This new interface will match classes whether they are abstract or
concrete. Casting as `AbstractType<MyConcrete>` will return a type that
isn't newable. This type will be used to match abstract classes in the
`get()` functions of `Injector` and `TestBed`.
Type isn't used yet so this isn't a breaking change.
Issue #26491
PR Close#29295
Prior to this change, all module metadata would be included in the
`defineNgModule` call that is set as the `ngModuleDef` field of module
types. Part of the metadata is scope information like declarations,
imports and exports that is used for computing the transitive module
scope in JIT environments, preventing those references from being
tree-shaken for production builds.
This change moves the metadata for scope computations to a pure function
call that patches the scope references onto the module type. Because the
function is marked pure, it may be tree-shaken out during production builds
such that references to declarations and exports are dropped, which in turn
allows for tree-shaken any declaration that is not otherwise referenced.
Fixes#28077, FW-1035
PR Close#29598
Introduces a new update schematic called "template-var-assignment"
that is responsible for analyzing template files in order to warn
developers if template variables are assigned to values.
The schematic also comes with a driver for `tslint` so that the
check can be used wtihin Google.
PR Close#29608
- moves all publicly exported instructions to their own files
- refactors namespace instructions to set state in `state.ts`
- no longer exports * from `instructions.ts`.
- `instructions.ts` renamed to `shared.ts` (old `shared.ts` contents folded in to `instructions.ts`)
- updates `all.ts` to re-export from public instruction files.
PR Close#29646
Queries can also be statically accessed within getters. e.g.
```ts
ngOnInit() {
this.myQueryGetter.doSomething();
}
```
In that case we need to check if the `myQueryGetter` definition accesses
a query statically. As we need to use the type checker for every property acess
within lifecylce hooks, the schematic might become slower than before, but considering
that this is a one-time execution, it is totally fine using the type-checker extensively.
PR Close#29609
- Adds `property` instruction
- Does _NOT_ add compiler changes to accommodate `property` instruction, that will be a follow up PR.
- Updates `select` instruction to set the selected index in state.
- Adds dev mode assertions around the selected index state.
Related #29527
PR Close#29513
When an @NgModule is imported more than once in the testing module (for
example it appears in the imports of more than one module, or if it's
literally listed multiple times), then TestBed had a bug where the
providers for the module would be overridden many times.
This alone was problematic but would not break tests. However, the original
value of the providers field of the ngInjectorDef was saved each time, and
restored in the same order. Thus, if the provider array was [X], and
overrides were applied twice, then the override array would become
[X, X'] and then [X, X', X, X']. However, on the second override the state
[X, X'] would be stored as original. The array would then be restored to
[X] and then [X, X'].
Each test, therefore, would continue to double the size of the providers
array for the module, eventually exhausting the browser's memory.
This commit adds a Set to track when overrides have been applied to a module
and refrain from applying them more than once.
PR Close#29571
Prior to this change, recompilation of AOT-compiled components in TestBed may fail when template override is requested. That was happening due to the `styleUrls` field defined for a Component, thus switching its state to "requires resolution" (i.e. having external resources) at compile time. This change avoids this issue by storing styles and resetting `styleUrls` field before recompilation. Once compilation is done, saved styles are patched back onto Component def.
PR Close#29555
- Adds `property` instruction
- Does _NOT_ add compiler changes to accommodate `property` instruction, that will be a follow up PR.
- Updates `select` instruction to set the selected index in state.
- Adds dev mode assertions around the selected index state.
Related #29527
PR Close#29513
forEach is slower as compared to a regular loop but more importantly
this change removes an anonymous function and thus makes stack traces
shorter and easier to read (important for perf analysis).
PR Close#29543
This commit removes code duplication where we had 2 versions of a
`flatten` utility. Moreover this change results in queries using
a non-recursive version of `flatten` which should result in a better
performance of query refresh operations.
PR Close#29547
Prior to this change, Ivy version of TestBed was not designed to support the logic to avoid recompilations - most of the Components/Directives/Pipes were recompiled for each test, even if there were no overrides defined for a given Type. Additional checks to avoid recompilation were introduced in one of the previous commits (0244a2433e), but there were still some corner cases that required attention. In order to support the necessary logic better, Ivy TestBed was rewritten/refactored. Main results of this rewrite are:
* no recompilation for Components/Directives/Pipes without overrides
* the logic to restore state between tests (isolate tests) was improved
* transitive scopes calculation no longer performs recompilation (it works with compiled defs)
As a result of these changes we see reduction in memory consumption (3.5-4x improvement) and pefromance increase (4-4.5x improvement).
PR Close#29483
The API changes are due to enabling strict checks in TypeScript (via `strict: true`).
The payload size changes in `polyfills.js` are due to more browser APIs being patched in recent versions (e.g. `fetch`, `customElement v1`).
