Improves stability of the Windows Bazel CI job by
installing Bazelisk globally.
Also makes the environment helpers more convenient by
evaluating the variable assignments directly, simplifying
some Bash logic.
PR Close#45431
Dedupes the Yarn run steps, avoiding the need to manually keep
this step in sync (e.g. with the timeout -- which is currently missing
for the windows job)
PR Close#45431
Recent changes in `rules_nodejs` caused the test case copy file actions
to be transitioned into the `exec` configuration, resulting in much larger
file paths. These paths break on Windows with the shell argument limit, and
with the path limit, causing errors like:
```
ERROR: C:/users/circleci/ng/packages/compiler-cli/test/compliance/test_cases/BUILD.bazel:9:12: Copying file packages/compiler-cli/test/compliance/test_cases/r3_compiler_compliance/components_and_directives/value_composition/structural_directives_if_directive_def.js failed: (Exit 1): cmd.exe failed: error executing command
cd /d C:/users/circleci/_bazel_circleci/u4uoan2j/execroot/angular
SET PATH=C:\Program Files\Git\usr\bin;C:\Program Files\Git\bin;C:\Windows;C:\Windows\System32;C:\Windows\System32\WindowsPowerShell\v1.0
SET RUNFILES_MANIFEST_ONLY=1
cmd.exe /C bazel-out\x64_windows-opt-exec-2B5CBBC6\bin\packages\compiler-cli\test\compliance\test_cases\test_cases--1973427149-cmd.bat
The system cannot find the path specified
```
https://app.circleci.com/pipelines/github/angular/angular/44038/workflows/4b530cb2-f232-4e1d-b35a-e6e085151d08/jobs/1140017
PR Close#45431
It is totally fine, and expected to use Git Bash for running Bazel
on Windows. In fact this is the most common setup for Bazel on Windows
and it's unrealistic to run without it.
This allows us to remove the old/legacy Powershell setup from CI
which is also quite flaky sometimes and does not reproduce how
Bazel is used on windows-users dev machines.
PR Close#45431
The Angular components repository can only start using Bazel Rules NodeJS v5
when `@angular/bazel` is published with support for it. To work around this
cycle we temporarily disable the unit tests job until we migrated the COMP
repo as well.
PR Close#45431
We updated buildifier and a few warnings became errors now. This commit
cleans up the failing unused variable instances, making the linter happy.
Additionally for the API extractor BUILD file, the package defaults
need to move to satisfy buildifier.
PR Close#45431
As mentioned in previous commits (check them for more details), `@bazel/typescript`
no longer contains `ts_library`-specific code, so we no longer need that dependency.
PR Close#45431
`@bazel/typescript` code moved to `@bazel/concatjs` for the tsc-wrapped code.
Note that this code is likely going to be removed anyway soon when we
move dts bundling from `ng_module` to `ng_package`.
PR Close#45431
Previously `tsc-wrapped` which is the foundation for `ngc-wrapped`, resided
in `@bazel/typescript`. It has been moved to `@bazel/concatjs` in rules_nodejs
so we need to account for that as part of our rules_nodejs v5 update.
PR Close#45431
Update `@bazel` packages to the latest 5.x version.
Some of the changes here are modeled after
angular/dev-infra@40c0ac8559.
Co-Authored-By: George Kalpakas <kalpakas.g@gmail.com>
PR Close#45431
We must read `Symbol.species` safely because `this` may be anything. For instance, `this`
may be an object without a prototype (created through `Object.create(null)`); thus
`this.constructor` will be undefined. One of the use cases is SystemJS creating
prototype-less objects (modules) via `Object.create(null)`. The SystemJS creates an empty
object and copies promise properties into that object (within the `getOrCreateLoad`
function). The zone.js then checks if the resolved value has the `then` method and invokes
it with the `value` context. Otherwise, this will throw an error: `TypeError: Cannot read
properties of undefined (reading 'Symbol(Symbol.species)')`.
PR Close#45369
.substr() is deprecated so we replace it with functions which work similarily but aren't deprecated
Signed-off-by: Tobias Speicher <rootcommander@gmail.com>
PR Close#45397
When we have an event listener inside an embedded view, we generate a `restoreView` call which saves the view inside of the LFrame. The problem is that we don't clear it until it gets overwritten which can lead to memory leaks.
These changes rework the generated code in order to generate a `resetView` call which will clear the view from the LFrame.
Fixes#42848.
PR Close#43075
This commit removes a bunch of methodss from `ReflectionCapabilities` as they
have gone unused. This also removes `Reflector` as it doesn't serve any purpose
and it is not exposed as public API, so can be safely removed.
