Previously, all asset-groups from `ngsw-config.json` were processed in
parallel. For each asset-group, we retrieved all files for the current
build, filtered out files that were already matched by other
asset-groups, determined which of the remaining files belonged to the
current asset-group and generated entries for the `ngsw.json` manifest.
This process was susceptible to race conditions when there were files
that would be matched by multiple asset-groups. This made the generation
of the `ngsw.json` manifest non-deterministic and violated the rule that
each file would belong to the first asset-group that matched it (based
on the asset-groups' order of appearance in `ngsw-config.json`), thus
leading to broken ServiceWorker behavior.
This commit fixes it by ensuring that the generation process is
deterministic and that asset-groups are processed in the proper order.
NOTE 1:
The generation process has been broken since the beginning, but we have
only noticed this recently. This is possibly related to the CLI's
switching from a virtual file system host (which has more consistent
timing characteristics) to the Node.js built-in `fs.promises` in
angular/angular-cli@d3bc530c10.
NOTE 2:
This commit also ensures that files in the `ngsw.json` hash-table are in
alphabetic order. Previously, the files were added to the hash-table in
blocks corresponding to each asset-group.
This change is not necessary (i.e. the order of keys in the hash-table
makes no difference in behavior), but it makes it easier to scan for a
file (for example, for debugging purposes).
PR Close#43679
Previously, when a router-outlet is conditionally shown with an ngIf, and a sub-route was re-attached
via a custom RouteReuseStrategy, router-outlet would try to mount a destroyed component into the view
if the router-outlet is destroyed and re-initialized.
This commit fixes it by unsetting context.attachRef when router-outlet is destroyed, so when the
router-outlet is being initialized again, it no longer sees an attachRef that it needs to mount to the
view.
Fixes#43696
PR Close#43697
This commit fixes an oversight in the JIT compilation of partial factory
declarations, where the literal `'invalid'` was not accounted for
(unlike the AOT linker).
Fixes#43609
PR Close#43619
Many of the tests in the router code use the deprecated loadChildren as a string. This
has been deprecated for years and can easily be changed to just a function that
returns the module.
PR Close#43578
When specifying the `deps` array in the `@Injectable` decorator to
inject dependencies into the injectable's factory function, it should
be possible to use an array literal to configure how the dependency
should be resolved by the DI system.
For example, the following example is allowed:
```ts
@Injectable({
providedIn: 'root',
useFactory: a => new AppService(a),
deps: [[new Optional(), 'a']],
})
export class AppService {
constructor(a) {}
}
```
Here, the `'a'` string token should be injected as optional. However,
the AOT compiler incorrectly used the array literal itself as injection
token, resulting in a failure at runtime. Only if the token were to be
provided using `[new Optional(), new Inject('a')]` would it work
correctly.
This commit fixes the issue by using the last non-decorator in the
array literal as the token value, instead of the array literal itself.
Note that this is a loose interpretation of array literals: if a token
is omitted from the array literal then the array literal itself is used
as token, but any decorator such as `new Optional()` would still have
been applied. When there's multiple tokens in the list then only the
last one will be used as actual token, any prior tokens are silently
ignored. This behavior mirrors the JIT interpretation so is kept as is
for now, but may benefit from some stricter checking and better error
reporting in the future.
Fixes#42987
PR Close#43226
Prior to this commit ngcc stored its package configuration in JavaScript
objects, which caused the builtin `Object` members to be found as
package configuration. This would subsequently crash as their shape was
not as expected.
This commit moves away from using raw JavaScript objects in favor of a
Map. To code was refactored such that `PartiallyProcessedConfig` is
now a class.
Fixes#43570
PR Close#43589
Native DOM events were previously not included in the completions
because the dom schema registry would filter out events completely. This
change updates the registry to include events in the private
element->property map and excludes events from lookups outside of the
new `allKnownEventsOfElement` function.
fixes https://github.com/angular/vscode-ng-language-service/issues/1479
PR Close#43299
Previously, when a version was found to be broken, any clients assigned
to that version were unassigned (and either assigned to the latest
version or to none if the latest version was the broken one). A version
could be considered broken for several reasons, but most often it is a
response for a hashed asset that eiher does not exist or contains
different content than the SW expects. See
https://github.com/angular/angular/issues/28114#issuecomment-923122967
for more details.
