In the previous patch () all the existing styling code was turned
off in favor of using the new refactored ivy styling code. This
patch is a follow up patch to that and removes all old, unused
styling code from the render3 directory.
PR Close#31193
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
Currently the `ɵɵdefineComponent` method has incorrect type
definitions the `directives` and `pipes` metadata property.
The incorrect types allow developers to pass in already instantiated
`DirectiveDef` or `ComponentDef` objects. This can cause unexpected
failures because the definition internally only expects `Type` objects
and now incorrectly tries to read the `ngDirectiveDef` or `ngComponentDef`
of existing definitions.
This issue has been unveiled by enabling the strict function parameter
types flag, where the directive definitions are determined from each array
element in the `directives` or `pipes` property (which can throw).
PR Close#30993
As part of FW-1265, the `@angular/core` package is made compatible
with the TypeScript `--strict` flag. This already unveiled a few bugs,
so the strictness flag seems to help with increasing the overall code health.
Read more about the strict flag [here](https://www.typescriptlang.org/docs/handbook/compiler-options.html)
PR Close#30993
Fixes Ivy matching directives against attribute bindings (e.g. `[attr.some-directive]="foo"`). Works by excluding attribute bindings from the attributes array during compilation. This has the added benefit of generating less code.
**Note:** My initial approach to implementing this was to have a different marker for attribute bindings so that they can be ignored when matching directives, however as I was implementing it I realized that the attributes in that array were only used for directive matching (as far as I could tell). I decided to drop the attribute bindings completely, because it results in less generated code.
PR Close#31541
Previously, multiple ViewContainerRef instances (obtained by injecting
ViewContainerRef multiple times) each had private state that could be out of
sync with actual LContainer, if views were inserted/removed/queried across
the different instances. In particular each instance had its own array which
tracked ViewRefs inserted via that instance.
This commit moves the ViewRefs array onto the LContainer itself, so that it
can be shared across multiple ViewContainerRef instances. A test is added
that verifies ViewContainerRefs now provide a consistent view of the
container.
FW-1377 #resolve
PR Close#30985
Factory functions written by the compiler optionally allow an explicit type
to be passed. If called with this type, an instance of the given type will
be created instead of the type for which the factory was generated. This is
used to power inheritance of Angular types, as if the constructor of a class
is inherited from its superclass, then the factory function of the
superclass must be used (it has all the DI info) to construct an instance of
the derived class.
This commit adjusts typings in a few places to allow factory functions to be
called with this extra type parameter.
PR Close#30855
When translated content was projected, all of the content was reappended, even the placeholders that had been removed in the translation.
To avoid that we added a new flag on `TNode` that specifies that a node is detached, in which case it should be ignored by the projection.
FW-1319 #resolve
PR Close#30783
Before this change we would systematically call LQueries.clone() when executting
elementStart / elementContainerStart instructions. This was often unnecessary as
LQueries can be mutated under 2 conditions only:
- we are crossing an element that has directives with content queries
(new queries must be added);
- we are descending into element hierarchy (creating a child element of an existing element)
and the current LQueries object is tracking shallow queries (shallow queries are removed).
With this PR LQueires.clone() is only done when needed and this gratelly reduces number
of LQueries object created: in the "expanding rows" benchmark number of allocated
(and often GCed just after!) LQueries is reduced from ~100k -> ~35k. This represents
over 1MB of memory that is not allocated.
PR Close#30664
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
`i18nAttributes` instructions always occur after the element instruction. This means that we need to treat `i18n-` attributes differently.
By defining a specific `AttributeMarker` we can ensure that we won't trigger directive inputs with untranslated attribute values.
FW-1332 #resolve
PR Close#30402
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
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
This is the final patch to migrate the Angular styling code to have a
smaller instruction set in preparation for the runtime refactor. All
styling-related instructions now work both in template and hostBindings
functions and do not use `element` as a prefix for their names:
BEFORE:
elementStyling()
elementStyleProp()
elementClassProp()
elementStyleMap()
elementClassMap()
elementStylingApply()
AFTER:
styling()
styleProp()
classProp()
styleMap()
classMap()
stylingApply()
PR Close#30318
This patch breaks up the existing `elementStylingMap` into
`elementClassMap` and `elementStyleMap` instructions. It also breaks
apart `hostStlyingMap` into `hostClassMap` and `hostStyleMap`
instructions. This change allows for better tree-shaking and reduces
the complexity of the styling algorithm code for `[style]` and `[class]`
bindings.
PR Close#30293
Stores the views that are part of a container directly on the `LContainer`, rather than maintaining a dedicated sub-array.
This PR resolves FW-1288.
PR Close#30179
Fixes `HostBinding` and `HostListener` declarations not being inherited from base classes that don't have an Angular decorator.
This PR resolves FW-1275.
PR Close#30158
Fixes view and content queries not being inherited in Ivy, if the base class hasn't been annotated with an Angular decorator (e.g. `Component` or `Directive`).
Also reworks the way the `ngBaseDef` is created so that it is added at the same point as the queries, rather than inside of the `Input` and `Output` decorators.
This PR partially resolves FW-1275. Support for host bindings will be added in a follow-up, because this PR is somewhat large as it is.
PR Close#30015
Previous to this change, we assumed embedded views could only be created after
their parent template node had completed processing. As a result, we only set
up query logic for containers after directives on the node were created.
