At the moment local compilation mode does not support custom decorators, and it leads to unhandled errors. In this change a compile time diagnostic is produced in local mode for custom decorators. This is a temporary solution since there are few custom decorators are in use in g3. Custom decorators will be eventually supported in local mode.
PR Close#53983
This adds initial support for extracting and rendering call and construct
signatures of classes, like within the new `InputFunction` for signal
inputs.
For now, signatures are a rare occasion and represented as class member
entries. In the future we might consider exposing this via its own entry
type, and field on the class/interface entry.
PR Close#54053
This fixes the definitions for signal-based inputs in the language
service and type checking symbol builder.
Signal inputs emit a slightly different output. The output works well
for comppletion and was designed to affect language service minimally.
Turns out there is a small adjustment needed for the definition symbols.
PR Close#54053
This commit separates `InputSignal` for input signals with transforms.
The reason being that most of the time, signal inputs are not using
transforms and the generics are rather confusing.
Especially for users with inferred types displayed in their IDEs, the
input signal types are seemingly complex, even if no transform is used.
For this reason, we are introducing a new type called
`InputSignalWithTransform`. This type will be used for inputs with
transforms, while non-transform inputs just use `InputSignal`.
A notable fact is that `InputSignal` extends `InputSignalWithTransform`,
with the "identity transform". i.e. there is no transform. This allows
us to share the code for input signals. In practice, we don't expect
users to pass around `InputSignal`'s anyway.
PR Close#54053
During the template parsing stage two-way bindings are split up into a property and event binding. All the downstream code treats these binding the same as their one-way equivalents. For some future work we'll have to distinguish between the two so these changes update the `BoundElementProperty.type` and `ParsedEvent.type` to include a `TwoWay` type. All existing call-sites have been updated to treat `TwoWay` the same as `Property`/`Regular`, but more specialized logic will be added in the future.
PR Close#54065
The Template Pipeline is a brand new backend for the Angular compiler, replacing `TemplateDefinitionBuilder`. It generates the Ivy instructions corresponding to an input template (or host binding). The Template Pipeline has an all-new design based on an intermediate representation compiled over many phases, which will allow us to experiment with compiler changes more easily in the future.
With this commit, the template pipeline can now be enabled in any project via the `useTemplatePipeline` TSConfig option. However, it is still disabled by default.
PR Close#54057
This commit introduces three additional diagnostics for queries:
- If a query (either using decorator or signal-based) is declared on a
static class member, a diagnostic is raised.
- If a signal-based query is mixed with a query decorator, a diagnostic
is raised. Similar to signal inputs.
- If a singal-based query is also declared in the directive/component
class decorator metadata, a diagnostic is raised.
PR Close#54019
Due to some refactorings, we were only checking the function name
and whether it originates from an import. We should also verify the
module. This seems like logic we lost in the refactorings.
PR Close#54019
This commit ensures that libraries can use signal-based queries, and the
partial compilation output will capture their metadata.
The linker is updated to support parsing this.
Two notes:
1. Older linker versions are not capable of parsing this, so the minimum
version for signal-based queries is adjusted when such are used.
2. We only emit `isSignal` metadata for queries when signal queries are
used. This enables libraries to continue supporting older linker
versions, if signal-based queries are not used.
PR Close#53978
Adds a compiler integration test for recognizing signal-based queries,
and emitting the expected output. Concrete output will be verified via
the compliance tests.
PR Close#53978
This commit uses the initializer API recognition that we built for
signal-based inputs, and teaches the compiler to recognize class members
that refer to `viewChild`, `viewChildren`, `contentChild` or
`contentChildren`. Those will declare signal-based view or content queries.
PR Close#53978
The new `input` API is recognized using class member initializers.
We want to support similar APIs for queries, using e.g. `viewChild`
or `viewChild.required`.
This commit extracts the input recognition API and makes it reusable,
so that the same logic can be used to detect queries on a class member.
Additional changes:
- replacing `coreModule` with the simpler `isCore` parameter. This is
more readable.
- support for detecting a list of API names on a single class member.
This allows us to detect possible query functions on the same class
member without having to check X times. We simply check for the
initializer API pattern and check if one API function name matches.
PR Close#53978
At the moment local compilation breaks for host directives because the current logic relies on global static analysis. This change creates a local version by cutting the diagnostics and copying the directive identifier as it is to the generated code without attempting to statically resolve it.