PR Close#45335
Update the `Dockerfile` used to create the preview server to use the
latest stable version of Debian (`bullseye`) and also update package
versions to latest versions.
Also, unpin the versions of installed packages (except for Node.js
related ones) as pinning proved problematic due to many packages
removing old versions from the official repositories.
NOTE:
This change will allow the preview server to be updated on the VM and
take advantage of recent fixes, such as #45349. Currently, the update
fails with the error:
```
E: Version '7.64.0-4+deb10u1' for 'curl' was not found
The command '/bin/sh -c apt-get update -y && apt-get install -y curl=7.64.0-4+deb10u1' returned a non-zero code: 100
```
PR Close#45390
Before this change, the macrotask for `setInterval(callback, ms)` was no
longer tracked by `TaskTrackingZoneSpec` after the `callback` was
invoked for the first time. Now the periodic macrotask is tracked until
it is cancelled, e.g. `clearInterval(id)`.
BREAKING CHANGE: in TaskTrackingZoneSpec track a periodic task until it is cancelled
The breaking change is scoped only to the plugin
`zone.js/plugins/task-tracking`. If you used `TaskTrackingZoneSpec` and
checked the pending macroTasks e.g. using `(this.ngZone as any)._inner
._parent._properties.TaskTrackingZone.getTasksFor('macroTask')`, then
its behavior slightly changed for periodic macrotasks. For example,
previously the `setInterval` macrotask was no longer tracked after its
callback was executed for the first time. Now it's tracked until
the task is explicitly cancelled, e.g with `clearInterval(id)`.
fixes 45350
PR Close#45391
Jasmine checks internally if `process` and `process.on` is defined. Otherwise,
it installs the browser rejection handler through the `global.addEventListener`.
This code may be run in the browser environment where `process` is not defined, and
this will lead to a runtime exception since Webpack 5 removed automatic Node.js polyfills.
PR Close#42260
PR Close#45392
The Bazel NodeJS rules provide two ways of accessing node modules:
* A linker which creates a `node_modules` directory in the execroot/or in the runfiles.
* A patched module resolution where no node modules directory necessarily needs to exist.
The first is the default in `rules_nodejs` and the second is technically the most idiomatic
resolution mechanism in Bazel (as it matches with a runfile resolution library).
The linker is prone to race conditions in persistent workers, or non-sandbox environments (like
windows). This is because the linker for all workers will operate on a shared `execroot` directory
and the same `node_modules` directory is modified all the time / potentially conflicting with other
linker processes from other concurrently-running workers.
We rely on the patched module resolution anyway, but just need to disable the unused linker to avoid
issues like the following:
```
---8<---8<--- Start of log, file at /private/var/tmp/_bazel_splaktar/280f06d55552a0d01f89f0955b5acd78/bazel-workers/worker-8-TypeScriptCompile.log ---8<---8<---
[link_node_modules.js] An error has been reported: [Error: ENOENT: no such file or directory, unlink 'node_modules'] {
errno: -2,
code: 'ENOENT',
syscall: 'unlink',
path: 'node_modules'
} Error: ENOENT: no such file or directory, unlink 'node_modules'
---8<---8<--- End of log ---8<---8<---
INFO: Elapsed time: 12.796s, Critical Path: 5.39s
INFO: 645 processes: 477 internal, 12 darwin-sandbox, 156 worker.
```
PR Close#45393
Drops support for TypeScript older than 4.6 and removes some workarounds in the compiler.
BREAKING CHANGE:
TypeScript versions older than 4.6 are no longer supported.
PR Close#45394
There was a subtle bug involving the opt-out class for FormBuilder, which I discovered during the ongoing migration. The types must be structurally the same, because people pass around FormBuilders, in addition to passing around the controls they produce. This PR ensures FormBuilder and UntypedFormBuilder are assignable to each other.
PR Close#45421
In early versions of Angular, it was sometimes necessary to provide a
`moduleId` to `@Component` metadata, and the common pattern for doing this
was to set `moduleId: module.id`. This relied on the bundler to fill in a
value for `module.id`.
However, due to the superficial similarity between `Component.moduleId` and
`NgModule.id`, many users ended up setting `id: module.id` in their
NgModules. This is an anti-pattern that has a few negative effects,
including preventing the NgModule from tree-shaking properly.
This commit changes the compiler to ignore `id: module.id` in NgModules, and
instead provide a warning which suggests removing the line entirely.
PR Close#45024
Previously, the `TraitCompiler` would naively consider a compilation as
failed if either analysis or resolution produced any diagnostics. This
commit adjusts the logic to only consider error diagnostics, which allows
warnings to be produced from `DecoratorHandler`s.
