This commit adds ngDevMode guard to show warnings only
in dev mode (similar to how things work in other parts of Ivy runtime code).
The ngDevMode flag helps to tree-shake these warnings from production builds
(in dev mode everything will work as it works right now) to decrease production bundle size.
PR Close#39964
This commit adds `ngDevMode` guard to run `checkNoChanges` only
in dev mode (similar to how things work in other parts of Ivy runtime code).
The `ngDevMode` flag helps to tree-shake this code from production builds
(in dev mode everything will work as it works right now) to decrease production bundle size.
PR Close#39964
When the compiler is invoked via ngc or the Angular CLI, its APIs are used
under the assumption that Angular analysis/diagnostics are only requested if
the program has no TypeScript-level errors. A result of this assumption is
that the incremental engine has not needed to resolve changes via its
dependency graph when the program contained broken imports, since broken
imports are a TypeScript error.
The Angular Language Service for Ivy is using the compiler as a backend, and
exercising its incremental compilation APIs without enforcing this
assumption. As a result, the Language Service has run into issues where
broken imports cause incremental compilation to fail and produce incorrect
results.
This commit introduces a mechanism within the compiler to keep track of
files for which dependency analysis has failed, and to always treat such
files as potentially affected by future incremental steps.
PR Close#39967
To avoid overwhelming a user with secondary diagnostics that derive from a
"root cause" error, the compiler has the notion of a "poisoned" NgModule.
An NgModule becomes poisoned when its declaration contains semantic errors:
declarations which are not components or pipes, imports which are not other
NgModules, etc. An NgModule also becomes poisoned if it imports or exports
another poisoned NgModule.
Previously, the compiler tracked this poisoned status as an alternate state
for each scope. Either a correct scope could be produced, or the entire
scope would be set to a sentinel error value. This meant that the compiler
would not track any information about a scope that was determined to be in
error.
This method presents several issues:
1. The compiler is unable to support the language service and return results
when a component or its module scope is poisoned.
This is fine for compilation, since diagnostics will be produced showing the
error(s), but the language service needs to still work for incorrect code.
2. `getComponentScopes()` does not return components with a poisoned scope,
which interferes with resource tracking of incremental builds.
If the component isn't included in that list, then the NgModule for it will
not have its dependencies properly tracked, and this can cause future
incremental build steps to produce incorrect results.
This commit changes the tracking of poisoned module scopes to use a flag on
the scope itself, rather than a sentinel value that replaces the scope. This
means that the scope itself will still be tracked, even if it contains
semantic errors.
PR Close#39967
Previously, if a trait's analysis step resulted in diagnostics, the trait
would be considered "errored" and no further operations, including register,
would be performed. Effectively, this meant that the compiler would pretend
the class in question was actually undecorated.
However, this behavior is problematic for several reasons:
1. It leads to inaccurate diagnostics being reported downstream.
For example, if a component is put into the error state, for example due to
a template error, the NgModule which declares the component would produce a
diagnostic claiming that the declaration is neither a directive nor a pipe.
This happened because the compiler wouldn't register() the component trait,
so the component would not be recorded as actually being a directive.
2. It can cause incorrect behavior on incremental builds.
This bug is more complex, but the general issue is that if the compiler
fails to associate a component and its module, then incremental builds will
not correctly re-analyze the module when the component's template changes.
Failing to register the component as such is one link in the larger chain of
issues that result in these kinds of issues.
3. It lumps together diagnostics produced during analysis and resolve steps.
This is not causing issues currently as the dependency graph ensures the
right classes are re-analyzed when needed, instead of showing stale
diagnostics. However, the dependency graph was not intended to serve this
role, and could potentially be optimized in ways that would break this
functionality.
This commit removes the concept of an "errored" trait entirely from the
trait system. Instead, analyzed and resolved traits have corresponding (and
separate) diagnostics, in addition to potentially `null` analysis results.
Analysis (but not resolution) diagnostics are carried forward during
incremental build operations. Compilation (emit) is only performed when
a trait reaches the resolved state with no diagnostics.
This change is functionally different than before as the `register` step is
now performed even in the presence of analysis errors, as long as analysis
results are also produced. This fixes problem 1 above, and is part of the
larger solution to problem 2.
PR Close#39967
This commit adds `ngDevMode` guard to show sanitization warnings only
in dev mode (similar to how things work in other parts of Ivy runtime code).
The `ngDevMode` flag helps to tree-shake these warnings from production builds
(in dev mode everything will work as it works right now) to decrease production bundle size.