PR Close#53877
At the moment when unified host is selected (through option `_useHostForImportGeneration`) the compiler always generates alias reexports. Such reexports are mainly generated to satisfy strict dependency condition for generated files. Such condition is no longer the case for G3. At the same time, these alias reexports make it impossible to mix locally compiled targets with globally compiled targets. More precisely, a globally compiled target may not be able to consume a locally compiled target as its dependency since the former may import from the alias reexports which do not exist in the latter due to local compilation mode. So, to make global-local compilation interop possible, it is required to be able to turn off alias reexport generation.
PR Close#53937
This commit adds extra logic to produce a diagnostic in case `@Component.deferredImports` contain types from imports that also bring eager symbols. This would result in retaining a regular import and generating a dynamic import, which would not allow to defer-load dependencies.
PR Close#53899
This commit update the logic to enable `register` and `resolve` phases for local compilation. Those phases will be useful for local compilation in certain cases (will be used in followup PRs).
PR Close#53901
Currently when the extended type check fails due to an import reference
that cannot be generated, the fatal diagnostic is not caught and
not properly exposed as a `ts.Diagnostic` that can be gracefully
handled. This is inconsistent to non-extended type checking diagnostics.
This is problematic because Angular CLI applications currently fail in
obscure ways because:
- the CLI does not expect `getDiagnosticsForFile` to actually throw
runtime errors.
- the CLI does not seem to properly print these errors given the
parallel workers and build excection, and those errors are
especially hard to debug because there is no `stack` for
`FatalDiagnosticError`'s.
Example: `MyDir` is not exported and the type check block cannot reference it.
PR Close#53896
This commit updates the `DeferredSymbolTracker` class to take info account the `onlyExplicitDeferDependencyImports` flag. The `DeferredSymbolTracker` class also exposes a new API to register import declarations as explicitly deferred, which will be used in followup commits.
PR Close#53591
This commit updates typechecker to store full Pipe metadata in internal data strctures, so that this information is available to more places in the code, which will be updated in a followup commit.
PR Close#53591
This commit updates a few places to extract the logic into a separate functions which will be reused in a few places in followup commits.
PR Close#53591
This allows us to ensure signal inputs and a potential JIT transform
remain single file compilation compatible. The consequences are that
options need to be statically analyzable more strictly, compared to
loosened restrictions with static interpretation where e.g. `alias`
can be defined through a shared variable.
PR Close#53872
Improves the recognition of the `input`/`input.required` functions to
not depend on external module resolution. This is useful for local
compilation and for transforms operating on a single file/ isolated
module.
PR Close#53808
Currently when someone declares a signal or non-signal input on a static
class member, the compiler will not yield any diagnostic. We can detect
these mistakes and report a diagnostic to help our users.
PR Close#53808
This commit addds two diagnostics for two scenarios where signal inputs
are declared incorrectly:
- a signal input is also annotated with `@Input` in the TypeScript
sources.
- a signal input is also declared in the `inputs` option of
`@Directive/`@Component`.
PR Close#53808
This commit adds a transform for supporting input signals in JIT
environments. The transform will be wired up for Angular CLI
applications automatically. An integration test verifies that this fixes
unit testing with signal inputs.
The transform basically will take the signal input metadata and
transform it into `@Input` decorators that can provide static
information to the Angular JIT runtime when the directive/component
definition is compiled.
PR Close#53808
This commit does two things:
- exposes `addImports` so that it can be used by transforms that we are
adding to the compiler. e.g. the signal input to `@Input` transform.
- `updates `addImports` to support/use the transform context factory.
This will allow us to write proper transforms using `addImports`. Also
leverages this in the Ivy JS/DTS transforms.
PR Close#53808
Moves the signal input class member extraction logic into the dedicated
input function file. This makes the code for signal inputs more
self-contained.
This commit then re-exposes the function as part of `ngtsc/annotations`
so that it can be used later for a transform that will take the signal
input metadata and translate it into a `@Input` decorator. This allows
us to remove code duplication and guarantees consistency/correctness
PR Close#53808
We recently landed a commit to introduce support for generic type
checking of signal inputs. For that we had to implement logic that
will generate imports for inline type constructors. This required
changes to the context logic and `TypeCtorOp` file-level op.
This commit ensures that everything is working as expected, specifically
in cases where an inline type ctor is generated and imports would be
needed to unwrap the class members for `InputSignal`.