This is a precursor commit to introducing such a warning. As such, the
logic here will be tested in the next commit.
PR Close#45024
Angular contains an NgModule registry, which allows a user to declare
NgModules with string ids and retrieve them via those ids, using the
`getNgModuleById` API.
Previously, we attempted to structure this registration in a clever fashion
to allow for tree-shaking of registered NgModules (that is, those with ids).
This sort of worked due to the accidental alignment of behaviors from the
different tree-shakers involved. However, this trick relies on the
generation of `.ngfactory` files and how they're specifically processed in
various bundling scenarios. We intend to remove `.ngfactory` files, hence
we can no longer rely on them in this way.
The correct solution here is to recognize that `@NgModule({id})` is
inherently declaring a global side-effect, and such classes should not
really be eligible for tree-shaking in the first place. This commit removes
all the old registration machinery, and standardizes on generating a side-
effectful call to `registerNgModuleType` for NgModules that have ids.
There is some risk here that NgModules with unnecessary `id`s may not
tree-shake as a result of this change, whereas they would have in previous
circumstances. The fix here should be to remove the `id` if it's not needed.
Specifying an `id` is a request that the NgModule be retained regardless of
any other references, in case it is later looked up by string id.
PR Close#45024
When `@angular/compiler` processes metadata and compiles a definition field,
it might also choose to return statements that are associated with that
definition, and should be included after the type being compiled. Currently,
the linker ignores these statements, as there are none generated that are
relevant in the linking operation.
A challenge to supporting such associated statements is that the linker
operates on "declare" expressions, and replaces those expressions with other
expressions. It does not have the capability to append statements after the
whole type. The linker actually faces this challenge with statements from
the `ConstantPool` as well, and solves this problem by generating an IIFE
expression that executes the statements and then returns the definition
expression.
Previously, an `EmitScope` processed the definition and converted it to an
expression, as well as collected constant statements from a `ConstantPool`.
A special `IifeEmitScope` implementation was used when emitting into a
context where top-level constant statements couldn't be added at all, and
uses the IIFE strategy in this case.
This commit adds blanket support for associated statements to the linker
using this IIFE strategy. The main `EmitScope` now uses the IIFE strategy to
emit associated statements, and `IifeEmitScope` has been renamed to
`LocalEmitScope`. Now, the `LocalEmitScope` represents constant statements
as associated statements to the main `EmitScope` implementation, so they
get included in the IIFE as well.
Tests are adjusted/added to cover this new behavior. This is a refactoring
commit because no live generated code is affected - there are no cases where
associated statements are present in linked definitions today.
PR Close#45024
The `compileNgModule` operation previously supported a flag `emitInline`,
which controlled whether template scoping information for the NgModule was
emitted directly into the compiled NgModule definition, or whether an
associated statement was generated which patched the information onto the
NgModule definition. Both options are useful in different contexts.
This commit changes this flag to an enum (and renames it), which allows for
a third option - do not emit any template scoping information. This option
is added to better represent the actual behavior of the Angular Linker,
which sometimes configures `compileNgModule` to use the side-effectful
statement generation but which does not actually emit such associated
statements. In other words, the linker effectively does not generate
scoping information for NgModules at all (in some configurations) and this
option more directly expresses that behavior.
This is a refactoring as no generated code is changed as a result of
introducing this flag, due to the linker's behavior of not emitting
associated statements.
PR Close#45024
Previously, the migration only migrated constructor calls. Now, the migration will rewrite every usage, in all contexts. Both ways are technically correct, but migrating all symbols is likely to produce clearer and more readable results.
PR Close#45311
warn developers when they are trying to animate non-animatable CSS
properties so that can more easily understand why something is not being
animated as they would expect it to
resolves#27577
PR Close#45212
Move to the CircleCI v2 api as the authentication fails for downloading artifacts using the v1 methods.
CircleCI v2 api now requires authentication to occur view the headers instead of being done in a
query parameter, all of the CircleCI interactions are now performed through one fetchFromCircleCi method
which ensures the token is provided in the headers as expected.
PR Close#45349
The animations package supports adding default parameter values to an animation that will be used as a fallback if some parameters aren't defined. The problem is that they're applied using a spread expression which means that any own property of the animation parameters will override the defaults, even if it resolves to null or undefined. This can lead to obscure errors like "Cannot read property toString of undefined" for an animation that looks like `{params: {foo: undefined}}` with defaults `{foo: 123}`.
I ran into this issue while debugging some test failures on Material.
These changes address the issue by:
1. Applying the defaults if the resolved value is null or undefined.
2. Updating the validation function to use a null check instead of `hasOwnProperty`.
PR Close#45339
