This commit is the final patch of the ivy styling algorithm refactor.
This patch swaps functionality from the old styling mechanism to the
new refactored code by changing the instruction code the compiler
generates and by pointing the runtime instruction code to the new
styling algorithm.
PR Close#30742
The Angular runtime frequently calls into user code (for example, when
writing to a property binding). Since user code can throw errors, calls to
it are frequently wrapped in a try-finally block. In Ivy, the following
pattern is common:
```typescript
enterView();
try {
callUserCode();
} finally {
leaveView();
}
```
This has a significant problem, however: `leaveView` has a side effect: it
calls any pending lifecycle hooks that might've been scheduled during the
current round of change detection. Generally it's a bad idea to run
lifecycle hooks after the application has crashed. The application is in an
inconsistent state - directives may not be instantiated fully, queries may
not be resolved, bindings may not have been applied, etc. Invariants that
the app code relies upon may not hold. Further crashes or broken behavior
are likely.
Frequently, lifecycle hooks are used to make assertions about these
invariants. When these assertions fail, they will throw and "swallow" the
original error, making debugging of the problem much more difficult.
This commit modifies `leaveView` to understand whether the application is
currently crashing, via a parameter `safeToRunHooks`. This parameter is set
by modifying the above pattern:
```typescript
enterView();
let safeToRunHooks = false;
try {
callUserCode();
safeToRunHooks = true;
} finally {
leaveView(..., safeToRunHooks);
}
```
If `callUserCode` crashes, then `safeToRunHooks` will never be set to `true`
and `leaveView` won't call any further user code. The original error will
then propagate back up the stack and be reported correctly. A test is added
to verify this behavior.
PR Close#31244
Projecting bare ICU expressions failed because we would assume that component's content nodes would be projected later and doing so at that point would be wasteful. But ICU nodes are handled independently and should be inserted immediately because they will be ignored by projections.
FW-1348 #resolve
PR Close#30696
The `flatten` function used `concat` and `slice` which created a lot of intermediary
object allocations. Because `flatten` is used from query any benchmark which
used query would exhibit high minor GC counts.
PR Close#30468
This is the first refactor PR designed to change how styling bindings
(i.e. `[style]` and `[class]`) behave in Ivy. Instead of having a heavy
element-by-element context be generated for each element, this new
refactor aims to use a single context for each `tNode` element that is
examined and iterated over when styling values are to be applied to the
element.
This patch brings this new functionality to prop-based bindings such as
`[style.prop]` and `[class.name]`.
PR Close#30469
Ivy uses R3Injector, but we are currently pulling in both the StaticInjector
(View Engine injector) and the R3Injector when running with Ivy. This commit
adds an ivy switch so calling Injector.create() pulls in the correct
implementation of the injector depending on whether you are using VE or Ivy.
This saves us about 3KB in the bundle.
PR Close#30219
We recently had an unexpected size regression in the hello world
tests because the CLI devkit released an RC that regressed us and
the dependencies were not pinned. This change ensures that we only
update dependencies like devkit deliberately, so we do not have
mysterious breakages caused by other packages.
PR Close#30152
Prior to this commit, we were pulling DebugNode and DebugElement
into production builds because BrowserModule automatically pulled
in NgProbe and thus getDebugNode. In Ivy, this is not necessary
because Ivy has its own set of debug utilities. We should use these
existing tools instead of NgProbe.
This commit adds an Ivy switch so we do not pull in NgProbe utilities
when running with Ivy. This saves us ~8KB in prod builds.
PR Close#30130
Master is red due to a size regression that was not caught before. We are making this change to bring master back to green state and will perform further investigation.
PR Close#30134
Previous to this commit, providing a component or directive in a test
module without @Injectable() would throw because the injectable factory
would not be found. Providing components in tests in addition to declaring
or importing them is not necessary, but it should not throw an error.
This commit ensures factory data is saved when component defs and directive
defs are created, which allows them to be processed by the module injector.
Note that bootstrapping is still required for this setup to work because
directiveInject() does not support cases where the view has not been
created. This case will be handled in a future commit.
PR Close#29945
The new styling algorithm in angular is designed to evaluate host
bindings stylinh priority in order of directive evaluation order. This,
however, does not work with respect to parent/sub-class directives
because sub-class host bindings are run after the parent host bindings
but still have priority. This patch ensures that the host styling bindings
for parent and sub-class components/directives are executed with respect
to the styling algorithm prioritization.
Jira Issue: FW-1132
PR Close#29602
Prior to this change, a module's imports and exports would be used verbatim
as an injectors' imports. This is detrimental for tree-shaking, as a
module's exports could reference declarations that would then prevent such
declarations from being eligible for tree-shaking.