PR Close#53808
As part of testing we did accidentally use `bitwiseAnd` for the input
flags, given we started without an extra flag for `HasTransform`.
This commit teaches the compiler to support emitting bitwise OR
and uses it when combining input flags, fully re-enabling transforms
for signal components after the new flag mechanism was introduced in
previous commits.
PR Close#53808
This commit adds additional type-check transform tests for signal
inputs. These tests verify some of the problems with covariance,
contravariance and bivariance that we were suspecting to be problematic
if we would assign `InputSignal`'s directly to the type constructors.
PR Close#53571
Currently compiling input transform in local mode breaks, since compiler does static analysis for the transform function, and this cannot be done in local mode if the function is imported from another compilation unit. In this fix the static analysis is ditched in local mode.
PR Close#53645
The diagnostic was catching the following case:
```ts
name = signal('Angular');
```
but not the following ones:
```ts
name = signal('Angular').asReadonly();
name = computed(() => 'Angular');
name!: Signal<string>
```
This was not catched in the tests because the type of `Signal` is different than the one actually used in core.
It turns out the real type forces the diagnostic to check both the `symbol.tsType.symbol` and the `symbol.tsType.aliasSymbol`.
PR Close#53585
Consider a case when an explicit `this` read is inside a template with a context that also provides the variable name being read:
```
<ng-template let-a>{{this.a}}</ng-template>
```
Clearly, `this.a` should refer to the class property `a`. However, in today's Angular, `this.a` will refer to `let-a` on the template context.
Amazingly, both TemplateDefinitionBuilder and the Typecheck block have the same bug, and are consistent with each other! This is because `ImplicitReceiver` extends `ThisReceiver` in the parser AST, which is an insane gotcha.
In this commit, I patch the template pipeline to emulate this behavior as well.
To actually fix this nastiness, we have to:
- Update `ingest.ts` in the Template Pipeline (see the corresponding comment)
- Check `type_check_block.ts` in the Typecheck block code (see the corresponding comment)
- Turn off legacy TemplateDefinitionBuilder
- Fix g3, and release in a major version
PR Close#53594
Whenever an input of a directive changes, the semantic symbol should
reflect this change for the type check API. This is important because
signal inputs require special output in the type checking blocks- hence
we need to ensure that such type checking blocks are re-generated
properly.
Test verify that incremental type-checking builds work as expected now.
PR Close#53521
Whenever a signal input is captured in a type check block, we will
insert an import. This will change the import graph so that the full
TypeScript program cannot be structurally re-used.
We can fix this trivially by ensuring the import graph remains stable,
by always generating an import to e.g. `@angular/core`. This fixes the
issue nicely for type-check block files. A test verifies this.
For inline code, such as TCB inline or the type constructors inline,
this fix is not applicable because we would change user-input source files,
adding new edges that would not exist for subsequent builds- causing the
program to be not re-used completely. One idea was to rely on the
existing edge that can be assumed to exist for directive code files.
This is true technically, but in practice TS does not deduplicate
imports- so our new namespace import when referencing our symbols will
invalidate the re-use. We will address this in a follow-up. There are a
couple of options, such as working with the TS team, updating the
existing edge, or inlining our helpers as well.
PR Close#53521
This commit adds the last remaining piece for signal input
type-checking. Bound values to signal inputs are already checked
properly at this point, but inference of generic directive/component
types through their inputs is not implemented.
This commit fixes this. To achieve this, there are a couple of potential
solutions. The generics of a directive are inferred based on input
value expressions using a so-called type constructor. The constructor
looks something like this:
```
const _ctor = <T>(v: Pick<Dir<T>, 'input1', 'input2'>) => Dir<T>;
_ctor({input1: expr1, input2: expr2});
```
This works very well for non-signal inputs where the class member is
directly holding the input values. For signal inputs, this does NOT
work because the class member will actually hold the `InputSignal`
instance. There are a couple of solutions to this:
1. Calling `_ctor` with an `InputSignal<typeof value>`
2. Converting the `_ctor` input signal fields to their write types
(unwrapping the input signals).
We've decided to go with the second option as TypeScript is very
sensitive with assignments and its checks. i.e. co-variance,
contravariance or bivariance. Semantically it makes more sense to unwrap
the input signal "write type" directly and "assign to it". This is safer
and conceptually also easier to follow. A type constructor continues to
only receive the "expresison values". This simplifies code as well.