However, assigning a client to a different version (or the network) in
the middle of a session, turned out to be more risky than keeping it on
the same version. For angular.io, for example, it has led to #28114.
This commit avoids making things worse when identifying a broken version
by keeping existing clients to their assigned version (but ensuring that
no new clients are assigned to the broken version).
NOTE:
Reloading the page generates a new client ID, so it is like a new client
for the SW, even if the tab and URL are the same.
PR Close#43518
Previously, a ServiceWorker test was using the
`MockServerState#sawRequestFor()` method. This method, however, only
returns a boolean indicating whether a request had been seen, but would
not throw an error in either case.
Since the intention was to assert that a specific request was made, this
commit switches to the correct `MockServerState#assertSawRequestFor()`
method.
PR Close#43518
Adds a test to the nullish coalescing diagnostic check to serve as
self-documentation on how it works with nullish coalescing on pipes that
are often misconfigured.
This also removes that non null assertion operator, which is incorrect
because there _are_ situations where a symbol cannot be retrieved.
PR Close#43419
The template type-checker has to emit type constructors for the
directives that are used in a template, where a type constructor's
declaration has to mirror the type parameter constraints as they were
originally declared. Therefore, the compiler analyzes whether a type
parameter constraint can be recreated, e.g. by generating imports for
any type references. Some type references cannot be recreated, in which
case the compiler has to fall back to a strategy where the type
constructor is created inline in the original source file (which comes
with a performance penalty).
There used to be an issue for type references to namespaced declarations.
The compiler is unable to emit such references such that an inline
type constructor should be used as fallback, but this did not happen.
This caused the attempt to emit the type reference to fail, as the
namespaced declaration cannot be located by the reference emitters.
This commit fixes the issue by using a stricter check to determine if a
type parameter requires an inline type constructor. The TypeScript
reflection host's `isStaticallyExported` logic was expanded to work for
any declaration instead of just classes, as e.g. type declarations can
also be referenced in a type parameter constraint.
Closes#43383
PR Close#43511
These token interfaces will make it easier to reason about tokens in the
parser and in specs.
Previously, it was never clear what items could appear in the `parts`
array of a token given a particular `TokenType`. Now, each token interface
declares a labelled tuple for the parts, which helps to document the token
better.
PR Close#43132
Previously, the way templates were tokenized meant that we lost information
about the location of interpolations if the template contained encoded HTML
entities. This meant that the mapping back to the source interpolated strings
could be offset incorrectly.
Also, the source-span assigned to an i18n message did not include leading
whitespace. This confused the output source-mappings so that the first text
nodes of the message stopped at the first non-whitespace character.
This commit makes use of the previous refactorings, where more fine grain
information was provided in text tokens, to enable the parser to identify
the location of the interpolations in the original source more accurately.
Fixes#41034
PR Close#43132
The tests were checking that the source-span of parsed HTML nodes were
accurate, but they were not checking the span when it includes the
"leading trivia", which are given by the `fullStart` rather than `start`
location.
PR Close#43132
When it was tokenized, text content is split into parts that can include
interpolations and encoded entities tokens.
To make this information available to downstream processing, this commit
adds these tokens to the `Text` AST nodes, with suitable processing.
PR Close#43132
The lexer now splits encoded entity tokens out from text and attribute value tokens.
Previously encoded entities would be decoded and the decoded value would be
included as part of the text token of the surrounding text. Now the entities
have their own tokens. There are two scenarios: text and attribute values.
Previously the contents of `<div>Hello & goodbye</div>` would be a single
TEXT token. Now it will be three tokens:
```
TEXT: "Hello "
ENCODED_ENTITY: "&", "&"
TEXT: " goodbye"
```
Previously the attribute value in `<div title="Hello & goodbye">` would be
a single text token. Now it will be three tokens:
```
ATTR_VALUE_TEXT: "Hello "
ENCODED_ENTITY: "&", "&"
ATTR_VALUE_TEXT: " goodbye"
```
- ENCODED_ENTITY tokens have two parts: "decoded" and "encoded".
- ENCODED_ENTITY tokens are always preceded and followed by either TEXT tokens
or ATTR_VALUE_TEXT tokens, depending upon the context, even if they represent
an empty string.