PR Close#39959
This expands the deprecation message that started to pop up in v11.0.3
after the landing of commit e148382bd0,
that deprecated the `{[key: string]: any}` type for the options property of the `FormBuilder.group` method.
It turns out that having a custom validator declared as
`{ validators: (group: FormGroup) => ValidationErrors|null }` works in practice,
but is now inferred by TS as the deprecated version of `group`
(because `FormGroup` is a subclass of `AbstractControl` that `ValidatorFn` expects).
We considered the possibility of tweaking the forms API to accept such validators,
but it turns out to generate too many changes in the framework or possible breaking changes for Angular users.
We settled for a more explicit deprecation message, elaborated with the help of @petebacondarwin.
This will hopefully help developers to understand why the deprecation warning is showing up
when they think they are already using the non-deprecated overload.
PR Close#39946
In the past, the legacy (VE-based) language service would use a
`UrlResolver` instance to resolve file paths, primarily for compiler
resources like external templates. The problem with this is that the
UrlResolver is designed to resolve URLs in general, and so for a path
like `/a/b/#c`, `#c` is treated as hash/fragment rather than as part
of the path, which can lead to unexpected path resolution (f.x.,
`resolve('a/b/#c/d.ts', './d.html')` would produce `'a/b/d.html'` rather
than the expected `'a/b/#c/d.html'`).
This commit resolves the issue by using Node's `path` module to resolve
file paths directly, which aligns more with how resources are resolved
in the Ivy compiler.
The testing story here is not great, and the API for validating a file
path could be a little bit prettier/robust. However, since the VE-based
language service is going into more of a "maintenance mode" now that
there is a clear path for the Ivy-based LS moving forward, I think it is
okay not to spend too much time here.
Closes https://github.com/angular/vscode-ng-language-service/issues/892
Closes https://github.com/angular/vscode-ng-language-service/issues/1001
PR Close#39917
At the moment, when creating a root module, a subscription to the
`onError` subject is also created. It captures the scope where `NgModuleRef`
is created and prevents it from being garbage collected. Also note that this
`NgModuleRef` has a reference to the root module instance (e.g. `AppModule`),
which also prevents it from being GC'd.
PR Close#39940
Since #39853, it is possible to deploy to multiple Firebase sites from a
single branch. In order to deploy to a site, we need to associate an
alias (`aio`) with a site. This is done via the `firebase target:apply`
command. However, when the command is called multiple times, it
associates the alias with many sites, which subsequently fails during
deployment ([example failure][1]), since the `firebase deploy` command
does not know what site to deploy to.
This commit fixes the deployment script by ensuring that any previous
association with the `aio` alias is cleared (via the
`firebase target:clear` command) before associating it with a new site.
[1]: https://circleci.com/gh/angular/angular/871020
PR Close#39948
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
This commit adds a few tests to verify that the `onDestroy` callbacks are invoked when `ComponentRef` instance
is destroyed and the logic is consistent between ViewEngine and Ivy.
PR Close#39876
Fix a case where, if the parent class had already been patched, it would
not patch the child class. In addition to checking if the method is
defined in the prototype, and not inherited, it also does the same for
the unpatched method.
PR Close#39850
Previously, test descriptions used `latest` to refer to the most
recent/highest version. This was ambiguous, because `latest` can also
refer to the stable version of a package (e.g. see `@latest` npm tag).
This commit replaces `latest` with `highest` (or `highest for major`) to
reduce ambiguity.
Discussed in:
https://github.com/angular/angular/pull/39853#discussion_r531730317
PR Close#39853
Currently there is an issue with redirecting `rc.angular.io` to
`angular.io` when there is no active RC. If a user has visited
`rc.angular.io` before and has a ServiceWorker registered for that
subdomain, they will never "see" the redirect to `angular.io`.
This commit fixes the problem by doing an additional deployment from the
stable branch to the `rc-angular-io-site` Firebase site when there is no
active RC. This additional deployment will ensure that:
1. Users will be temporarily redirected from `rc.angular.io` to
`angular.io`.
2. Users with a registered ServiceWorker (who don't see the redirect)
will have their ServiceWorker unregistered on the next visit.
3. The content on both sites is identical.
See #39760 for more details on the problem and the solution.
NOTE:
As mentioned in #39760, for this fix to take affect, we need to remove
the redirect from `rc.angular.io` to `angular.io` in the Firebase
console for site `rc-angular-io-site`.
