The `canceledNavigationResolution: 'computed'` option does not correctly
assign page IDs or restore them when redirects result in navigating to
the current URL. This change ensures that the page IDs are still
incremented and restored correctly in this scenario.
PR Close#49793
The previous fix for squashing empty children didn't quite work when the
existing route had segments. The result would be that the segments from
the existing route were dropped from the final URL.
PR Close#49691
The navigation transition clears the current navigation in the finalize
operator of the current navigation Observer. This commit both completes
the `bootstrapDone` observable and updates the transition to only take 1
emit from the completed navigation. Either of these changes on their own
would fix the issue. The latter is a preventative measure in case a
mistake like the former is made again.
fixes#49567
PR Close#49572
The preloading strategy did not handle a `loadComponent` on a route with
a static `children`. It only preloaded children if they were also
`loadChildren` or both were not lazy loaded.
fixes#49558
PR Close#49571
Add a more specific error message when defining a lazy-loaded route using
`loadComponent` and passing it a NgModule instead of a standalone component,
when the user should actually be using `loadChildren`.
PR Close#49164
The details of this commit are pretty thoroughly described in the tests
and code comments. In short, it is sometimes ambiguous where to apply commands in
a `SegmentGroup` tree that is created from an `ActivatedRoute` when
dealing with empty paths. This adjusts the strategy to tolerate more
ambiguity while still allowing developers to be explicit.
This is a fix-forward for b/265215141
PR Close#48734
`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 ESM js files need to be referenced, and the router tests rely
on `async/await` with change detection- so a special rule is required
that downlevels `async/await` to generators (similar to the Angular CLI)
PR Close#48521
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
For every `ts_library` target we expose a shorthand that grants
access to the JS files because `DefaultInfo` of a ts library
only exposes the `.d.ts` files.
We rename this away from `es2015` since in practice it's a much
higher target these days. Additionally we no longer use the devmode
output but rather use the prodmode output which has the explicit
`.mjs` output- compatible with ESM.
PR Close#48521
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
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 `initNavigated` flag is not necessary. The way `enabledBlocking`
works is by blocking any navigations from finishing until the
application has bootstrapped. The bootstrap is unblocked as soon as we
hit the `afterPreactivation` step. After that point,
`afterPreactivation` is always unblocked because the `bootstrapDone`
Subject is completed/stopped. There is no need for an additional
variable to track this information.
PR Close#48064
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
This commit makes several updates to the bootstrap tests of the Router.
The change that made the real difference was providing
`MockPlatformLocation`. Other changes include:
* using `async/await` instead of `done`
* not providing APP_BASE_HREF (recent router refactors made this unnecessary)
* using some functional guards to simplify some tests
PR Close#48176
If `history.pushState()` or `history.replaceState()` were called manually without including the `navigationId` field the state was being incorrectly discarded - that logic was for maintaining the original behavior of `NavigationStart.restoredState`.
Improves on #28176, fully fixes#28108, see also #28954
PR Close#48033
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
Due to being only 1 letter away from `provideRouter`, it is quite
possible that developers may accidentally use `provideRoutes` rather
than `provideRouter` in the `boostrapApplication` function. This change
will warn developers when `provideRoutes` is used without the `Router`.
PR Close#47896
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
Currently, the scroll event is fired immediately after the
`NavigationEnd`. However, this is problematic because a change detection
has not been able to run, meaning that Angular will not yet have run the update
block of the component templates being rendered as part of the navigation.
This change delays the scroll event using a `setTimeout`, which will
allow Angular's change detection to run before the scroll restoration is
performed.
fixes#24547
PR Close#47563
This commit adds the `provideLocationMocks()` function that returns mocks for the `Location` and `LocationStrategy` classes. This function can be used in tests to configure an environment where it's possible to fire simulated location events (helpful when testing Router configuration).
PR Close#47674
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
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 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
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
This commit creates and exposes the APIs required to use the Angular Router without importing `RouterModule`.
The newly added APIs are tree-shakable and you can add features using special functions rather than using `ExtraOptions` to control the providers via an internal switch in Router code.
```
const appRoutes: Routes = [];
bootstrapApplication(AppComponent,
{
providers: [
provideRouter(appRoutes,
withDebugTracing(), // enables debug tracing feature
withInMemoryScrolling() // enables scrolling feature
]
}
);
```
This "features" pattern allows for router behavior to evolve in a backwards compatible and tree-shakable way in the future. This approach also makes features more discoverable.