It's worth noting that the unwrapping as per option 2 also comes at a
cost. We need to be able to generate imports in type constructors. This
was not possible until the previous commit because inline type constructors
did not have an associated type-check block `Environment` and we were
missing access to expression translation and correct import generation.
Overall, solution 2 is now implemented as works as expected. This commit
adds additional unit tests to ensure this.
PR Close#53521
For signal inputs we are looking at generating additional code inside
type constructors. This code is planned to reference an external type
from `@angular/core` to unwrap `InputSignal`'s class fields.
The existing `Environment` class contains helpers for emitting such
references / and translating them from the output AST. We extract
this logic into a superclass for only emitting references. A similar
type already existed to avoid circular dependencies- but now we have
actual use-cases to populate this as a base class.
This allows us to create more-suitable minimal emit environments
when we e.g. generate type constructors inline- which are not
part of any type check block. The existing `Environment` class is scoped
to type check blocks and therefore was not suitable.
PR Close#53521
Signal inputs do not need coercion members for their transforms. That is
because the `InputSignal` type- which is accessible in the class member-
already holds the type of potential "write values". This eliminates the
need for coercion members which were simply used to somehow capture this
write type (especially when libraries are consumed and only `.d.ts` is
available).
We can simplify this, and also significantlky loosen restrictions
of transform functions- given that we can fully rely on TypeScript for
inferring the type. There is no requirement in being able to
"transplant" the type into different places- hence also allowing
supporting transform functions with generics, or overloads.
In a follow-up commit, once more parts are place, there will be some
compliance tests to ensure these new "loosend restrictions".
PR Close#53521
This commit ensures that the type-check diagnostic testing
infrastructure is prepared to validate signal inputs. i.e. providing the
necessary "mocks" in the fake "d.ts" of `@angular/core`.
The commit then sets up a Golang-style table driven testing environment
that allows us to validate/verify signal input type-checking in a
readable way.
With this infrastructure set up, this commit defines an initial set
of unit tests for type checking of input signals.
PR Close#53521
This commit introduces the initial type-checking for signal inputs.
To enable type-checking od signal inputs, there are a couple of tricks
needed. It's not trivial as it would look like at first glance.
Initial attempts could have been to generate additional statements in
type-checking blocks for signal inputs to simply call a method like
`InputSignal#applyNewValue`. This would seem natural, as it would match
what will happen at runtime, but this would break the language-service
auto completion in a highly subtle way. Consider the case where multiple
directives match the same input. Consider the directives have some
overlap in accepted input values, but they also have distinct diverging
values, like:
```ts
class DirA {
value = input<'apple'|'shared'>();
}
class DirB {
value = input<'orange'|'shared'>();
}
```
In such cases, auto completion for the binding expression should suggest
the following values: `apple`, `shared`, `orange` and `undefined`.
The language service achieves this by getting completions in the
type-check block where the user expression would live. This BREAKS if
we'd have multiple places where the expression from the user is used.
Two different places, or more, surface additional problems with
diagnostic collection. Previously diagnostics would surface the union
type of allowed values, but with multiple places, we'd have to work with
potentially 1+ diagnostics. This is non-ideal.
Another important consideration is test coverage. It might sound
problematic to consider the existing test infrastructure as relevant,
but in practice, we have thousands of diagnostic type check block tests
that would greatly benefit if the general emit structure would still
match conceptually. This is another bonus argument on why changing the
way inputs are applied is probably an option we should consider as a
last resort.
Ultimately, there is a good solution where we unwrap directive signal
inputs, based on metadata, and access a brand type field on the
`InputSignal`. This ensures auto-completion continues to work as is, and
also the structure of type check blocks doesn't change conceptually. In
future commits we also need to handle type-inference for generic signal
inputs.
Note: Another alternative considered, in terms of using metadata or not.
We could have type helpers to unwrap signal inputs using type helpers
like: `T extends InputSignal<any, WriteT> ? WriteT : T`. This would
allow us to drop the input signal metadata dependency, but in reality,
this has a few issues:
- users might have `@Input`'s passing around `InputSignal`'s. This is
unlikely, but shows that the solution would not be fully correct.
- we need the metadata regardless, as we plan on accessing it at runtime
as well, to distinguish between signal inputs and normal inputs when
applying new values. This was not clear when this option was
considered initially.
PR Close#53521
Payam Valadkhan