This change adjust the equality comparator used by NgSwitch - now it
defaults to === from previously used ==. This change is based on the
following reasoning:
- align behaviour with the built-in switch block);
- improve performance (avoid type coercion);
- enable better type-checking.
BREAKING CHANGE:
the NgSwitch directive now defaults to the === equality operator,
migrating from the previously used ==. NgSwitch expressions and / or
individual condition values need adjusting to this stricter equality
check. The added warning message should help pinpointing NgSwitch
usages where adjustements are needed.
Fixes#33873
PR Close#51504
This change removes the `mutate` method from the `WritableSignal` interface and
completely drops it from the public API surface.
The initial API proposal for Angular signals included the mutate method, allowing
in-place modification of JS objects, without changing their references (identity).
This was based on the reasoning that identity change on modification is not necessary
as we can send the “modified” notification through the signals graph.
Unfortunately the signal-specific change notification is lost as soon as we read
signal value outside of a reactive context (outside of a reactive graph).
In other words - any code outside of the Angular signals library can’t know
that an object is modified.
Secondly, to make the mutate method work, we’ve defaulted the signal value equality function
to the one that considers non-primitive values as always different.
This is unfortunate for people working with immutable data structures
(this is notably the case for the popular state management libraries)
as the default equality function de-optimizes memoization in computed,
making the application less performant.
Given the above reasons we prefer to remove the mutate method in the signals library -
at least for now. There are just too many sharp edges and tradeoffs that we don’t fully
understand yet.
BREAKING CHANGE:
The `mutate` method was removed from the `WritableSignal` interface and completely
dropped from the public API surface. As an alternative please use the update method and
make immutable changes to the object.
Example before:
```typescript
items.mutate(itemsArray => itemsArray.push(newItem));
```
Example after:
```typescript
items.update(itemsArray => [itemsArray, …newItem]);
```
PR Close#51986
This commit reorganizes the Angular code a bit, and moves signals into a
newly defined `@angular/core/primitives` location. This will be used inside
g3 to allow non-Angular targets to depend on the signals core without
incurring a dependency on the whole framework.
PR Close#51986
This commit refactors the signals API surface of Angular out of the
//packages/core/src/signals package. This is done in preparation of moving
the core signals package into a new 'primitives' package that's decoupled
from the public API.
PR Close#51986
The code already supports `DatePipeConfig` to have only some properties
set, and not all. But the typing disallows it. This aligns the typing
with the code.
PR Close#51287
`provideClientHydration()` accepts new `HydrationFeature` : `HttpTransferCacheOptions`.
`withHttpTransferCacheOptions` accepts an option object:
* `includeHeaders` : list of headers entries to keep in the cache with the request
* `filter` a callback to determine if a request should be cached
* `includePostRequests`: to include POST requests in the allowed methods
Implements some of the features requested in #50117
PR Close#52029
Fixes that the compiler wasn't picking up pipes used inside defer block triggers as dependencies. We had implemented the `visitDeferredTrigger` visitor method, but it wasn't being called, because we weren't going through the `visitAll` method of the deferred block. We don't use `visitAll`, because child nodes have to be processed differently than the connected blocks and triggers.
Fixes#52068.
PR Close#52071
The `_enabledBlockTypes` config option was removed recently, since we've enabled @-syntax by default. This commit removes `_enabledBlockTypes` references from the `compiler-cli` test cases.
PR Close#52066
This commit adds a logic to produce a warning in case HttpClient doesn't use fetch during SSR.
It's recommended to use `fetch` for performance and compatibility reasons.
PR Close#52037
This change removes the `mutate` method from the `WritableSignal` interface and
completely drops it from the public API surface.
The initial API proposal for Angular signals included the mutate method, allowing
in-place modification of JS objects, without changing their references (identity).
This was based on the reasoning that identity change on modification is not necessary
as we can send the “modified” notification through the signals graph.
Unfortunately the signal-specific change notification is lost as soon as we read
signal value outside of a reactive context (outside of a reactive graph).
In other words - any code outside of the Angular signals library can’t know
that an object is modified.
Secondly, to make the mutate method work, we’ve defaulted the signal value equality function
to the one that considers non-primitive values as always different.
This is unfortunate for people working with immutable data structures
(this is notably the case for the popular state management libraries)
as the default equality function de-optimizes memoization in computed,
making the application less performant.
Given the above reasons we prefer to remove the mutate method in the signals library -
at least for now. There are just too many sharp edges and tradeoffs that we don’t fully
understand yet.
BREAKING CHANGE:
The `mutate` method was removed from the `WritableSignal` interface and completely
dropped from the public API surface. As an alternative please use the update method and
make immutable changes to the object.
