Currently when ESBuild bundles an application importing from
`@angular/core`, the signals library will be discovered during
export analysis. ESBuild will come across the constants for the reactive
signal graph- and end up considering some of these as side-effects given
the pattern of using a spread assignment for extending from e.g.
`REACTIVE_NODE` (a similar issue may occur if we e.g. extend from the
computed reactive node).
See more details on the issue: https://github.com/evanw/esbuild/issues/3392
Even though, ESBuild preserves these constants now, and all of its
dependencies— Terser will consider these as side-effect free and
eliminate these constants. This may require multiple passes though, and
might not be sufficient, depending on the chain of reactive node
extensions. E.g. in the signals branch we noticed some constants
unnecessarily being preserved.
PR Close#51809
This commit adds support for zone.js 0.14.0 and drops support for older versions
BREAKING CHANGE: Angular now required `zone.js` version `~0.14.0`
PR Close#51774
upgrade the warning for lazy-loaded lcp images when using NgOptimizedImage to an error
BREAKING CHANGE:
Previously when NgOptimizedImage directive detected that an LCP image is lazy-loaded, a console warning was produced. Now the directive throws an error to make it more discoverable in a console. If you receive this error, refer to this guide for additional information: https://angular.io/guide/image-directive#step-4-mark-images-as-priority
PR Close#51748
Terser does not treat `Symbol` as side-effect free- so if we end up with
a symbol export being loaded, it will result in the symbol being
retained.
We noticed this in the signals prototyping where symbols exported
from `computed` ended up appearing in symbol bundling tests.
PR Close#51776
This commit moves the destroy logic from 'effect' in the lower-level
'watch' so this implementation is shared among varius watch implementations.
PR Close#51757
Certain code patterns and tools in Google (and possibly 3P world) lead to the situation that a component is bootstrapped/rendered without its ng-module being loaded in the browser. Technically speaking this should be an anti-pattern since the ng-module could contain some runtime logic (e.g., providing something, calling some services, etc) and its not being loaded leads to unexpected behaviour. However, in many cases ng-module is an empty class and its only usage is for providing scope, and since in AoT full compilation mode we already hard-code dependencies into components so we can get away with not loading the ng-module. But in AoT local compilation mode it is not possible to get away since the component's dependencies are computed in runtime and the presence of the corresponding ng-module in the browser is needed. For this reason in this change it is forbidden to attempt to render a component without first loading its ng-module in local compilation mode and an explicit error message is created to make this situation clear. This error message can help with catching such cases when running TGP in Google.
It would be an interesting question as to whether to ban this situation in full compilation mode as well, as it is error prone and these errors are sometimes very hard to debug.
PR Close#51726
This commit adds explicit type annotations to the reactive node prototype objects,
such that the prototypes are type-checked against the interface they are supposed
to (partially) implement. This also allows IDEs to better track usages of reactive
node properties, improving code navigation.
PR Close#51722
When a producer is no longer used, the consumer has to update its internal data structure
that keeps track of all producers. There used to be an issue where only half of the stale
producers would actually be removed from this data structure, as the intended upper bound
of the number of producers to remove would decrease with each removed producer, therefore
not reaching all producers that should be removed from the data structure.
This commit fixes the issue by truncating the arrays directly, without going through
individual `pop` operations. An assertion that would catch the inconsistent state in
the internal data structures of the signal graph has been introduced.
PR Close#51722
BREAKING CHANGE: Node.js v16 support has been removed and the minimum support version has been bumped to 18.13.0.
Node.js v16 is planned to be End-of-Life on 2023-09-11. Angular will stop supporting Node.js v16 in Angular v17. For Node.js release schedule details, please see: https://github.com/nodejs/release#release-schedule
PR Close#51755
Adds support for passing in `@Component.styles` as a string. Also introduces a new `styleUrl` property on `@Component` for providing a single stylesheet. This is more convenient for the most common case where a component only has one stylesheet associated with it.
PR Close#51715
Currently internally Angular has some customized tsconfig files, because we don't align with the tsconfig of the rest of g3. These changes enable `noImplicitReturns` and `noPropertyAccessFromIndexSignature` to align better with the internal config.
