In the new behavior Angular removes applications from the testability registry when the
root view gets destroyed. This eliminates a memory leak, because before that the
TestabilityRegistry holds references to HTML elements, thus they cannot be GCed.
PR Close#22106
PR Close#39876
`ElementRef` is declared in ViewEngine but it sub-classed in Ivy. This creates a circular
dependency between ViewEngine `ElementRef` which needs to declare `__NG_ELEMENT_ID__` and
ivy factory which needs to create it. The workaround used to be to pass the `ElementRef`
through stack but that created a very convoluted code. This refactoring simply bundles the
two files together and removes the stack workaround making the code simpler to follow.
PR Close#39621
IMPORTANT: `HEADER_OFFSET` should only be refereed to the in the `ɵɵ*` instructions to translate
instruction index into `LView` index. All other indexes should be in the `LView` index space and
there should be no need to refer to `HEADER_OFFSET` anywhere else.
PR Close#39233
Before this refactoring/fix the ICU would store the current selected
index in `TView`. This is incorrect, since if ICU is in `ngFor` it will
cause issues in some circumstances. This refactoring properly moves the
state to `LView`.
closes#37021closes#38144closes#38073
PR Close#39233
`TNodeType.View` was created to support inline views. That feature did
not materialize and we have since removed the instructions for it, leave
an unneeded `TNodeType.View` which was still used in a very
inconsistent way. This change no longer created `TNodeType.View` (and
there will be a follow up chang to completely remove it.)
Also simplified the mental model so that `LView[HOST]`/`LView[T_HOST]`
always point to the insertion location of the `LView`.
PR Close#38707
`previousOrParentTNode` stores current `TNode`. Due to inconsistent
implementation the value stored would sometimes belong to the current
`TView` and sometimes to the parent. We have extra logic which accounts
for it. A better solution is to just ensure that `previousOrParentTNode`
always belongs to current `TNode`. This simplifies the mental model
and cleans up some code.
PR Close#38707
This commit replaces an assert with more descriptive error message that is thrown in case `<ng-template>` or `<ng-container>` is used as host element for a Component.
Resolves#35240.
PR Close#35916
Currently when bootstrapped component is being removed using `ComponentRef.destroy` or `NgModuleRef.destroy` methods, DOM nodes may be retained in the DOM tree. This commit fixes that problem by always attaching host element of the internal root view to the component's host view node, so the cleanup can happen correctly.
Resolves#36449.
PR Close#37600
The _tViewNode field (that was marked as internal) on the ViewRef is not
necessery as a reference to a relevant TView is available as a local
variable.
PR Close#36814
This commit fixes 2 separate issues related to root nodes retrieval from
embedded views with `<ng-content>`:
1) we did not account for the case where there were no projectable nodes
for a given `<ng-content>`;
2) we did not account for the case where projectable nodes for a given
`<ng-content>` were represented as an array of native nodes (happens in
the case of dynamically created components with projectable nodes);
Fixes#35967
PR Close#36051
In View Engine, host element of dynamically created component received attributes and classes extracted from component's selector. For example, if component selector is `[attr] .class`, the `attr` attribute and `.class` class will be add to host element. This commit adds similar logic to Ivy, to make sure this behavior is aligned with View Engine.
PR Close#34481
Prior to this change, element namespace was not set for host elements of dynamically created components that resulted in incorrect rendering in a browser. This commit adds the logic to pick and set correct namespace for host element when component is created dynamically.
PR Close#35136
- Adds `TView` into `LFrame`, read the `TView` from `LView` on `enterView`.
- Before this change the `TView` was ofter looked up from `LView` as `lView[TVIEW]`. This is suboptimal since reading from an Array, requires that the read checks array size before the read. This means that such a read has a much higher cost than reading from the property directly. By passing in the `TView` explicitly it makes the code more explicit and faster.
- Some rearrangements of arguments so that `TView` would come before `LView` for consistency.
PR Close#35069
NOTE: This change must be reverted with previous deletes so that it code remains in build-able state.
This change deletes old styling code and replaces it with a simplified styling algorithm.
The mental model for the new algorithm is:
- Create a linked list of styling bindings in the order of priority. All styling bindings ere executed in compiled order and than a linked list of bindings is created in priority order.
- Flush the style bindings at the end of `advance()` instruction. This implies that there are two flush events. One at the end of template `advance` instruction in the template. Second one at the end of `hostBindings` `advance` instruction when processing host bindings (if any).
