Prior to this fix, incremental rebuilds could fail to type check due to
missing ambient types from auto-discovered declaration files in @types
directories, or type roots in general. This was caused by the
intermediary `ts.Program` that is created for template type checking,
for which a `ts.CompilerHost` was used which did not implement the
optional `directoryExists` methods. As a result, auto-discovery of types
would not be working correctly, and this would retain into the
`ts.Program` that would be created for an incremental rebuild.
This commit fixes the issue by forcing the custom `ts.CompilerHost` used
for type checking to properly delegate into the original
`ts.CompilerHost`, even for optional methods. This is accomplished using
a base class `DelegatingCompilerHost` which is typed in such a way that
newly introduced `ts.CompilerHost` methods must be accounted for.
Fixes#38979
PR Close#39011
We weren't resolving a path correctly which resulted in an error on Windows.
For reference, here's the error. Note the extra slash before `C:`:
```
Error: ENOENT: no such file or directory, scandir '/C:/bazel_output_root/yxvwd24o/external/npm/node_modules/typescript'
at Object.readdirSync (fs.js:854:3)
```
PR Close#39005
In the past, the docs examples were configured to not use AOT
compilation in development mode (only in production mode). This was an
artifact of when JIT was the default in development mode.
Now that AOT is the default (even in development mode) for new CLI apps,
this commit configures all docs examples to always use AOT compilation.
(This has been made possible by fixing the `component-interaction` docs
example to correctly run in AOT mode in an earlier commit.)
PR Close#39001
Previously, some of the e2e tests of the `component-interaction` docs
example were disabled because they were failing.
This commit fixes and re-enables them.
PR Close#39001
This commit refactors the e2e tests of the `component-interaction` docs
example to improve readability and make them easier to maintain.
Changes include:
- Switch from `element.all().get(0)` to `element()` when there is only
one such element on the page.
- Switch from `Promise#then()` to `async/await`.
- Move `ElementFinder`s at the top of the test (instead of having them
interleaved with expectations).
- Load the page before every test (i.e. in a `beforeEach()` instead of
`beforeAll()`) to prevent state from each test leaking into the
subsequent tests.
- Order imports alphabetically.
PR Close#39001
Previously, the `component-interaction` docs example was configured to
run e2e tests on CI in development mode (in contrast to the default for
all docs examples, which is to run e2e tests in production mode). This
was necessary due to the following reasons:
- One of the components, `CountdownTimerComponent`, which is used by
`CountdownLocalVarParentComponent` and
`CountdownViewChildParentComponent`, was triggering a periodic
asynchronous task (via `setInterval()`), which prevented the app from
stabilizing and caused tests to fail.
- In order to prevent this from happening, the example's `AppModule` had
special provisioning to not include the problematic components in its
declarations when testing.
- Since this had to be determined dynamically at runtime (via inspecting
the URL query params), the `AppModule`'s config could not be
statically evaluated in AOT compilation.
This commit fixes the example to make it compatible with AOT compilation
and removes the custom test command from its `example-config.json`
(allowing it to be run with the default e2e test command, i.e. in
production mode).
PR Close#39001
This commit updates the docs examples to Angular v10.1.3. In addition to
updating the dependencies versions, it also updates the project's
structure and config to more closely match what a new v10 CLI app would
look like. See, also, the [diff][1] between a basic v9.1.4 CLI app and a
v10.1.3 one.
[1]: https://github.com/cexbrayat/angular-cli-diff/compare/9.1.4..10.1.3
PR Close#38993
BlacklistedStackFrames to InternalZoneJsStackFrames along with other related
symbols renamed with the same changes (with appropriate casing style).
PR Close#38978
Previously, the `retry` example did not work as intended. The `retry`
operator was called before the exception occured, thus not retrying the
`ajax` request.
This commit moves the `retry` operator into the correct order to ensure
that the failed request is retried.
PR Close#38905
This commit updates the symbols in the TemplateTypeCheck API and methods
for retrieving them:
* Include `isComponent` and `selector` for directives so callers can determine which
attributes on an element map to the matched directives.
* Add a new `TextAttributeSymbol` and return this when requesting a symbol for a `TextAttribute`.