Example before:
```typescript
items.mutate(itemsArray => itemsArray.push(newItem));
```
Example after:
```typescript
items.update(itemsArray => [itemsArray, …newItem]);
```
PR Close#51821
Two key refactors to enable deeper language service support for blocks:
(1) We now generate accurate source spans for the various block types. Additionally, all the top-level source spans for a block are now *inclusive* of all the connected or descending blocks. This helps the language service visit connected blocks.
(2) The language service's template visitor was previously skipping over the AST nodes corresponding to several block types. We are now careful to visit all such nodes.
PR Close#52038
Fixes a bug with REMOVE_STYLES_ON_COMPONENT_DESTROY when a component is destroyed and re-created, its previous stylesheets will not be re-used and instead a new stylesheet will still be created each time.
PR Close#52019
This commit adds runtime code to support `after` and `minimum` parameters in the `@placeholder` and `@loading` blocks. The code uses the `TimerScheduler` service added earlier for `on timer` triggers.
PR Close#52009
Adds an `UnknownBlock` node to the Ivy AST to represent blocks that haven't been recognized by the compiler. This will make it easier to integrate blocks into the language service.
PR Close#52047
Adds some logic to treat incomplete blocks as empty blocks so that we can recover from them. Also logs an error about the incomplete block.
PR Close#52047
Updates the lexer to parse blocks as incomplete, instead of throwing errors. This will allow us to better handle them further down in the pipeline.
PR Close#52047
Revert (with improvements of): dcf18dc74c
We recently landed a change that allows `toSignal` to be called
from within reactive contexts (e.g. `effect`/`computed`). After
more thorough investigatio and consideration with the team, we
feel like allowing `toSignal` to be called in such contexts is
encouraging non-ideal / hard-to-notice code patterns.
e.g. a new subscription to an observable is made every time `toSignal`
is invoked. There is no caching done here. Additionally, multiple new
subscriptions can trigger unintended side-effects- that may slow down
the app, result in incorrect/unexpected behavior or perform unnecessary
work.
Users should instead move the `toSignal` call outside of the `computed`
or `effect` and then read the signal values from within their `computed`. e.g.
```ts
computed(() => {
const smth = toSignal(coldObservable$)
return smth() + 2;
}
```
--> should instead be:
```ts
const smth = toSignal(coldObsverable$);
computed(() => smth() + 2);
```
In cases where a new subscription for each invocation is actually intended, a manual
subscription can be made. That way it's also much more obvious to users
that they are triggering side-effects every time, or causing new
subscriptions.
PR Close#52049
GC and render events can happen _while_ running scripts as well as
outside of the script blocks. The new metric entries capture both
the gc and render time happening in the scrip blocks.
PR Close#50771
This adds API doc extraction for interfaces, largely using the same code paths for classes. The primary difference between classes and interfaces is that classes have member _declarations_ while interfaces have member _signatures_. This largely doesn't matter for the purposes of extraction, but the types are distinct with no common base types, so we have to do a fair amount of type unioning and aliasing.
PR Close#52006
Upgrade the existing warning so it now logs an error instead, when an LCP element is determined to not be usings the `priority` attribute. Error is logged, not thrown.
PR Close#52004
When the `ts.Project` creates the language service plugin (in this case,
the Angular Language Service), it sets the project's language service to
the new language service returned by the plugin create:
https://sourcegraph.com/github.com/microsoft/TypeScript@b12af0fa2bbd4b015e59adcfb49988cea7f919a1/-/blob/src/server/project.ts?L2035-2044
The project may be reloaded in response to various events, such as a
change to the tsconfig file, which then recreates the plugin. When this
happens, the language service that gets passed to the plugin `create`
function will not be the typescript language service, but rather the
previous instance of the new language service returned by the last call
to `create`.
This commit ensures that subsequent calls to `create` for the
`NgLanguageService` plugin for a project after the first call are able
to retrieve and hold on to the _TypeScript_ language service.
fixes https://github.com/angular/vscode-ng-language-service/issues/1923
PR Close#51912
A couple tests were already passing, and just needed to be enabled. This includes tests pertaining to:
* ng-template
* host binding styling slots
* and host animation bindings
* some literal tests (which were missing some $foo$ escaped names)
We add pipeline-specific versions of the following tests, and enable them:
* A local refs test. The consts for the element attributes and the consts for local reference are collected in the reverse order, but the emitted template is functionally the same.
* A safe accesstest. Consider the expression `$any(val)?.foo`. `TemplateDefinitionBuilder` extracts a temporary variable: `($tmp_0_0$ = $ctx$.val) == null ? null : $tmp_0_0$.foo`. It presumably does this because it considers the `$any(...)` to be a function call. However, this is not a real call, so Template Pipeline safely ignores it and declines to generate a temporary.