PR Close#28219
In View Engine, we used to generate empty QueryLists for content queries on root
components (though we did not actually support populating these lists). We need
to keep this behavior in Ivy for backwards compatibility. Otherwise, components
that are sometimes used as root will fail if they are relying on content query
results to always be defined.
PR Close#29514
In some cases ivy expects projectable nodes to be passed in a different order
to ViewEngine. Specifically, ivy expects the catch-all ("*") to be at index
0, whereas ViewEngine expects it to be at its position at which it was parsed
in the template.
This commit adds one test that breaks under ivy and others that just describe
more accurately what happens in corner cases.
PR Close#27791
ivy's bindingUpdated instruction is using the assertNotEqual check to make
sure that NO_CHANGE value (of type Object) is not passed as a value to be
dirty-checked. In practice it means that any value passed as a binding
value would be compared to the NO_CHANGE object.
It turns out that the assertNotEqual is using == and given
that binding values are of different type and we always compare it to the
NO_CHANGE object we were doing lots of type coercion. It resulted in calls
to expensive types conversions and calls to Object.toString().
A profiler reported ~15% of the self time spent in the assertNotEqual
but it turns out that removing type coercion speeds up Material Chips with
input scenario much more (~40ms down to ~20ms).
This PR introduces new assert method `assertNotSame` that uses strict equality
check. The new assertion is used in binding instructions to compare to
NO_CHANGE object reference.
PR Close#29470
Queries can be also used statically within the "ngDoCheck" and "ngOnChanges" lifecylce hook.
In order to properly detect all queries, we need to also respect these lifecycle hooks.
Resolves FW-1192
PR Close#29492
In ReflectionCapabilities, when checking for own parameters of a type, inherit the types properly for classes that do have a constructor, but the constructor takes no declared parameters and just delegates to super(...arguments). This removes the need to declare trivial constructors in such classes to make them work properly in JIT mode. Without this, DI fails and injectables are undefined.
PR Close#29232
This fix corrects a bug where we were passing a binding _value_
in place of an expected binding index. This reulted in the binding
value being compared to an array length and buggy type coercion.
Fixing this bug speeds up test scenario by ~10-15%.
PR Close#29476
Previously, only directives and services with generic type parameters
would emit `any` as generic type when emitting Ivy metadata into .d.ts
files. Pipes can also have generic type parameters but did not emit
`any` for all type parameters, resulting in the omission of those
parameters which causes compilation errors.
This commit adds support for pipes with generic type arguments and emits
`any` as generic type in the Ivy metadata.
Fixes#29400
PR Close#29403
Previously, the transitive scope calculation could lead into re-compiling
the same module multiple times. This fix ensures we cannot get into this loop.
It should be fixed more completely (e.g. more cases) once FW-1178 is resolved.
PR Close#29402
This patch is the first of a few patches which separates the
styling logic between template bindings (e.g. <div [style])
from host bindings (e.g. @HostBinding('style')). This patch
in particular introduces a series of host-specific styling
instructions and changes the existing set of template styling
instructions not to accept directives. The underyling code (which
communicates with the styling algorithm) still works as it did
before.
This PR also separates the styling instruction code into a separate
file and moves over all other instructions into an dedicated
instructions directory.
PR Close#29292
Prior to this change, we always recompile all Components/Directives/Pipes even if they were AOT-compiled and had no overrides. This is causing problems in case we try to recompile a Component with "templateUrl" or "styleUrls" (which were already resolved in case of AOT) and generally this unnecessary work that TestBed was doing is not required. This commit adds extra logic to check whether a Component/Directive/Pipe already have compiled NG def (like ngComponentDef) and whether there are no overrides present - in this case recompilation is skipped. Recompilation is also skipped in case a Component/Directive has only Provider overrides - in this situation providers resolver function is patched to reflect overrides. Provider overrides are very common in g3, thus this code path ensures no full recompilation.
PR Close#29294
Fixes the incorrect failure message or the TSLint rule that
is used within Google. The TSLint rule is not part of the
public schematic code.
Additionally in order to make it easier to understand what
action the developer needs to take, we rather print out the
expected "static: true/false" statement instead of saying that
a query needs to be static or dynamic. Dynamic is ambiguous, as
there is no `dynamic: true` option.
PR Close#29320
Just updating comments in query-related things to make it easier for the next person that has to grok this for the first time.
Also adds a demo from @mhevery to one of the query specs
Related #29031
PR Close#29342
Carlos Ortiz García
JoostK
Paul Gschwendtner
Ben Lesh
Jon Grunewald
Marc Laval
Judy Bogart
Kara
Alex Rickabaugh
Andrew Kushnir
Miško Hevery
Pawel Kozlowski
JiaLiPassion
Kara Erickson
Kristiyan Kostadinov
Pete Bacon Darwin
Bolek Szewczyk
Cameron Steffen
Greg Magolan
Matias Niemelä
Rob Gregorius