Commit Graph

19417 Commits

Author SHA1 Message Date
Kristiyan Kostadinov 7e33cb9626 fix(compiler-cli): generating invalid setClassMetadata call in ES5 for class with custom decorator (#39527)
When a class with a custom decorator is transpiled to ES5, it looks something like this:

```
var SomeClass = (function() {
  function SomeClass() {...};
  var SomeClass_1 = __decorate([Decorator()], SomeClass);
  SomeClass = SomeClass_1;
  return SomeClass;
})();
```

The problem is that if the class also has an Angular decorator that refers to the class itself
(e.g. `{provide: someToken, useClass: SomeClass}`), the generated `setClassMetadata` code will
be emitted after the IIFE, but will still refer to the intermediate `SomeClass_1` variable from
inside the IIFE. This happens, because we generate the `setClassMetadata` call directly from
the source AST which contains identifiers that TS will rename when it emits the ES5 code.

These changes resolve the issue by looking through the metadata AST and cloning any `Identifier`
that is referring to the class. Since TS doesn't have references to the clone, it won't rename
it when transpiling to ES5.

Fixes #39509.

PR Close #39527
2020-11-03 14:52:59 -08:00
Joey Perrott be64f98076 fix(dev-infra): run fetches in the git client in quiet mode (#39503)
When fetch is run in normal mode, the `git-fetch-pack` plumbing command
outputs progress to stderr.  Since this is unnecessary progress
information for ng-dev usages, it should be suppressed instead.

PR Close #39503
2020-11-02 16:41:47 -08:00
Joey Perrott db1c48ff36 ci: only check testing directories within packages/ for fw-testing (#39536)
The fw-testing group should only match on testing directories within the
framework packages, previously it was also matching in other areas of
the codebase.

PR Close #39536
2020-11-02 16:39:44 -08:00
Keen Yee Liau 3cab823222 docs(bazel): Mention Alex Eagle's article on migration doc (#39507)
Alex Eagle wrote an external article on our decision to move Bazel out of
Angular repo, and it's useful for users who want to know more about what's next.

PR Close #39507
2020-11-02 13:34:35 -08:00
Pete Bacon Darwin 5f1e9758f5 build(docs-infra): ensure that deployment works on CI (#39535)
The actual "main" part of the script that is executed was using
an uninitialized variable. This is fixed and a test is added to
check.

PR Close #39535
2020-11-02 12:24:34 -08:00
Eduard Bondarenko ad62edac9c docs: update forwardRef example (#39519)
Replace the deprecated api ReflectiveInjector.resolveAndCreat with Injector.create

PR Close #39519
2020-11-02 11:40:17 -08:00
Alex Rickabaugh eaace44d57 refactor(language-service): refactor `HybridVisitor` and expand its capabilities (#39505)
This commit takes the `HybridVisitor` in the language service and gives it
the ability to return not just a node but the template context in which it
appears. In the future, more context regarding where a node appears in the
template might become necessary (ex: the microsyntax container for binding
nodes), and this refactoring enables that.

In the process, `HybridVisitor` is renamed and the concept of a
`TemplateTarget` interface is introduced to contain the results of this
operation.

PR Close #39505
2020-11-02 10:29:50 -08:00
Alex Rickabaugh 643c96184c refactor(language-service): introduce DisplayParts abstraction for Ivy (#39505)
This commit refactors the QuickInfo abstraction shared between the VE and
Ivy services and used to implement hover tooltips (quick info), which was
extracted from the VE code in commit faa81dc. The new DisplayParts
abstraction is more general and can be used to extract information needed by
various LS functions (e.g. autocompletion).

This commit effectively reverts faa81dc, returning the original code to the
VE implementation as the Ivy code is now diverged.

PR Close #39505
2020-11-02 10:29:50 -08:00
George Kalpakas cf48d508af build(docs-infra): upgrade `firebase-tools` to v8.14.1 (#39470)
This commit upgrades `firebase-tools` (used to deploy angular.io to
Firebase hosting) to the latest version. This gives us access to the
latest fixes/improvements and also new features, such as
[version cloning][1].

[1]:
https://firebase.google.com/docs/hosting/manage-hosting-resources#cli-commands-cloning

PR Close #39470
2020-11-02 07:57:51 -08:00
George Kalpakas a55a5696a1 build(docs-infra): deploy angular.io to new Firebase sites (#39470)
Previously, the documentation for each major Angular version was hosted
on each own Firebase project. This required creating a new project for
each major release and increased the administrative/maintenance cost.

Now that Firebase supports hosting [multiple websites][1] as part of the
same project, we are switching to deploying all major versions to sites
created on `angular-io` project.