- Each binding instructions effectively updates the string to represent the string at that location. Because most of the bindings are additive, this is a cheap strategy in most cases. In rare cases the strategy requires removing tokens from the styling up to this point. (We expect that to be rare case)S Because, the bindings are presorted in the order of priority, it is safe to resume the processing of the concatenated string from the last change binding.
PR Close#34616
Prior to this change, ComponentFactory.create function invocation in Ivy retained the content of the host element (in case host element reference or CSS seelctor is provided as an argument). This behavior is different in View Engine, where the content of the host element was cleared, except for the case when ShadowDom encapsulation is used (to make sure native slot projection works). This commit aligns Ivy and View Engine and makes sure the host element is cleared before component content insertion.
PR Close#33487
When debugging `LView`s it is easy to get lost since all of them have
the same name. This change does three things:
1. It makes `TView` have an explicit type:
- `Host`: for the top level `TView` for bootstrap
- `Component`: for the `TView` which represents components template
- `Embedded`: for the `TView` which represents an embedded template
2. It changes the name of `LView` to `LHostView`, `LComponentView`, and
`LEmbeddedView` depending on the `TView` type.
3. For `LComponentView` and `LEmbeddedView` we also append the name of
of the `context` constructor. The result is that we have `LView`s which
are name as: `LComponentView_MyComponent` and `LEmbeddedView_NgIfContext`.
The above changes will make it easier to understand the structure of the
application when debugging.
NOTE: All of these are behind `ngDevMode` and will get removed in
production application.
PR Close#33449
`LFrame` stores information specifice to the current `LView` As the code
enters and leaves `LView`s we use `enterView()` and `leaveView()`
respectively to build a a stack of `LFrame`s. This allows us to easily
restore the previous `LView` instruction state.
PR Close#33178
Currently Ivy stores the element attributes into an array above the component def and passes it into the relevant instructions, however the problem is that upon minification the array will get a unique name which won't compress very well. These changes move the attributes array into the component def and pass in the index into the instructions instead.
Before:
```
const _c0 = ['foo', 'bar'];
SomeComp.ngComponentDef = defineComponent({
template: function() {
element(0, 'div', _c0);
}
});
```
After:
```
SomeComp.ngComponentDef = defineComponent({
consts: [['foo', 'bar']],
template: function() {
element(0, 'div', 0);
}
});
```
A couple of cases that this PR doesn't handle:
* Template references are still in a separate array.
* i18n attributes are still in a separate array.
PR Close#32798
After a series of recent refactorings `enterView` and `leaveView` became
identical. This PR merges both into one concept of view selectio (similar
to a node selection). This reduces number of concepts and code size.
PR Close#32263
In VE the `Sanitizer` is always available in `BrowserModule` because the VE retrieves it using injection.
In Ivy the injection is optional and we have instructions instead of component definition arrays. The implication of this is that in Ivy the instructions can pull in the sanitizer only when they are working with a property which is known to be unsafe. Because the Injection is optional this works even if no Sanitizer is present. So in Ivy we first use the sanitizer which is pulled in by the instruction, unless one is available through the `Injector` then we use that one instead.
This PR does few things:
1) It makes `Sanitizer` optional in Ivy.
2) It makes `DomSanitizer` tree shakable.
3) It aligns the semantics of Ivy `Sanitizer` with that of the Ivy sanitization rules.
4) It refactors `DomSanitizer` to use same functions as Ivy sanitization for consistency.
PR Close#31934
This commit fixes a bug where we assumed all dynamically created
components would have tag-name selectors, so we passed through the
"tag name" to the renderer as the first index of the selector. For
components with attribute selectors, the tag name would be "", so
the renderer would try to create an element with tag name "" and
throw. Now we default to a "div" element the same way that View
Engine did.
Closes#31785
PR Close#31812
The way the `ComponentFactory.create` is set up at the moment is that if a `rootSelectorOrNode` is passed in, the root context will be injected instead of creating dedicated one for the component. As far as I can tell, there doesn't seem to be a reason to do this and nothing seems to break because of it. These changes switch to always create the root context.
PR Close#31661
Prior to this commit, host element of a view created via TestBed.createComponent was not attached to the component's host, making it problematic to use TestBed.createComponent API in component factories, which might be used for testing purposes only. This behavior is observed in google3 app tests and was supported by VE, so this commit aligns Ivy and VE.
PR Close#31318
Currently in Ivy whenever we encounter a new namespace, we set it in the global state so that all subsequent nodes are created under the same namespace. Next time a template is run the namespace will be reset back to HTML.
