Parsing styling is now simplified to be used like so:
```
for (let i = parseStyle(text); i <= 0; i = parseStyleNext(text, i)) {
const key = getLastParsedKey();
const value = getLastParsedValue();
...
}
```
This change makes it easier to invoke the parser from other locations in the system without paying the cost of creating and iterating over `Map` of styles.
PR Closes#34418
This change moves information from instructions to declarative position:
- `ɵɵallocHostVars(vars)` => `DirectiveDef.hostVars`
- `ɵɵelementHostAttrs(attrs)` => `DirectiveDef.hostAttrs`
When merging directives it is necessary to know about `hostVars` and `hostAttrs`. Before this change the information was stored in the `hostBindings` function. This was problematic, because in order to get to the information the `hostBindings` would have to be executed. In order for `hostBindings` to be executed the directives would have to be instantiated. This means that the directive instantiation would happen before we had knowledge about the `hostAttrs` and as a result the directive could observe in the constructor that not all of the `hostAttrs` have been applied. This further complicates the runtime as we have to apply `hostAttrs` in parts over many invocations.
`ɵɵallocHostVars` was unnecessarily complicated because it would have to update the `LView` (and Blueprint) while existing directives are already executing. By moving it out of `hostBindings` function we can access it statically and we can create correct `LView` (and Blueprint) in a single pass.
This change only changes how the instructions are generated, but does not change the runtime much. (We cheat by emulating the old behavior by calling `ɵɵallocHostVars` and `ɵɵelementHostAttrs`) Subsequent change will refactor the runtime to take advantage of the static information.
PR Close#34683
This adds `insertTStyleValue` but does not hook it up to anything yet.
The purpose of this function is to create a linked-list of styling
related bindings. The bindings can be traversed during flush.
The linked list also keeps track of duplicates. This is important
for binding to know if it needs to check other styles for reconciliation.
PR Close#34004
This change introduces class/style reconciliation algorithm for DOM elements.
NOTE: The code is not yet hooked up, it will be used by future style algorithm.
Background:
Styling algorithm currently has [two paths](https://hackmd.io/@5zDGNGArSxiHhgvxRGrg-g/rycZk3N5S)
when computing how the style should be rendered.
1. A direct path which concatenates styling and uses `elemnent.className`/`element.style.cssText` and
2. A merge path which uses internal data structures and uses `element.classList.add/remove`/`element.style[property]`.
The situation is confusing and hard to follow/maintain. So a future PR will remove the merge-path and do everything with
direct-path. This however breaks when some other code adds class or style to the element without Angular's knowledge.
If this happens instead of switching from direct-path to merge-path algorithm, this change provides a different mental model
whereby we always do `direct-path` but the code which writes to the DOM detects the situation and reconciles the out of bound write.
The reconciliation process is as follows:
1. Detect that no one has modified `className`/`cssText` and if so just write directly (fast path).
2. If out of bounds write did occur, switch from writing using `className`/`cssText` to `element.classList.add/remove`/`element.style[property]`.
This does require that the write function computes the difference between the previous Angular expected state and current Angular state.
(This requires a parser. The advantage of having a parser is that we can support `style="width: {{exp}}px" kind of bindings.`)
Compute the diff and apply it in non destructive way using `element.classList.add/remove`/`element.style[property]`
Properties of approach:
- If no out of bounds style modification:
- Very fast code path: Just concatenate string in right order and write them to DOM.
- Class list order is preserved
- If out of bounds style modification detected:
- Penalty for parsing
- Switch to non destructive modification: `element.classList.add/remove`/`element.style[property]`
- Switch to alphabetical way of setting classes.
PR Close#34004
Fixes Ivy detecting changes inside child embedded views, even though they're detached.
Note that there's on subtlety here: I made the changes inside `refreshDynamicEmbeddedViews` rather than `refreshView`, because we support detecting changes on a detached view (evidenced by a couple of unit tests), but only if it's triggered directly from the view's `ChangeDetectorRef`, however we shouldn't be detecting changes in the detached child view when something happens in the parent.
Fixes#34816.
PR Close#34846
This patch removes the need for the styleSanitizer() instruction in
favor of passing the sanitizer into directly into the styleProp
instruction.
This patch also increases the binding index size for all style/class bindings in preparation for #34418
PR Close#34480
remove unnecessary underscore suffix and the corresponding TODO comments,
because the rollup bug was fixed: github.com/rollup/rollup/issues/2047
PR Close#34757
Unlike in View Engine, we currently reset the dirty state of
components in the check no changes change detection cycle.
This means that components cannot be marked as dirty from
view lifecycle hooks because the dirty state is reset and
the lifecycle hooks do not run in the check no changes CD cycle.
PR Close#34495
A while ago we made a pass through all instructions to make sure that none of them call directly into other instructions, however it seems like missed the `pipeBind*` since they still call into the pure functions. The result is that we have some unnecessary duplicated accesses of global state like `getLView` which are called twice in a row with nothing changing.