This is part of the work needed to prepare angular.io for our
[new versioning/branching process][2] (also tracked in #39366).

[1]: https://firebase.google.com/docs/hosting/multisites
[2]: https://docs.google.com/document/d/197kVillDwx-RZtSVOBtPb4BBIAw0E9RT3q3v6DZkykU

PR Close #39470
2020-11-02 07:57:51 -08:00
George Kalpakas 699824a2a5 refactor(docs-infra): break up `deploy-to-firebase.js` script into functions (#39470)
This commit breaks up the code in `deploy-to-firebase.js` script, that
we use for deploying angular.io to production, to smaller functions
(instead of a monolithic block). This makes the script easier to
maintain and also makes testing individual operations easier.

The commit also updates the `deploy-to-firebase.spec.js` spec file to
take advantage of the standalone functions to speed up testing by
calling the corresponding function instead of having to spawn a new
process and run the `deploy-to-firebase.js` script with the `--dry-run`
flag.

NOTE: Before updating the tests, I verified that the updated
      `deploy-to-firebase.js` script passed the old tests.

PR Close #39470
2020-11-02 07:57:51 -08:00
George Kalpakas 1aaf556815 build(docs-infra): add support for RC deployments to deployment script (#39470)
This commit updates the angular.io deployment script
(`deploy-to-firebase.js`) to support deploying release-candidate
versions.

This is part of the work needed to prepare angular.io for our
[new versioning/branching process][1] (also tracked in #39366).

[1]: https://docs.google.com/document/d/197kVillDwx-RZtSVOBtPb4BBIAw0E9RT3q3v6DZkykU

PR Close #39470
2020-11-02 07:57:51 -08:00
George Kalpakas 51fe0fe622 build(docs-infra): add new angular.io build configuration for release-candidates (#39470)
This commit adds a new build configuration/deployment mode (`rc`) for
deploying release-candidate versions to angular.io.

This is part of the work needed to prepare angular.io for our
[new versioning/branching process][1] (also tracked in #39366).

[1]: https://docs.google.com/document/d/197kVillDwx-RZtSVOBtPb4BBIAw0E9RT3q3v6DZkykU

PR Close #39470
2020-11-02 07:57:51 -08:00
George Kalpakas 0f584694cc refactor(docs-infra): use Sass mixin to simplify the creation of deployment mode themes (#39470)
Different deployment modes (such as `archive` and `next`) are identified
by the different colors used in prominent elements of the page, such as
the topbar and the footer.

Previously, the necessary styles for creating such a deployment mode
"theme" were duplicated for each mode.

This commit simplifies the creation/modification of a deployment mode
theme by introducing a Sass mixin that generates the necessary styles
(when provided with necessary theme colors).

PR Close #39470
2020-11-02 07:57:51 -08:00
George Kalpakas eaf7d8d69f test(docs-infra): ensure `deploy-to-firebase` tests pass locally (#39470)
Previously, the `deploy-to-firebase.js` script and the accompanying
`deploy-to-firebase.spec.js` spec file were using the `origin` remote
alias in certain commands. This works fine on CI, where `origin` points
to the `angular/angular` GitHub repo, but might not work locally.

This commit ensures that the correct remote is used by explicitly
specifying it by the URL, thus ensuring that the tests will behave
identically on CI and locally.

PR Close #39470
2020-11-02 07:57:51 -08:00
George Kalpakas 5143d52d2b build(docs-infra): switch `deploy-to-firebase.sh` script to JS (#39470)
This commit switches the `deploy-to-firebase.sh` script, that we use for
deploying angular.io to production, from Bash to JavaScript. This makes
the script easier to maintain.

For the same reasons, it also switches the `deploy-to-firebase.test.sh`
script, that we use for testing the `deploy-to-firebase` script, from
Bash to JavaScript (using jasmine as the test runner).

Finally, this commit also updates ShellJS to the latest version to get
better error messages (including the actual error) when `exec()` fails.

NOTE: Before switching the test script to JS, I verified that the new
      `deploy-to-firebase.js` script passed the tests with the old
      `deploy-to-firebase.test.sh` script.

PR Close #39470
2020-11-02 07:57:51 -08:00
JoostK 8317042483 perf(core): do not recurse into modules that have already been registered (#39514)
When registering an NgModule based on its id, all transitively imported
NgModules are also registered. This commit introduces a visited set to
avoid traversing into NgModules that are reachable from multiple import
paths multiple times.