This breaks down if the last node that was rendered was under the SVG or MathML namespace and we create a component through `ViewContainerRef.create`, because the next template function hasn't run yet and it hasn't had the chance to update the namespace. The result is that the root node of the new component will retain the wrong namespace and may not end up rendering at all (e.g. if we're trying to show a `div` inside the SVG namespace). This issue has the potential to affect a lot of apps, because all components inserted through the router also go through `ViewContainerRef.create`.
PR Close#31232
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
Currently in Ivy we throw when attempting to destroy a `ComponentRef` that has been destroyed, however in ViewEngine we didn't which can cause some tests to break. These changes remove the error to match ViewEngine.
These changes resolve FW-1379.
PR Close#31022
With View engine it was possible to declare multiple projection
definitions and to programmatically project nodes into the slots.
e.g.
```html
<ng-content></ng-content>
<ng-content></ng-content>
```
Using `ViewContainerRef#createComponent` allowed projecting
nodes into one of the projection defs (through index)
This no longer works with Ivy as the `projectionDef` instruction only
retrieves a list of selectors instead of also retrieving entries for
reserved projection slots which appear when using the default
selector multiple times (as seen above).
In order to fix this issue, the Ivy compiler now passes all
projection slots to the `projectionDef` instruction. Meaning that
there can be multiple projection slots with the same wildcard
selector. This allows multi-slot projection as seen in the
example above, and it also allows us to match the multi-slot node
projection order from View Engine (to avoid breaking changes).
It basically ensures that Ivy fully matches the View Engine behavior
except of a very small edge case that has already been discussed
in FW-886 (with the conclusion of working as intended).
Read more here: https://hackmd.io/s/Sy2kQlgTE
PR Close#30561
There is an encoding issue with using delta `Δ`, where the browser will attempt to detect the file encoding if the character set is not explicitly declared on a `<script/>` tag, and Chrome will find the `Δ` character and decide it is window-1252 encoding, which misinterprets the `Δ` character to be some other character that is not a valid JS identifier character
So back to the frog eyes we go.
```
__
/ɵɵ\
( -- ) - I am ineffable. I am forever.
_/ \_
/ \ / \
== == ==
```
PR Close#30546
Prior to this change, components created via TestBed.createComponent in the same test were placed into the same root context, which caused problems in conjunction with fixture.autoDetectChanges usage in the same test. Specifically, change detection was triggered immediately for created component (starting from the 2nd one) even if it was not required/desired. This commit makes Ivy and VE behavior consistent: now every component created via TestBed.createComponent is isolated from each other. Current solution uses host element id naming convention, which is not ideal, but helps avoid public API surface changes at this point (we might revisit this approach later).
Note: this commit also adds extra tests to verify bootstrap and change detection behavior in case of multiple components in `bootstrap` array in @NgModule, to make sure this behavior is aligned between Ivy and VE.
PR Close#29981
The `Δ` caused issue with other infrastructure, and we are temporarily
changing it to `ɵɵ`.
This commit also patches ts_api_guardian_test and AIO to understand `ɵɵ`.
PR Close#29850
- 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
Google3 detected circular references here, so splitting up this rather hodge-podge list of functions into slightly better organizational units.
PR Close#28382
- TView no longer stores childIndex
- LView now as CHILD_HEAD and CHILD_TAIL
TView used to store the head of the list, therefor all LViews had to have the same head, which is incorrect.
PR Close#28382
Accounts for schemas in when validating properties in Ivy.
This PR resolves FW-819.
A couple of notes:
* I had to rework the test slightly, in order to have it fail when we expect it to. The one in master is passing since Ivy's validation runs during the update phase, rather than creation.
* I had to deviate from the design in FW-819 and not add an `enableSchema` instruction, because the schema is part of the `NgModule` scope, however the scope is only assigned to a component once all of the module's declarations have been resolved and some of them can be async. Instead, I opted to have the `schemas` on the component definition.
PR Close#28637
Prior to this change we used current injector implementation for module injector, which was causing problems and produces circular dependencies in case the same token is referenced (with @SkipSelf flag) in the `deps` array. The origin of the problem was that once `directiveInject` implementation becomes active, it was used for module injector as well, thus searching deps in Component/Directive DI scope. This fix sets `injectInjectorOnly` implementation for module injector to resolve the problem.
PR Close#28667
`LView` `HOST` was set in most cases right after creating `LView`.
This makes the API cleaner by explicitly passing it ont `createLView`.
PR Close#28461
Previously, bootstrapping a component with render3 would create a chained injector with the test bed ngModule instead of the ngModule that the component belongs to.
Now when a component belongs to an ngModule, we use that for the chained injector, ensuring the correct injection of any providers that this ngModule contains.
FW-776 #resolve
PR Close#28183