Since an injector actually only needs NgModule references as its imports,
we may safely filter out any declarations from the list of module exports.
This makes them eligible for tree-shaking once again.
PR Close#29598
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
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
Fixes the `ngOnDestroy` hook on a component or directive being called twice, if the type is also registered as a provider.
This PR resolves FW-1010.
PR Close#28470
Prior to this fix, both the `NgStyle` and `NgClass` directives made use
of `Renderer2` and this dependency raised issues for future versions of
Angular that cannot inject it. This patch ensures that there are two
versions of both directives: one for the VE and another for Ivy.
Jira Issue: FW-882
PR Close#28711
This commit adds support for the `static: true` flag in
`ViewChild` queries. Prior to this commit, all `ViewChild`
queries were resolved after change detection ran. This is
a problem for backwards compatibility because View Engine
also supported "static" queries which would resolve before
change detection.
Now if users add a `static: true` option, the query will be
resolved in creation mode (before change detection runs).
For example:
```ts
@ViewChild(TemplateRef, {static: true}) template !: TemplateRef;
```
This feature will come in handy for components that need
to create components dynamically.
PR Close#28811
Prior to this fix if a root component was instantiated it create host
bindings, but never render them once update mode ran unless one or more
slot-allocated bindings were issued. Since styling in Ivy does not make
use of LView slots, the host bindings function never ran on the root
component.
This fix ensures that the `hostBindings` function does run for a root
component and also renders the schedlued styling instructions when
executed.
Jira Issue: FW-1062
PR Close#28664
There are two ways to bootstrap an Ivy app: with `platformBrowserDynamic().bootstrapModule` and `renderComponent`.
To distinguish between these two approaches we call the `platformBrowserDynamic().bootstrapModule` way `ivy-compat` and the `renderComponent` way just `ivy`.
PR Close#28372
In ESM5 decorated classes can be indicated by calls to `__decorate()`.
Previously the `ReflectionHost.findDecoratedClasses()` call would identify
helper calls of the form:
```
SomeClass = tslib_1.__decorate(...);
```
But it was missing calls of the form:
```
SomeClass = SomeClass_1 = tslib_1.__decorate(...);
```
This form is common in `@NgModule()` decorations, where the class
being decorated is referenced inside the decorator or another
member.
This commit now ensures that a chain of assignments, of any length,
is now identified as a class decoration if it results in a call to
`__decorate()`.
Fixes#27841
PR Close#27848
This adds compilation of @NgModule providers and imports into
ngInjectorDef statements in generated code. All @NgModule annotations
will be compiled and the @NgModule decorators removed from the
resultant js output.
All @Injectables will also be compiled in Ivy mode, and the decorator
removed.
PR Close#22458
InjectionToken can be created with an ngInjectableDef, and previously
this allowed the full expressiveness of @Injectable. However, this
requires a runtime reflection system in order to generate factories
from expressed provider declarations.
Instead, this change requires scoped InjectionTokens to provide the
factory directly (likely using inject() for the arguments), bypassing
the need for a reflection system.
Fixes#22205
PR Close#22207
This commit bundles 3 important changes, with the goal of enabling tree-shaking
of services which are never injected. Ordinarily, this tree-shaking is prevented
by the existence of a hard dependency on the service by the module in which it
is declared.
Firstly, @Injectable() is modified to accept a 'scope' parameter, which points
to an @NgModule(). This reverses the dependency edge, permitting the module to
not depend on the service which it "provides".
Secondly, the runtime is modified to understand the new relationship created
above. When a module receives a request to inject a token, and cannot find that
token in its list of providers, it will then look at the token for a special
ngInjectableDef field which indicates which module the token is scoped to. If
that module happens to be in the injector, it will behave as if the token
itself was in the injector to begin with.
Thirdly, the compiler is modified to read the @Injectable() metadata and to
generate the special ngInjectableDef field as part of TS compilation, using the
PartialModules system.
Additionally, this commit adds several unit and integration tests of various
flavors to test this change.
PR Close#22005
Both Firefox and Safari are vulnerable to XSS if we use an inert document
created via `document.implementation.createHTMLDocument()`.
Now we check for those vulnerabilities and then use a DOMParser or XHR
strategy if needed.
Further the platform-server has its own library for parsing HTML, so we
sniff for that (by checking whether DOMParser exists) and fall back to
the standard strategy.
Thanks to @cure53 for the heads up on this issue.
PR Close#17019
Implement NgOnChangesFeature, ViewContainerRef, TemplateRef,
and the renderEmbeddedTemplate instruction, and wire together the
pieces required for the ngForOf directive to work.
PR Close#21430