The `generateUniqueIdentifier` helper relies on the internal `identifiers` property of a `ts.SourceFile`.
As this is not a public-facing API, it is not included in the externs provided to Closure Compiler.
Consequently, it is susceptible to being renamed during advanced optimizations, which would lead to runtime failures.
To prevent this, the property is now accessed via a string literal (`sf['identifiers']`). This change ensures that
the property name is preserved through the compilation process. An explanatory comment has been added to clarify the
necessity of this approach for future reference.
PR Close#62517
Currently when a diagnostic happens in a template, we always say "Error occurs in ...", however we have other types of diagnostics as well.
PR Close#62479
The mock file system currently relies on some NodeJS logic. This blocks
the convenient use of the mock file system in the browser. The file
system is useful for creating tests to verify e.g. that our transforms
can work in the browser.
PR Close#62493
The static fields may be part of advanced compilations where the
receiver would change as part of the static field collapsing
optimization.
The optimization requires us to use the explicit class name, over
`this` that would change to e.g. `globalThis`.
PR Close#62493
The commit introduce the distinction between
- class like decorators (like Component, Interface, NgModule. They are formatted like classes/interfaces, with each attribute being documents.
- function like decorators (Attribute, Host, Optional...)
PR Close#60411
Currently when Tsurge runs in g3, it creates a bare bones Angular
compiler plugin. The tsconfigs from compilation units may set options
like "useHostForImportGeneration", but the Ngtsc logic doesn't enable
because the `fileNameToModuleName` method is not defined on the host.
This can break reference emission for Tsurge analyzers/programs and
result in subtle differences to real `ng_module` compilations. This
commit fixes this by making the method available in 1P Tsurge.
Notably, reference emission can occur during analysis— so even if
migrations aren't "emitting TS -> JS" output.
PR Close#62447
When executing the compiler-cli in the browser, or some parts of it—
modules like `url` are stubbed out with `{}`. This commit accounts for
this.
PR Close#62431
Use the packaged versions of the packages instead of the local ts_project dependencies to prevent multiple versions of the deps to enter test bundles
PR Close#62413
Since we know that DOM properties won't go to an inputs, we can move the remapping logic to the compiler, saving us some processing on the client.
PR Close#62421
Add support for mode and redirect options in Angular's HttpClient based on fech provider to enable control CORS behavior and redirect handling
PR Close#62315
In this commit, the conditional branching around `ɵloadImpl` is removed from client-side code, as `ɵloadImpl` is never defined in client bundles. This makes the logic simpler and improves tree-shaking, allowing the `from()` import to be dropped from the common bundle in browser builds.
PR Close#62191
In this commit, the use of `images.forEach` was replaced with a traditional `for (let i = 0; ...)` loop to ensure compatibility with Domino. Domino may return a custom internal data structure (e.g., a lazily evaluated query result) that does not fully support standard iteration protocols like `forEach` or `for...of`.
Naturally, this would never happen in any browser. It occurs only in unit tests when using `renderApplication()`, because Domino's `querySelectorAll` returns the following:
```js
var nodes = select(selector, context);
return nodes.item ? nodes : new NodeList(nodes);
```
In certain unit tests, it returns an object like:
`{ root: document, filter: function(e) { ... }, lastModTime: 1, done: true, cache: [] }`,
which means that the object does not have a `forEach` method.
As a result, this causes an `afterAll` error: `forEach is not a function`.
PR Close#62290
Your blog post signaled that afterNextRender and afterEveryRender
are now stable:
https://angular.love/angular-20-whats-new#Signal%20related%20APIs
However, only 1 of the overloads of those was marked as stable.
I detected this because angular-eslint errors on calls even to
to the publicApi overload - and while that should probably be
fixed on their end, your announcement wasn't specific about only
one overload being stable, so I assume this was an oversight.
angular-eslint related issue:
https://github.com/angular-eslint/angular-eslint/issues/2534
PR Close#62153
Currently when there's a parser error in interpolated text, the compiler reports an error on the entire text node. This can be really noisy in long strings.
These changes switch to reporting the errors on the specific expressions that caused them.
PR Close#62258
This way, an arbitrary service can implement Angular's service requirements without a hard dependency on @angular/core
ex:
class Foo {
bar = inject(Bar);
}
registerInjectable(Foo);
PR Close#62087
In this commit, the `ngServerMode` check is moved outside the RxJS `pipe()` to ensure that server-only logic is excluded from client bundles. Previously, the `tap()` operator and its closure were always included in the output, even though `ngServerMode` was false on the client and the side effect was never triggered.
By guarding the observable chain earlier, this reduces the RxJS stack frame depth, which simplifies debugging by avoiding unnecessary operator noise in client-side stack traces.
The resulting logic is also easier to reason about and avoids evaluating `HttpResponse` instances where not needed.
PR Close#62238
Joey Perrott