Fixes #39487

PR Close #39514
2020-11-02 07:51:18 -08:00
JoostK 40bf1e0475 test(compiler-cli): remove spurious `console.error` call from a test (#39321)
Reporting the source file text to the console was left as a debugging
artifact.

PR Close #39321
2020-11-02 07:50:41 -08:00
JoostK 5d731354d0 perf(compiler-cli): only generate template context declaration when used (#39321)
The variable declaration for a template context is only needed when it
is referenced from somewhere, so the TCB operation to generate the
declaration is marked as optional.

PR Close #39321
2020-11-02 07:50:41 -08:00
JoostK 3b0b7d2210 fix(compiler-cli): report missing pipes when `fullTemplateTypeCheck` is disabled (#39320)
Even if `fullTemplateTypeCheck` is disabled should missing pipes still
be reported, as was the case in View Engine.

Fixes #38195

PR Close #39320
2020-10-30 18:01:51 -07:00
Andrew Scott 3af585c004 fix(router): Ensure all outlets are used when commands have a prefix (#39456)
When there is a primary outlet present in the outlets map and the object is also prefixed
with some other commands, the current logic only uses the primary outlet and ignores
the others. This change ensures that all outlets are respected at the
segment level when prefixed with other commands.

PR Close #39456
2020-10-30 16:23:03 -07:00
Andrew Scott ff7a62ee21 refactor(router): Small refactor of createUrlTree and extra tests (#39456)
This commit has a small refactor of some methods in create_url_tree.ts
and adds some test cases, including two that will fail at the moment but
should pass. A follow-up commit will make use of the refactorings to fix
the test with minimal changes.

PR Close #39456
2020-10-30 16:23:03 -07:00
Alan Agius f12157c145 docs: update CLI workspace config documentation (#39427)
- Improves JSON formatting
- Add reference to font optimization
- Removes `="true"` from boolean command line args.
These are redundant and it can be confusing to why
boolean values need to be provided via a CLI.

PR Close #39427
2020-10-30 12:39:30 -07:00
Kristiyan Kostadinov cbc0907bfd fix(compiler): preserve this.$event and this.$any accesses in expressions (#39323)
Currently expressions `$event.foo()` and `this.$event.foo()`, as well as `$any(foo)` and
`this.$any(foo)`, are treated as the same expression by the compiler, because `this` is considered
the same implicit receiver as when the receiver is omitted. This introduces the following issues:

1. Any time something called `$any` is used, it'll be stripped away, leaving only the first parameter.
2. If something called `$event` is used anywhere in a template, it'll be preserved as `$event`,
rather than being rewritten to `ctx.$event`, causing the value to undefined at runtime. This
applies to listener, property and text bindings.

These changes resolve the first issue and part of the second one by preserving anything that
is accessed through `this`, even if it's one of the "special" ones like `$any` or `$event`.
Furthermore, these changes only expose the `$event` global variable inside event listeners,
whereas previously it was available everywhere.

Fixes #30278.

PR Close #39323
2020-10-30 10:49:15 -07:00
Andrew Kushnir 4e68254514 fix(common): change the week-numbering year format from `r` -> `Y` (#39495)
This commit updates the week-numbering year format from `r` -> `Y` based on the description in
http://www.unicode.org/reports/tr35/tr35-dates.html#dfst-year.

Note: this is not a breaking change, since the week-numbering year format was introduced in
v11.0.0-next.3 (984ed39195)
and the major version that contains that change was not released yet.

PR Close #39495
2020-10-30 10:24:43 -07:00
Andrew Scott beb935613e refactor(compiler-cli): Store inline templates and styles in the resource registry (#39482)
The Language Service is not only interested in external resources, but
also inline styles and templates. By storing the expression of the
inline resources, we can more easily determine if a given position is
part of the inline template/style expression.

PR Close #39482
2020-10-30 09:58:21 -07:00
ayazhafiz 3241d922fc refactor(compiler): parse bindings "by hand" rather than via regex (#39375)
To support recovery of malformed binding property names like `([a)`,
`[a`, or `()`, the binding parser needs to be more permissive w.r.t. the
kinds of bindings it can detect. This is difficult to do maintainably
with a regex, but is trivial with a "hand-rolled" string parser. This
commit refactors render3's binding attribute parsing to use this method
for multi-delimited bindings (namely via the `()`, `[]`, and `[()]`)
syntax, making the way recovery of malformed bindings in a future patch.

Note that we can keep using a regex for prefix-only binding syntax
(e.g. `bind-`, `ref-`) because validation of the binding is complete
once we have matched the prefix, and the only thing left to do is check
that the binding identifier is non-empty, which is trivial.