The newly added APIs can be used in any application today (doesn't require an application to be bootstrapped using standalone-based APIs).
Note: APIs added in this commit are released in the "Developer Preview" mode, read more about this mode in Angular docs: https://angular.io/guide/releases#developer-preview
PR Close#47010
Add wider typings to setter of preserveFragment, skipLocationChange and replaceUrl inputs of routerLink directives and coerce them to boolean
PR Close#47101
The current Router APIs require guards/resolvers to be present in the DI tree. This is because we want to treat all guards/resolvers equally and some may require dependencies. This requirement results in quite a lot of boilerplate for guards. Here are two examples:
```
const MY_GUARD = new InjectionToken<any>('my_guard');
…
providers: {provide: MY_GUARD, useValue: () => window.someGlobalState}
…
const route = {path: 'somePath', canActivate: [MY_GUARD]}
```
```
@Injectable({providedIn: 'root'})
export class MyGuardWithDependency {
constructor(private myDep: MyDependency) {}
canActivate() {
return myDep.canActivate();
}
}
…
const route = {path: 'somePath', canActivate: [MyGuardWithDependency]}
```
Notice that even when we want to write a simple guard that has no dependencies as in the first example, we still have to write either an InjectionToken or an Injectable class.
With this commit router guards and resolvers can be plain old functions.
For example:
```
const route = {path: 'somePath', component: EditCmp, canDeactivate: [(component: EditCmp) => !component.hasUnsavedChanges]}
```
Additionally, these functions can still use Angular DI with `inject` from `@angular/core`.
```
const route = {path: 'somePath', canActivate: [() => inject(MyDependency).canActivate()]}
```
PR Close#46684
In Angular 14, we introduced the `loadComponent` API for a `Route` to
allow lazy loading of a routed component in addition to the existing
`loadChildren` which allows lazy loading of child routes. As a result,
the `preload` method of the `PreloadingStrategy` needs to sometimes be
called even when there is a `canLoad` guard on the `Route`. `CanLoad`
guards block loading of child routes but _do not_ block loading of the
component.
This change updates the conditional checks in the internal preloader to
skip calling the `PreloadingStrategy.preload` when there is only a
`loadChildren` callback with a `canLoad` guard an no `loadComponent`.
In this case, the callback passed to the `preload` method is already
effectively a no-op so it's not necessary to call it at all.
resolves#47003
PR Close#47007
This commit refactors the `NgModuleRef` implementation to drop functions required by the `EnvironmentInjector` interface. Previously the idea was that the `NgModuleRef` can act as an Injector to facilitate easier transition to standalone. However, from the mental model perspective, the `NgModuleRef` has the `injector` field, which is the correct injector reference and can be used is needed as an `EnvironmentInjector`.
PR Close#46896
Karma jasmine updated the `jasmine-core` dependency. Jasmine is now more
strict when:
* The done callback is invoked multiple times
* The done callback is used, while a promise is also returned
* The done callback is treated as error when e.g. a number is returned
as first argument. This was the case with `requestAnimationFrame`.
The Router transition observable pipe keeps an outer reference to a `t`
variable for use in the `catchError` operator. However, this variable is
not updated with intermediate state. This commit fixes that so the
`catchError` can access properties that get updated in intermediate
states. Specifically, `RouterStateSnapshot` in the `NavigationError` for
now but could be more in the future.
PR Close#46842
Before this commit, the `NavigationCancellationCode` would always be set
to `Redirect` when encountering a "navigationcancelingError". However,
this error can also be thrown when `CanLoad` guars reject. This commit
ensures these cancellation errors have a code as well so this mistake
cannot be made again.
PR Close#46752
The existing logic does something similar but in a more roundabout way.
It reads _the whole array_. If it encounters a pending value, it ignores
the remaining ones. If it hasn't encountered a pending value by the time
it hits false/UrlTree, it returns that result.
The new logic is the same, but reverses what we're looking for. Instead
of processing the whole array, we stop when we encounter an initial
value. When we encounter one that isn't `true`, that gets returned. If
we get to the end and everything was `true`, return `true`.
PR Close#46745
PR Close#46745
Andrew Scott