PR Close#51728
Previously effects were queued as they became dirty, and this queue was
flushed at various checkpoints during the change detection cycle. The result
was that change detection _was_ the effect runner, and without executing CD,
effects would not execute. This leads a particular tradeoff:
* effects are subject to unidirectional data flow (bad for dx)
* effects don't cause a new round of CD (good/bad depending on use case)
* effects can be used to implement control flow efficiently (desirable)
This commit changes the scheduling mechanism. Effects are now scheduled via
the microtask queue. This changes the tradeoffs:
* effects are no longer limited by unidirectional data flow (easy dx)
* effects registered in the Angular zone will trigger CD after they run
(same as `Promise.resolve` really)
* the public `effect()` type of effect probably isn't a good building block
for our built-in control flow, and we'll need a new internal abstraction.
As `effect()` is in developer preview, changing the execution timing is not
considered breaking even though it may impact current users.
PR Close#51049
The View Transitions API enables easy animations when transitioning between different DOM states. This commit adds an opt-in feature to the Router which runs the component activation and deactivation logic in the document.startViewTransition callback. If the browser does not support this API, route activation and deactivation will happen synchronously.
resolves#49401
PR Close#51314
Improves the error handling story for after*Render by delegating errors to an ErrorHandler, so that one failure does not break every callback.
PR Close#51662
For redirects, most of the logic between wildcard and regular redirects
is identical. Combine what can be combined there. Once that was done,
the first part of the matching for redirects looks almost identical to the regular
matching. Create helpers and combine simplify code where possible there
as well.
PR Close#51543
This commit updates the logic to add `prefetch on idle` support for defer blocks. Previously, the `on idle` logic was already implemented for the main loading and rendering. This commit reuses the same logic to bring it to the prefetching mechanism.
PR Close#51629
This flag allows message event listeners to prevent callbacks from executing within the NgZone if they contain a special `__ignore_ng_zone__` flag.
This functionality is built with Angular DevTools in mind, where it prevents an infinite change detection loop in inspecting applications that have message event listeners:
CD -> Inspected app emits componentTreeDirty event to DevTools -> DevTools emits event to get new component Tree from Inspected app -> Inspected app message event listener fires -> CD
PR Close#51339
These changes build on top of #51514 to add support for advanced expressions inside the `track` parameter of `for` loop blocks. There are two different outputs that the compiler can generate:
1. If the tracking function only references the item or `$index`, the compiler generates a pure arrow function as a constant references in the `repeaterCreate` instruction.
2. If the tracking function has references to properties outside of the `for` loop block, the compiler will rewrite those references to go through `this` and generate a function declaration. The runtime will `bind` the declaration to the current component instance so that the rewritten `this` references are resolved correctly.
Advanced tracking expression come with the following limitations to ensure the best possible performance:
1. They can only reference the item, `$index` and properties directly on the component instance. This means that there'll be an error when accessing this like local template variables and references. While we could get this to work, we would have to traverse the context tree at runtime which will degrade the performance of the loop, because it's a linear time operation that is performed on each comparison. Furthermore, allowing local references would require us re-evaluate the list when any one of them has changed.
2. Pipes aren't allowed inside the tracking function.
3. Object literals and pipes used inside the tracking expression will be recreated on each invocation.
PR Close#51618
Adds an instruction that allows us to access the containing component instance directly instead of having to traverse the context tree. This will be necessary for the tracking function of `for` loop blocks.
PR Close#51618
This commit updates the runtime implementation of defer blocks to avoid their triggering on the server. This behavior was described in the RFC (https://github.com/angular/angular/discussions/50716, see "Server Side Rendering Behavior" section): only a placeholder is rendered on the server at this moment. This commit also updates the logic to make sure that the placeholder content is hydrated after SSR.
PR Close#51530
This commit fixes a bug in the change detection algorithm that would
ignore the `OnPush`/dirty flag of a component's host when it is created
dynamically. That is, `OnPush` components that were not marked dirty but
were created as embedded views would have their host bindings and `ngDoCheck`
function always run even if they were not dirty.
BREAKING CHANGE: `OnPush` components that are created dynamically now
only have their host bindings refreshed and `ngDoCheck run` during change
detection if they are dirty.
Previously, a bug in the change detection would result in the `OnPush`
configuration of dynamically created components to be ignored when
executing host bindings and the `ngDoCheck` function. This is
rarely encountered but can happen if code has a handle on the
`ComponentRef` instance and updates values read in the `OnPush`
component template without then calling either `markForCheck` or
`detectChanges` on that component's `ChangeDetectorRef`.
PR Close#51356
This commit refactors the router internals to track state inside a separate
`StateManager`. This helps open the door to managing the state in
different, swappable providers. The current interface needed by the
Router is: `currentUrlTree`, `rawUrlTree`, `browserUrlTree`,
`routerState`, and `handleNavigationEvent` (other properties are because
some router properties are writeable when they shouldn't be). This is
a suprisingly small interface and can hopefully be made smaller in the
future (i.e. removing 1 or more of the 3 tracked of `UrlTree`s).