Part of #38596

PR Close #39375
2020-10-30 09:40:56 -07:00
Andrew Kushnir c83b2ad87f docs: tView.preOrderHooks and tView.preOrderCheckHooks docs update (#39497)
This commit updates the docs for the `tView.preOrderHooks` and `tView.preOrderCheckHooks` TView
fields. Current docs are not up-to-date as it was pointed out in #39439.

Closes #39439.

PR Close #39497
2020-10-29 16:25:39 -07:00
Kristiyan Kostadinov fe343d8d96 refactor(compiler): clean up i18n attribute generation logic (#39498)
This is follow-up from [an earlier discussion](https://github.com/angular/angular/pull/39408#discussion_r511908358).
After some testing, it looks like the type of `Element.attributes` was correct in specifying that it
only has `TextAttribute` instances. This means that the extra checks that filter out `BoundAttribute`
instances from the array isn't necessary. There is another loop a bit further down that actually
extracts the bound i18n attributes.

PR Close #39498
2020-10-29 16:07:50 -07:00
JiaLiPassion 4a8d5ae970 fix(core): markDirty() should only mark flags when really scheduling tick. (#39316)
Close #39296

Fix an issue that `markDirty()` will not trigger change detection.

The case is for example we have the following component.

```
export class AppComponent implements OnInit {
  constructor(private router: Router) {}

  ngOnInit() {
    this.router.events
      .pipe(filter((e) => e instanceof NavigationEnd))
      .subscribe(() => ɵmarkDirty(this));
  }
}

export class CounterComponent implements OnInit, OnDestroy {
  ngOnInit() {
    this.countSubject.pipe(takeUntil(this.destroy)).subscribe((count) => {
      this.count = count;
      ɵmarkDirty(this);
    });
  }
```

Then the app navigate from `AppComponent` to `CounterComponent`,
so there are 2 `markDirty()` call at in a row.

The `1st` call is from `AppComponent` when router changed, the
`2nd` call is from `CounterComponent.ngOnInit()`.

And the `markDirty()->scheduleTick()` code look like this

```
function scheduleTick(rootContext, flags) {
    const nothingScheduled = rootContext.flags === 0 /* Empty */;
    rootContext.flags |= flags;
    if (nothingScheduled && rootContext.clean == _CLEAN_PROMISE) {
      rootContext.schedule(() => {
	...
        if (rootContext.flags & RootContextFlags.DetectChanges)
          rootContext.flags &= ~RootContextFlags.DetectChanges;
          tickContext();

        rootContext.clean = _CLEAN_PROMISE;
        ...
      });
```

So in this case, the `1st` markDirty() will
1. set rootContext.flags = 1
2. before `tickContext()`, reset rootContext.flags = 0
3. inside `tickContext()`, it will call `CounterComponent.ngOnint()`,
   so the `2nd` markDirty() is called.
4. and the `2nd` scheduleTick is called, `nothingScheduled` is true,
   but rootContext.clean is not `_CLEAN_PROMISE` yet, since the `1st` markDirty tick
   is still running.
5. So nowhere will reset the `rootContext.flags`.
6. then in the future, any other `markDirty()` call will not trigger the tick, since
   `nothingScheduled` is always false.

So `nothingScheduled` means no tick is scheduled, `rootContext.clean === _CLEAN_PROMISE`
means no tick is running.
So we should set the flags to `rootContext` only when `no tick is scheudled or running`.

PR Close #39316
2020-10-29 16:07:10 -07:00
JiaLiPassion 170af0740d release: cut the zone.js-0.11.3 release (#39440)
PR Close #39440
2020-10-29 15:40:29 -07:00
Igor Minar 17070af417 build: update to CLI 11.0.0-rc.1 (#39432)
This release fixed the previously found size regressions.

PR Close #39432
2020-10-29 13:47:12 -07:00
Igor Minar 904b213954 build: update to @angular/cli@11.0.0-rc.0 (#39432)
This updates just the cli packages, the material and cdk packages
will be updated separately.

PR Close #39432
2020-10-29 13:47:12 -07:00
ayazhafiz e3365724f2 feat(compiler): recover expression parsing in more malformed pipe cases (#39437)
This commit handles the following cases:
- incomplete pipes in a pipe chain
- incomplete arguments in a pipe chain
- incomplete arguments provided to a pipe
- nested pipes

The idea is to unconditionally recover on the presence of a pipe, which
should be okay because expression parsing can be independently between
pipes.

PR Close #39437
2020-10-29 12:38:40 -07:00
Andrew Kushnir 8d324ec314 ci: update NgBot config to include more labels that indicate that ticket is triaged (#39494)
This commit updates the config of NgBot to treat tickets that have "needs clarification" or
"needs reproduction" labels on them as triaged.