The HTML parser has been modified to recombine these tokens to allow this
refactoring to have limited effect in this commit. Further refactorings
to use these new tokens will follow in subsequent commits.
PR Close#43132
The lexer now splits interpolation tokens out from text tokens.
Previously the contents of `<div>Hello, {{ name}}<div>` would be a single
text token. Now it will be three tokens:
```
TEXT: "Hello, "
INTERPOLATION: "{{", " name", "}}"
TEXT: ""
```
- INTERPOLATION tokens have three parts, "start marker", "expression"
and "end marker".
- INTERPOLATION tokens are always preceded and followed by TEXT tokens,
even if they represent an empty string.
The HTML parser has been modified to recombine these tokens to allow this
refactoring to have limited effect in this commit. Further refactorings
to use these new tokens will follow in subsequent commits.
PR Close#43132
Previously, the regular expression used by the compiler's ShadowDOM CSS
shim to strip comments from CSS text was susceptible to [catastrophic
backtracking][1], which could lead to exponential (O(2^n)) increase in
complexity/execution time. More specifically, this would be triggered if
the processed text contained an unterminated comment with lots of
leading whitespace (i.e. `/*`, followed by lots of whitespace characters
and no closing `*/`).
Although such input is unlikely in real-world scenarios, this commit
improves the regular expression to not be susceptible to this issue.
[1]: https://www.regular-expressions.info/catastrophic.html
PR Close#43456
No test was added for this case because I can't think of a test case to
write that would work.
This change updates the code for a consistent mental model of setting
the `browserUrlTree`. It's only meant to track the `UrlTree` that the
`UrlHandlingStrategy` is set to extract, not the full `rawUrl`. Notice
that everywhere else, the `browserUrlTree` is set to
`urlAfterRedirects`, which is computed based on the extracted URL, _not_
the `rawUrl`.
PR Close#43426
There are situations where the Router does not currently clean up failed navigations
correctly. While this is problematic on its own, we can mitigate some of
the damage by processing any URL when we get a navigation request when
the internal router state is out of sync.
Each of the added tests would fail without this change.
fixes#34795
PR Close#43424
* Document `currentUrlTree`, `browserUrlTree`, and `rawUrlTree`
* Add a couple `TODO` comments to investigate changes based on understanding
PR Close#43391
This restores the `finalize` function to directly call
`resetUrlToCurrentUrlTree`, as it was before efb440eb2f.
This allows us to simplify the `restoreHistory` function because it no
longer needs to handle the call from `finalize` where it should not
reset the internal router state.
PR Close#43382
The TODO comment suggesting to verify that the target element to
scroll to needs to be an anchor does not longer seems under consideration
so it can be removed
resolves#43348
PR Close#43378
Previously, the error message was not added to `Unhandled Navigation Error`
Before
```
main.js:1 Unhandled Navigation Errors:
```
After
```
main.js:1 Unhandled Navigation Errors: SecurityError: Failed to execute 'replaceState' on 'History': A history state object with URL 'http://localhost:4200/' cannot be created in a document with origin 'http://127.0.0.1:8080' and URL 'http://127.0.0.1:8080/'.
```
PR Close#43291
Add an abstract class that has a default implementation of the run
function and visits all nodes. Authors of template checks can extend
this class and override the `visitNode` method to return diagnostics.
Refs #42966
PR Close#43232
Currently the error message functions are defined as static methods on a class which means that as soon as one of them is used somewhere, all of them have to be retained. This isn't a problem at the moment, because all of them are behind `ngDevMode` checks, but it's error prone and it's easy to fix.
These changes move them out into functions so that they can be imported individually. It also has the advantage of allowing Webpack to minify the function names.
PR Close#43223
The current naming causes tsickle to emit `/** @implements {ZoneDelegate} */ class ZoneDelegate ...`, which is a cyclic definition error in Closure Compiler. Additionally, the name shadowing is also confusing to humans.
PR Close angular#43019
PR Close#43019
Previously, the decorator transformer was annotating the synthesized properties with TS type annotations. However, because it ran after the JSDoc transformer, the TS types were just dropped from the emitted JS. Attempting to move the decorator transformer before the JSDoc transformer causes tsickle crashes because synthetic AST fragments are not attached to a SourceFile node.
PR Close#43021
Kristiyan Kostadinov