These changes move the common functionality into a shared file and make the pipe instructions call into them with the global state instead.
PR Close#33714
This commit improves `ExpressionChangedAfterChecked` error message for attributes by including attribute name and the content of the entire expression that contains interpolation(s). In order to achieve that, metadata is now stored in `TData` array when `attribute` and `attributeInterpolate` instructions are being called (similar to `property` and `propertyInterpolate` instructions).
PR Close#34505
Follow-up from [this discussion](https://github.com/angular/angular/pull/33419#discussion_r339296216). In Ivy we don't use the schema to validate tag names, but instead we use feature detection to figure out whether an element is supported. While this should generally be more accurate, it'll also end up throwing for some more innocent cases. E.g. now Ivy throws an error for `main` elements in IE which is accurate since IE doesn't support the element, but is annoying since there is no functionality attached.
These changes switch to logging a warning instead, similarly to what we're doing for unknown properties.
PR Close#34524
The main logic of the `InheritDefinitionFeature` is to go through the prototype chain of a given Component and merge all Angular-specific information onto that Component def. The problem happens in case there is a Component in a hierarchy that also contains the `InheritDefinitionFeature` (i.e. it extends some other Component), so it inherits all Angular-specific information from its super class. As a result, the root Component may end up having duplicate information inherited from different Components in hierarchy.
Let's consider the following structure: `GrandChild` extends `Child` that extends `Base` and the `Base` class has a `HostListener`. In this scenario `GrandChild` and `Child` will have `InheritDefinitionFeature` included into the `features` list. The processing will happend in the following order:
- `Child` inherits `HostListener` from the `Base` class
- `GrandChild` inherits `HostListener` from the `Child` class
- since `Child` has a parent, `GrandChild` also inherits from the `Base` class
The result is that the `GrandChild` def has duplicated host listener, which is not correct.
This commit introduces additional logic that checks whether we came across a def that has `InheritDefinitionFeature` feature (which means that this def already inherited information from its super classes). If that's the case, we skip further fields-related inheritance logic, but keep going though the prototype chain to look for super classes that contain other features (like NgOnChanges), that we need to invoke for a given Component def.
PR Close#34244
Prior to this change, the ExpressionChangedAfterChecked error thrown in Ivy was missing useful information that was available in View Engine, specifically: missing property name for proprty bindings and also the content of the entire property interpolation (only a changed value was displayed) if one of expressions was changed unexpectedly. This commit improves the error message by including the mentioned information into the error text.
PR Close#34381
We have a couple of cases where we use something like `typeof Node === 'function'` to figure out whether we're in a worker context. This works in most browsers, but IE returns `object` instead of `function`. I've updated all the usages to account for it.
PR Close#34305
TypeScript 3.7 flags `if` conditions that implicitly coerce a function/method definition. While checking for the `template` presence on a def (actually checking whether we work with Component) in `saveNameToExportMap`, the `if` condition had implicit type coercion. This commit updates the condition to use the `isComponentDef` function (that checks `def.template` against `null` internally) to avoid compilation errors with TypeScript 3.7.
PR Close#34335
Occasionally a factory function needs to be generated for an "invalid"
constructor (one with parameters types which aren't injectable). Typically
this happens in JIT mode where understanding of parameters cannot be done in
the same "up-front" way that the AOT compiler can.
This commit changes the JIT compiler to generate a new `invalidFactoryDep`
call for each invalid parameter. This instruction will error at runtime if
called, indicating both the index of the invalid parameter as well as (via
the stack trace) the factory function which was generated for the type being
constructed.
Fixes#33637
PR Close#33739
Prior to this commit, calling change detection for destroyed views resulted in errors being thrown in some cases. This commit adds a check to make sure change detection is invoked for non-destroyed views only.
PR Close#34241
Prior to this commit, i18n runtime code failed with the exception saying that no provider was found for ChangeDetectorRef for a pipe used in ICU. The problem happened because the underlying `createViewRef` function was not taking into account IcuContainer as a valid TNodeType. This commit updates the `createViewRef` function to return corresponding ViewRef for TNodeType.IcuContainer.
PR Close#34198
Includes a few minor performance improvements:
* In the `nextContext` instruction we assign a new LView to the `LFrame.contextLView` and then we immediately look it up to get its context. We don't need to do that since we know the view that was assigned already.
* Removes the default value for the `level` parameter of `nextContextImpl` because it generates more code in es5 and is internal-only.
* Removes the default parameter from `setActiveHostElement` since it generates extra code and it's an internal function.
* Makes a check in `setElementExitFn` more strict since we're guaranteed for the value to match the type.
PR Close#34183
Prior to this commit, the unknown element can happen twice for AOT-compiled components: once during compilation and once again at runtime. Due to the fact that `schemas` information is not present on Component and NgModule defs after AOT compilation, the second check (at runtime) may fail, even though the same check was successful at compile time. This commit updates the code to avoid the second check for AOT-compiled components by checking whether `schemas` information is present in a logic that executes the unknown element check.