PR Close #39494
2020-10-29 12:25:47 -07:00
Bjarki f54662e931 fix(core): mark Trusted Types as declarations (#39471)
Angular-internal type definitions for Trusted Types were added in #39211.
When compiled using the Closure compiler with certain optimization
flags, identifiers from these type definitions (such as createPolicy)
are currently uglified and renamed to shorter strings. This causes
Angular applications compiled in this way to fail to create a Trusted
Types policy, and fall bock to using strings.

To fix this, mark the internal Trusted Types definitions as declarations
using the "declare" keyword. Also convert types to interfaces, for
the reasons explained in https://ncjamieson.com/prefer-interfaces/

PR Close #39471
2020-10-29 11:58:48 -07:00
Michael Ziluck 8af32fcb16 docs(forms): update ngModel documentation (#39481)
This commit improves the ngModel docs, specifically:
- clarifies purpose of the name attribute in ngModelOptions
- clarifies on the interaction with a parent form or lack thereof
- fix inconsistency with analogy for two-way binding
- cleans up some typos and extra wordiness
- clarifies language around common properties
- adds missing preposition to commit message format origins
PR Close #39481
2020-10-29 11:18:54 -07:00
Joey Perrott d6a765909b fix(dev-infra): update generated ng-dev compiled output (#39474)
Update to the new ng-dev generated script.

PR Close #39474
2020-10-29 09:55:26 -07:00
Joey Perrott 8106a7a11b fix(dev-infra): clean up output of `caretaker check` command (#39474)
Suppress the logging of the git command executed during the caretaker
check process.

PR Close #39474
2020-10-29 09:55:26 -07:00
Joey Perrott af8e547d86 feat(dev-infra): allow suppression of `GitClient`s verbose logging (#39474)
Some usages of the `GitClient` are better served by suppressing the
logging of lines that express what commands are being run. Many usages
of `GitClient` are contained within tools which are best served by
keeping the output clean as mostly read actions are occurring.

PR Close #39474
2020-10-29 09:55:26 -07:00
Joey Perrott 059b1ca322 build: create yarn command for running local version of ng-dev (#39474)
For better development experience of the dev-infra work, the `ng-dev:dev`
command runs the transpiled version of `ng-dev` making iterative
development easier.

PR Close #39474
2020-10-29 09:55:26 -07:00
Shadab Ansari 75bfc400cd docs: Match browser name with custom launcher name (#39480)
The browser being launched needs to match the custom launcher name.
Otherwise Karma would still trigger the original Chrome executable without the flags.

PR Close #39480
2020-10-29 08:39:14 -07:00
Alan Agius dd0ba3f4ab docs: rename `ng xi18n` to `ng extract-i18n` (#39337)
In Angular CLI version 11, xi18n has been changed from `ng xi18n` to `ng extract-i18n`.

PR Close #39337
2020-10-28 14:42:27 -07:00
Kapunahele Wong e3d2f46bbd docs: move template ref vars doc to concepts section clarify, add scope section, and update headers (#31195)
Fixes #31186. This commit adds more context about the behavior
of template reference variables in nested templates and moves
doc into concepts section.

PR Close #31195
2020-10-28 14:41:44 -07:00
David Shevitz 60e377ed77 docs: archive user-input topic (#39309)
PR Close #39309
2020-10-28 14:41:08 -07:00
Alex Rickabaugh bdec818b3a docs: release notes for the v11.0.0-rc.1 release 2020-10-28 14:20:29 -07:00
Alex Rickabaugh e42d3ecfcc docs: release notes for the v10.2.1 release 2020-10-28 13:54:14 -07:00
Jessica Janiuk 6d51ae2571 docs(core): add jessicajaniuk to pullapprove list (#39473)
This just adds jessicajessica to the pullapprove.yml file

PR Close #39473
2020-10-28 11:14:49 -07:00
Thibaut ACKERMANN 62a58e23b6 docs(router): Fix ActivationEnd link (#39469)
ActivationEnd goes to ActivationEnd and not to ActivationStart section

PR Close #39469
2020-10-28 11:06:04 -07:00
George Kalpakas 246d50e934 build(docs-infra): upgrade cli command docs sources to bf9a92299 (#39465)
Updating [angular#master](https://github.com/angular/angular/tree/master) from
[cli-builds#master](https://github.com/angular/cli-builds/tree/master).

##
Relevant changes in
[commit range](241058c1e...bf9a92299):

**Modified**
- help/build.json

PR Close #39465
2020-10-28 11:04:56 -07:00