide routes in this way. The harm `provideRoutes`
will cause due to being only 1 letter away from `provideRouter` vastly
outweighs the benefit of a slightly shorter provider.
PR Close#47896
This commit switches `provideRouter()` to return the new
`EnvironmentProviders` wrapper type, preventing it from being accidentally
(or intentionally) included in `@Component.providers`.
PR Close#47669
This commit introduces a new type `EnvironmentProviders` which can be used
in contexts where Angular accepted `Provider`s destined for
`EnvironmentInjector`s. This includes contexts such as `@NgModule.providers`
and `Route.providers`.
The new type is useful for preventing such providers from accidentally
ending up in `@Component.providers`. It can be used as the return type of
provider functions (such as `provideRouter`) to enforce this safety.
Because `Provider` allows `any[]` nested arrays, the compile-time safety
provided by `EnvironmentProviders` is easily circumvented. However, the
runtime shape of `EnvironmentProviders` is not compatible with component
injectors and will result in a runtime error if it leaks through (NG0207).
A new function `makeEnvironmentProviders` is used to construct this new type
from an array of providers.
The existing `importProvidersFrom` operation previously returned a very
similar type `ImportedNgModuleProviders` which had the same goal. This
machinery is switched over to use the new `EnvironmentProviders` interface
instead (in fact, `ImportedNgModuleProviders` is now just an alias to
`EnvironmentProviders`).
PR Close#47669
This commit updates the `RouterLink` to extend the selector to also include `<a>` and `<area>` elements, which were previously matched by the `RouterLinkWithHref` directive. The code of the directives was merged together (since there was a lot of duplication) and this commit finalizes the merge. The `RouterLinkWithHref` becomes an alias of the `RouterLink` directive.
Now there is no need to import and use the `RouterLinkWithHref` class, the `RouterLink` directive will match any element that has the `routerLink` attribute.
DEPRECATED:
The `RouterLinkWithHref` directive is deprecated, use the `RouterLink` directive instead. The `RouterLink` contains the code from the `RouterLinkWithHref` to handle elements with `href` attributes.
PR Close#47630
The `relativeLinkResolution` is no longer supported. The only behavior
now is the correct behavior.
BREAKING CHANGE: `relativeLinkResolution` is no longer configurable in
the Router. This option was used as a means to opt out of a bug fix.
PR Close#47623
This commit updates the `RouterLinkWithHref` class to further align with the `RouterLink` class by removing own constructor from the `RouterLinkWithHref` class.
PR Close#47619
When using `loadChildren` or `loadComponent`, a common pattern is to pass
a function that returns a `Promise` from a dynamic import:
```typescript
{
path: 'lazy',
loadComponent: () => import('./lazy-file').then(m => m.LazyCmp),
}
```
The `.then` part of the expression selects the particular exported
component symbol from the dynamically imported ES module.
ES modules can have a "default export", created with the `export default`
modifier:
```typescript
@Component({...})
export default class LazyCmp { ... }
```
This default export is made available to dynamic imports under the well-
known key of `'default'`, per the ES module spec:
https://tc39.es/ecma262/#table-export-forms-mapping-to-exportentry-records
This commit adds a feature to the router to automatically dereference such
default exports. With this logic, when `export default` is used, a `.then`
operation to select the particular exported symbol is no longer required:
```typescript
{
path: 'lazy',
loadComponent: () => import('./lazy-file'),
}
```
The above `loadComponent` operation will automatically use the `default`
export of the `lazy-file` ES module.
This functionality works for `loadChildren` as well.
PR Close#47586
This commit breaks out the code in @angular/router needed to support the
deprecated `loadChildren` signature into a separate file, which reduces the
opportunity for conflicts when patching that behavior in g3.
PR Close#47586
This commit updates the 'name' of `RouterOutlet` to be an `Input` rather
than an attribute. Note that this change does not affect `[attr.name]=`
because those already would not have worked. The static name was only
read in the constructor and if it wasn't available, it would use
'PRIMARY' instead.
fixes#12522
BREAKING CHANGE: Previously, the `RouterOutlet` would immediately
instantiate the component being activated during navigation. Now the
component is not instantiated until the change detection runs. This
could affect tests which do not trigger change detection after a router
navigation. In rarer cases, this can affect production code that relies
on the exact timing of component availability.
PR Close#46569
This commit updates the `RouterLink` directive to use a renderer to apply an `href` attribute value. Previously, the `RouterLinkWithHref` used a host binding for this and that prevented further merge of the `RouterLink` and `RouterLinkWithHref` directives.
PR Close#47605
Currently, the `RouterLink` and `RouterLinkWithHref` classes share a lot of common code (with some special logic around handling `<a>`-related scenarios). This commit unifies the logic of the mentioned directives by moving the necessary handling to the `RouterLink` directive and making it a parent one for the `RouterLinkWithHref` class (i.e. class RouterLinkWithHref extends RouterLink).
This is the first step in upcoming unification to merge both directives and just keep `RouterLink` one (it'd be done in followup PRs).
PR Close#47500
ActivatedRouteSnapshot data gets mutated in the resolve phase of the Router. The title is assigned as part of this.
As a result, the title must be a getter in order to pick up the value that was note available during the class creation.
fixes#47459
BREAKING CHANGE: The title property is now required on ActivatedRouteSnapshot
PR Close#47481
Paul Gschwendtner