PR Close#34024
Since config=ivy now sets the define=compile flag and the define=angular_ivy_enabled
flag to cause usage of Ivy, we can update all of the documentation and scripts that
reference compile=aot to use config=ivy.
PR Close#33983
Prior to this commit, there was a runtime check in i18n logic to make sure "other" case is always present in an ICU. That was not a requirement in View Engine, so ICUs that previously worked may produce errors. This commit removes that restriction and adds support for ICUs without "other" cases.
PR Close#34042
Move a view only if it would end up at a different place.
Otherwise we would do unnecessary processing like DOM manipulation, query notifications etc.
Thanks to @pkozlowski-opensource for the change.
PR Close#34052
When inserting a `viewRef` it is possible to not provide
an `index`, which is regarded as appending to the end of
the container.
If the `viewRef` already exists in the container, then
this results in a move. But there was a fault in the logic
that computed where to insert the `viewRef` that did not
account for the fact that the `viewRef` was already in
the container, so the insertion `index` was outside the
bounds of the array.
Fixes#33924
PR Close#34052
Previously if a type was returning itself it would cause an infinite loop in tsickle. We worked around it with a type that alises to `any`. Now that the issue has been resolved in tsickle, we can clean up the workaround.
PR Close#34019
This commit fixes a compatibility bug where pre-order lifecycle
hooks (onInit, doCheck, OnChanges) for directives on the same
host node were executed based on the order the directives were
matched, rather than the order the directives were instantiated
(i.e. injection order).
This discrepancy can cause issues with forms, where it is common
to inject NgControl and try to extract its control property in
ngOnInit. As the NgControl directive is injected, it should be
instantiated before the control value accessor directive (and
thus its hooks should run first). This ensures that the NgControl
ngOnInit can set up the form control before the ngOnInit
for the control value accessor tries to access it.
Closes#32522
PR Close#34026
The root view case is already covered by the existing code in the
getRenderParent function so no need to have an explicit checks
(and associated memory reads) again.
PR Close#33988
The assertion that we have in the `directiveInject` instruction is too restrictive and we came across some pattern where it throws unnecessarily. This commit removes that assertion for now and more detailed investigation is needed to decide is we need to restrict the set of TNodeType again.
This commit also adds a test which triggered the TNodeType.View to come up in the `directiveInject` instruction, so it might be useful to avoid regressions during further refactoring.
PR Close#33948
NgModule compilation in JIT mode (that is also used in TestBed) caches module scopes on NgModule defs (using `transitiveCompileScopes` field). Module overrides (defined via TestBed.overrideModule) may invalidate this data by adding/removing items in `declarations` list. This commit forces TestBed to recalculate transitive scopes in case module overrides are present, so TestBed always gets the most up-to-date information.
PR Close#33787
NgModules in Ivy have a definition which contains various different bits
of metadata about the module. In particular, this metadata falls into two
categories:
* metadata required to use the module at runtime (for bootstrapping, etc)
in AOT-only applications.
* metadata required to depend on the module from a JIT-compiled app.
The latter metadata consists of the module's declarations, imports, and
exports. To support JIT usage, this metadata must be included in the
generated code, especially if that code is shipped to NPM. However, because
this metadata preserves the entire NgModule graph (references to all
directives and components in the app), it needs to be removed during
optimization for AOT-only builds.
Previously, this was done with a clever design:
1. The extra metadata was added by a function called `setNgModuleScope`.
A call to this function was generated after each NgModule.
2. This function call was marked as "pure" with a comment and used
`noSideEffects` internally, which causes optimizers to remove it.
The effect was that in dev mode or test mode (which use JIT), no optimizer
runs and the full NgModule metadata was available at runtime. But in
production (presumably AOT) builds, the optimizer runs and removes the JIT-
specific metadata.
However, there are cases where apps that want to use JIT in production, and
still make an optimized build. In this case, the JIT-specific metadata would
be erroneously removed. This commit solves that problem by adding an
`ngJitMode` global variable which guards all `setNgModuleScope` calls. An
optimizer can be configured to statically define this global to be `false`
for AOT-only builds, causing the extra metadata to be stripped.
A configuration for Terser used by the CLI is provided in `tooling.ts` which
sets `ngJitMode` to `false` when building AOT apps.
PR Close#33671
TNode.inputs are initialised during directives resolution now so we know early
if a node has directives with inputs or no. We don't need to use undefined value
as an indicator that inputs were not resolved yet.
PR Close#33798
Before this change a public name of a directive's input
was stored in 2 places:
- as a key of an object on TNode.index;
- as a value of PropertyAliasValue at the index 1
This PR changes the data structure so the public name is stored
only once as a key on TNode.index. This saves one array entry
for each and every directive input.
PR Close#33798
Adds support for chaining of `styleProp`, `classProp` and `stylePropInterpolateX` instructions whenever possible which should help generate less code. Note that one complication here is for `stylePropInterpolateX` instructions where we have to break into multiple chains if there are other styling instructions inbetween the interpolations which helps maintain the execution order.
PR Close#33837