However, this assumption didn't take into account the case where a directive
on a template node could create views in its constructor.
This commit fixes query logic to work with views created in constructors.
In that case, we need to create a query container before the new view is
rendered so query results in the view aren't lost. But since the query container
is created before directives have completed processing, we also have to ensure
that query results gathered later on the template node are inserted before that
query container. Otherwise, query results in embedded views will clobber query
results on template nodes.
This splice mode may be slightly slower than the normal matching for queries on
containers, but we should only fall back to this strategy in the edge case where
views are created in constructors. (We should encourage developers to create
views in ngOnInit instead).
PR Close#29983
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
- Removes `@publicApi` annotation from ivy instructions
- Adds new `@codeGenApi` annotation to ivy instructions
- Updates ts_api_guardian to support the new annotation properly
PR Close#29820
Currently in Ivy we pass both the raw and parsed selectors to the projectionDef instruction, because the parsed selectors are used to match most nodes, whereas the raw ones are used to match against nodes with the ngProjectAs attribute. The raw selectors add a fair bit of code that won't be used in most cases, because ngProjectAs is somewhat rare.
These changes rework the compiler not to output the raw selectors in the projectionDef, but to parse the selector in ngProjectAs and to store it on the TAttributes. The logic for matching has also been changed so that it matches the pre-parsed ngProjectAs selector against the list of projection selectors.
PR Close#29578
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, 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
- Remove an extra type `ViewOrElement`, which even had the same numeric value as `View`.
- Updates comment to remove part about alleged bit-masking that we could be doing here.
We aren't using this with bitmasks, and if we were, everything would be a `NodeType.Container`,
because it's value was `0`.
- Updates the number values to be simple, human-readable integers, since we're not using these
with any kind of bit-manipulation.
- Add comments about each type.
PR Close#29343
Angular Ivy interprets inline static style/class attribute values as instructions that
are processed whilst an element gets created. Because these inline style values are
referenced by style/class bindings, their inline style values are applied at a later
stage. Despite them being eventually applied, their values should be applied earlier
before any directives are instantiated so that directive code can rely on any inline
style/class changes.
This patch ensures that all static style/class attribute values are applied (rendered)
on the element before directives are instantiated.
Jira Issue: FW-1133
PR Close#29269
The content projection mechanism is static, in that it only looks at the static
template nodes before directives are matched and change detection is run.
When you have a selector-based content projection the selection is based
on nodes that are available in the template.
For example:
```
<ng-content selector="[some-attr]"></ng-content>
```
would match
```
<div some-attr="..."></div>
```
If you have an inline-template in your projected nodes. For example:
```
<div *ngIf="..." some-attr="..."></div>
```
This gets pre-parsed and converted to a canonical form.
For example:
```
<ng-template [ngIf]="...">
<div some-attr=".."></div>
</ng-template>
```
Note that only structural attributes (e.g. `*ngIf`) stay with the `<ng-template>`
node. The other attributes move to the contained element inside the template.
When this happens in ivy, the ng-template content is removed
from the component template function and is compiled into its own
template function. But this means that the information about the
attributes that were on the content are lost and the projection
selection mechanism is unable to match the original
`<div *ngIf="..." some-attr>`.
This commit adds support for this in ivy. Attributes are separated into three
groups (Bindings, Templates and "other"). For inline-templates the Bindings
and "other" types are hoisted back from the contained node to the `template()`
instruction, so that they can be used in content projection matching.
PR Close#29041
This commit adds a new `AttributeMarker` type that will be used, in a
future commit, to mark attributes as coming from an inline-template
expansion, rather than the element that is being contained in the template.
PR Close#29041
Angular supports having a component extend off of a parent component.
When this happens, all annotation-level data is inherited including styles
and classes. Up until now, Ivy only paid attention to static styling
values on the parent component and not the child component. This patch
ensures that both the parent's component and child component's styling
data is merged and rendered accordingly.
Jira Issue: FW-1081
PR Close#29015
For efficiency reasons we often put several different data types (`RNode`, `LView`, `LContainer`,
`StylingContext`) in same location in `LView`. This is because we don't want to pre-allocate
space
for it because the storage is sparse. This file contains utilities for dealing with such data
types.
How do we know what is stored at a given location in `LView`.
- `Array.isArray(value) === false` => `RNode` (The normal storage value)
- `Array.isArray(value) === true` => than the `value[0]` represents the wrapped value.
- `typeof value[TYPE] === 'object'` => `LView`
- This happens when we have a component at a given location
- `typeof value[TYPE] === 'number'` => `StylingContext`
- This happens when we have style/class binding at a given location.
- `typeof value[TYPE] === true` => `LContainer`
- This happens when we have `LContainer` binding at a given location.
NOTE: it is assumed that `Array.isArray` and `typeof` operations are very efficient.
PR Close#28947
`LView`, `LContainer`, `StylingContext` are all arrays which wrap either
an `HTMLElement`, `LView`, `LContainer`, `StylingContext`. It is often
necessary to retrieve the correct type of element from the location
which means that we often have to wrap the arrays. Logically it makes
more sense if the thing which we are wrapping is at `0` location. Also
it may be more performant since data is more local which may result in
more L2 cache hits in CPU.
PR Close#28947
- 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
- Removes CONTAINER_INDEX
- LView[PARENT] now contains LContainer when necessary
- Removes now unused arguments to methods after refactor
PR Close#28382
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