Fixes#39760
PR Close#39853
Previously, a branch would be deployed as `archive` even if it had a
major version that was equal/higher than that of the stable branch (as
long as it was not the RC branch - i.e. not the most recent minor
branch). For example, with `11.0.x` as the stable branch and `12.0.x`
as the RC branch, `11.1.x` would be deployed as archive.
Theoretically, we should never find ourselves in such a situation.
Typically, there will only be at most one minor branch most recent than
the stable one (and that branch will be the RC branch). However, it
is possible under unusual circumstances.
This commit adds additional checks to guard against this problem. It
also refactors the code in preparation of fixing an issue with
`rc.angular.io` redirects in the presence of a ServiceWorker, which will
require identifying whether there is an active RC version or not.
See #39760 for more details.
PR Close#39853
Previously, the latest commit for branch may be computed multiple times
in the `deploy-to-firebase.js` tests.
This commit avoids the unnecessary re-computations by computing the
latest commits for the necessary branches at the beginning and using the
computed values throughout the tests.
PR Close#39853
Previously, the `deploy-to-firebase.js` script would only perform one
deployment operation on each run.
This commit adds support for performing multiple deployment operations.
NOTE:
In a subsequent commit, this will be leveraged fix an issue with
`rc.angular.io` redirects in the presence of a ServiceWorker by
deploying the same artifacts to multiple Firebase projects/sites.
See #39760 for more details.
PR Close#39853
Previously, the `deploy()` function in `deploy-to-firebase.js` would
also perform other operations (beyond deploying), such as building the
app, checking the generated payload size, testing the PWA score of the
deployed app.
This commit decouples these operations, so that deploying can be
performed independently.
NOTE:
In a subsequent commit, this will be leveraged fix an issue with
`rc.angular.io` redirects in the presence of a ServiceWorker by
deploying the same artifacts to multiple Firebase projects/sites.
See #39760 for more details.
PR Close#39853
Since we turned off the Selenium Promise Manager in #39600, the AIO e2e
tests have started flaking on CI. After trying out several things, the
only change that seems to eliminate the flakiness is turning the
Selenium Promise Manager back on (see #39873 for more details).
This commit turns the Selenium Project Manager on to get rid of the
flakiness.
Fixes#39872
PR Close#39905
In #32497, an option was introduced to retry failed docs examples tests
a second time. This was done to work around some then recently
introduced flakiness. After inspecting ~50 recent CI docs examples jobs,
all tests passed on the first try.
This commit gets rid of the retry attempt. This will avoid covering up
any new flakiness introduces in the future.
PR Close#39905
This commit edits the copy on the interpolation page.
Tightens language, clarifies headers, and streamlines text.
There are no content or code changes.
PR Close#38687
Currently we convert objects to strings using `'' + value` which is quickest,
but it stringifies the value using its `valueOf`, rather than `toString`. These
changes switch to using `String(value)` which has identical performance
and calls the `toString` method as expected. Note that another option
was calling `toString` directly, but benchmarking showed it to be slower.
I've included the benchmark I used to verify the performance so we have it
for future reference and we can reuse it when making changes to `renderStringify`
in the future.
Also for reference, here are the results of the benchmark:
```
Benchmark: renderStringify
concat: 2.006 ns(0%)
concat with toString: 2.201 ns(-10%)
toString: 237.494 ns(-11741%)
toString with toString: 121.072 ns(-5937%)
constructor: 2.201 ns(-10%)
constructor with toString: 2.201 ns(-10%)
toString mono: 14.536 ns(-625%)
toString with toString mono: 9.757 ns(-386%)
```
Fixes#38839.
PR Close#39843
We currently only wrap the event listener in the function which ensures
ancestors are marked for check when the listener is placed on an element
that has a native method for listening to an event. We actually need to do
this wrapping in all cases so that events that are attached to non-rendered
template items (`ng-template` and `ng-container`) also mark ancestors for check
when they receive the event.
fixes#39832
PR Close#39833
DEPRECATION:
Mark the {[key: string]: any} type for the options property of the FormBuilder.group method as deprecated.
Using AbstractControlOptions gives the same functionality and is type-safe.
PR Close#39769
This adds angulararchitects.io to the tranings section in the
resources.json file for angular.io. It also adds a description
for angulararchitects.io.
PR Close#37139
The article previously referred was out of date. Most of the article content
was already duplicated here, or described how to generate multiple bundles. But
creating multiple bundles manually is now obsolete and the parameter `--localize`
should be preferred instead, because it is much faster.
This commit integrate the only interesting part left, relating to webserver configuration.
Co-authored-by: David Shevitz <dshevitz@google.com>
PR Close#36553
arturovt