PR Close#48481
This commit switches the signals library from a bidirectional symmetric
dependency graph using weak references, to a bidirectional _asymmetric_
graph which uses strong references. This is made possible with a reference
counting algorithm which only tracks producer -> consumer references for
effect-like "live" consumers, preventing memory leaks.
The new algorithm should be simpler and faster than the previous
implementation as weak references are fairly slow to create and traverse.
A tradeoff is that non-live consumers must now poll their producers when
read, as they cannot rely on dirty notifications.
As part of this refactoring, the `ReactiveNode` class is replaced with an
interface instead, and methods are moved to standalone functions. This is
paired with instantiating individual signals/computeds via `Object.create`
against a prototype node which contains static or initial values. This
technique, in conjunction with the rest, greatly improves the performance
of node creation.
PR Close#51226
In preparation for adding support for phases to after*Render, which will increase the implementation size, this commit splits out the optional logic so that it can be tree-shaken and dynamically loaded.
PR Close#51541
The data `Observable` is not updated unless there have been changes to
the object. The current diffing does not look at `symbol` keys of the
object but the `title` property is stored as a private `symbol`. This
commit updates the object diffing to include symbols.
fixes#51401
PR Close#51561
- Emphasized the importance of using the same InjectionToken instance for both provider and injection call.
- Added examples to illustrate correct usages to prevent NullInjectorError.
PR Close#51386
Fixes that there was code duplication between the primary entry-point,
the testing entry-point and the rxjs-interop entry-point.
This code duplication resulted in additional code size (really
neglibible here because rxjs-interop did not duplicate large parts of
core, and `testing` is not used in production).
On the other hand though, the duplication resulted in a subtle JIT
dependency tracking issue due to the `depsTracker` no longer being a
singleton. This caused test failures as in:
https://github.com/angular/angular/pull/51415.
PR Close#51500
With the new control flow and defer blocks it'll be common for several template instructions to be declare one after another. These changes add support for chaining to the `template` instruction which will allow us to save some bytes.
PR Close#51546
Adds the initial implementation to generate the instructions for the `for` loop block.
**Note:** the expressions we support in the `track` paramateter are currently limited to tracking by identity or index, or a specific property of the item. Supporting more advanced expression will require additional work that I'll do in a follow-up PR.
PR Close#51514
This commit fixes an issue where serialization of a view container fails in case it uses a component host as an anchor. This fix is similar to the fix from #51247, but for cases when we insert a component (that acts as a host for a view container) deeper in a hierarchy.
Resolves#51318.
PR Close#51456
This commit adds runtime implementation of a basic preloading mechanism for defer blocks. The base prefetching logic invokes a dependency loading function (generated by the compiler) when a corresponding `prefetch when` condition is triggered. The `prefetch on` triggers would be implemented in followup PRs.
We plan to explore additional prefetching techniques and will followup with more PRs later (based on the research).
PR Close#51529
This commit adds an initial implementation of the `{#defer}` block runtime, which supports the `when` conditions. More conditions and basic prefetching support will be added in followup PRs.
PR Close#51347
Rather than the navigation transitions managing Router state, this
commit updates the Router to manage its own state based on ongoing
transition events. In the future, this can be abstracted even further to
have a totally separate class that manages the Router state. This would
allow the potential for swapping state manager implementations rather
than having to implement all types of state management in a single place.
One finding during tests was that unexpected errors thrown by the state management code moved
to the Router here will no longer be caught by the transition pipe's `catchError`.
This only includes calls to the following public Api methods:
* `go`, `replaceState`, `historyGo`, `isCurrentPathEqualTo` on `Location`
* `UrlSerializer.serialize`
* `UrlHandlingStrategy.merge`.
None of these methods should throw if the router is expected to function.
These might throw when tests include incomplete mocks, which is not
supported, or in cases where the actual browser methods like
`replaceState` would throw. This will already result in unexpected/unsupported
behavior. The failure case here is now arguably better - the navigation
itself still completes but the state update (either updating Router
internal state or updating the browser URL) fails separately and is
unhandled.
PR Close#48427
Updates the control flow tests to use the compiler instead of manually-written instructions. Also adds a couple of tests that I was using along the way to verify that things work as expected.
PR Close#51380
Paul Gschwendtner