* Another local refs test. AttributeMarker.Template is emitted at the end of the const array (instead of the middle)
PR Close#51950
Consider an `ng-template` which is generated as a result of a structural directive:
```
<div *ngFor="let inner of items"
(click)="onClick(inner)"
[title]="getTitle()"
>
```
This should logically expand into something like the following:
```
<ng-template [ngForOf]="..." >
<div (click)="..." [title]="..."></div>
</ng-template>
```
Note that the `(click)` handler and the `[title]` property are only present on the inner div, *not* on the enclosing generated `ng-template`.
Previously, Template Pipeline would place these bindings on *both* the tempate and the inner element.
However, we can't just remove them completely, because these bindings should still be matchable on the generated `ng-template` (which is very surprising, but nonetheless true).
We resolve this issue with two improvements:
(1) The ingestion step is now much smarter about determining not only if a binding is on a template element, but whether it actually targets that template element.
(2) We use `ExtractedAttributeOp` directly, rather than going through `BindingOp`, to cause the `ng-template` to still receive these bindings in its `consts` array for matching purposes.
PR Close#51950
For components, the parser already extracts the `important` property (and it is later disregarded). However, because host bindings use a totally separate parsing code path, this was never happing for host bindings.
Here, we add some code to the host style parsing phase to drop the `!important` suffix.
We could solve this category of problems for good by parsing host bindings with the same code as template bindings.
PR Close#51950
Previously, we always generated temporary variable declarations at the beginning of each view's update block. This is wrong, for two reasons:
1. Temporaries can be used in the create block
2. When listeners use temporaries, we should declare them inside the listener.
Now, we always place temporaries at the beginning of the enclosing OpList, and recursively try to generate them when we find a listener.
PR Close#51950
Currently, if there are 2 nested @defer blocks with the same dependency, Angular throws an error at runtime to indicate that there was a duplicate component def in the registry. This commit updates the logic to only append dependencies when they didn't previously exist in the registry.
PR Close#51964
#51885 patched a call site that threw an error but there were 2 others call that needed to be wrapped in the same way by a try/catch.
`initializeFully` is part of the calls in `responseWith(handleFetch)`.
Same #51885, throwing `SwCriticalError`allows the driver to fallback to `safeFetch` and ensure `responseWith` doesn't fail.
Fixes#50378
PR Close#51960
Currently the TCB for aliased `if` blocks looks something like this:
```
// Markup: `@if (expr; as alias) { {{alias}} }
if (block.condition) {
var alias = block.condition;
"" + alias;
}
```
The problem with this approach is that the type of `alias` won't be narrowed. This is something that `NgIf` currently supports.
These changes resolve the issue by emitting the variable outside the `if` block and using the variable reference instead:
```
// Markup: `@if (expr; as alias) { {{alias}} }
var alias = block.condition;
if (alias) {
"" + alias;
}
```
PR Close#51952
The context of an embedded view ref at some point was switched from a
getter to an actual assignable property. This is something we reverted
with the previous commit as it introduces additional complexity for our
generated code (in terms of closures capturing the `ctx`).
This change impacted the template outlet code because we actively relied
on swapping out the full context if the user changes it. Previousl,
before we allowed to swap out the context (in v16), we mutated the
initial view context if it didn't change structurally- and in other
cases the view was re-created. We improved this performance aspect with
the changes to allow for the context to be swapped out + actually also
fixed a bug where the initial context object was mutated and the user
could observe this change.
This commit adjusts for context not being replacable- while still
keeping the bugs fixed and preserving the performance wins of not
having to destroy/re-create the view whenever the context changes.
Benchmarks: https://hackmd.io/J0Ci_JzxQ0K1AA1omXhIQQ
PR Close#51887
This partially reverts commit a3e17190e7
and deprecates behavior added.
The context of an embedded view ref at some point was switched from a
getter to an actual assignable property. This is something we revert
as it introduces additional complexity for our generated code
(in terms of closures capturing the `ctx`), creates technical
limitations for Angular's internals and the usage pattern is rarely
used (and can be addressed via simple assignments, `Object.assign` or
the use of a proxy if replacing the full context object is still
desirable)
DEPRECATED: Swapping out the context object for `EmbeddedViewRef`
is no longer supported. Support for this was introduced with v12.0.0, but
this pattern is rarely used. There is no replacement, but you can use
simple assignments in most cases, or `Object.assign , or alternatively
still replace the full object by using a `Proxy` (see `NgTemplateOutlet`
as an example).
Also adds a warning if the deprecated
PR Close#51887
When adding a new view flag, you currently need to adjust the last number of the last
3 flags. All of these share the same number so the shifting ones can just use
the base-10 IndexWithinInitPhaseShift.
PR Close#51839
Andrew Scott