* When requesting a symbol for `PropertyWrite` and `MethodCall`, use the
`nameSpan` to retrieve symbols.
* Add fix to retrieve generic directives attached to elements/templates.
PR Close#38844
`NodeInjector` is store in expando as a list of values in an array. The
offset constant into the array have been brought together into a single
`NodeInjectorOffset` enum with better documentation explaining their usage.
PR Close#38707
This change makes `getPreviousOrParentTNode` return `TNode|null` (rather
than just `TNode`) which is more reflective of the reality. The
`getPreviousOrParentTNode` can be `null` upon entering the `LView`.
PR Close#38707
`TNodeType.View` was created to support inline views. That feature did
not materialize and we have since removed the instructions for it, leave
an unneeded `TNodeType.View` which was still used in a very
inconsistent way. This change no longer created `TNodeType.View` (and
there will be a follow up chang to completely remove it.)
Also simplified the mental model so that `LView[HOST]`/`LView[T_HOST]`
always point to the insertion location of the `LView`.
PR Close#38707
Host `TNode` was passed into `getOrCreateTNode` just so that we can
compute weather or not we are a root node. This was needed because
`previousOrParentTNode` could have `TNode` from `TView` other then
current `TView`. This is confusing mental model. Previous change
ensured that `previousOrParentTNode` must always be part of `TView`,
which enabled this change to remove the unneeded argument.
PR Close#38707
`previousOrParentTNode` stores current `TNode`. Due to inconsistent
implementation the value stored would sometimes belong to the current
`TView` and sometimes to the parent. We have extra logic which accounts
for it. A better solution is to just ensure that `previousOrParentTNode`
always belongs to current `TNode`. This simplifies the mental model
and cleans up some code.
PR Close#38707
Creates a tool for staging and publishing releases as per the
new branching and versioning that has been outlined in the following
document. The tool is intended to be used across the organization to
ensure consistent branching/versioning and labeling:
https://docs.google.com/document/d/197kVillDwx-RZtSVOBtPb4BBIAw0E9RT3q3v6DZkykU/edit#heading=h.s3qlps8f4zq7dd
The tool implements the actions as outlined in the following
initial plan: https://hackmd.io/2Le8leq0S6G_R5VEVTNK9A.
The implementation slightly diverged in so far that it performs
staging and publishing together so that releasing is a single
convenient command. In case of errors for which re-running the
full command is not sufficient, we want to consider adding
recover functionality. e.g. when the staging completed, but the
actual NPM publishing aborted unexpectedly due to build errors.
PR Close#38656
Introduces a new command for `ng-dev release`, so that the NPM
dist tag can be set for all configured NPM packages. This command
can be useful in case a manual tag needs to be set, but it is
primarily used by the release tooling when a new stable version
is cut, and when the previous patch branch needs to be set as LTS
version through a `v{major}-lts` dist tag.
It is necessary to have this as a command so that the release tool
can execute it for old branches where other packages might have been
configured. This is similar to the separate `ng-dev build` command
that we created.
Note that we also added logic for spawning a process conveniently
with different "console output" modes. This will be useful for
other command invocations in the release tool and it's generally
better than directly using native `child_process` as that one doesn't
log to the dev-infra debug log file.
PR Close#38656
Adds a command for building all release packages. This command
is primarily used by the release tool for building release output
in version branches. The release tool cannot build the release packages
configured in `master` as those packages could differ from the
packages available in a given version branch. Also, the build process
could have changed, so we want to have an API for building
release packages that is guaranteed to be consistent across branches.
PR Close#38656
Moves the existing `ng-dev release env-stamp` command into a
subfolder so that the staging/publish tool can have its own
dedicated folder (without being polluted by the env-stamp logic).
Every subcommand should be in its own folder.
PR Close#38656
Adds a new folder to dev-infra where shared testing utilities
could be placed in. This commit already adds initial testing
utilities for dealing with the `GitClient` and SemVer versions.
The `GitClient` in the testing utilities simulates actual Git
behavior in a virtual manner. It's not complete at all, but can
be extended based on our needs. The currently implemented commands
are the most basic ones that we'd need for our release tooling.
PR Close#38656
Adds a method for printing active release trains for a configured
project. This is helpful for the release tool that will print
the active release trains. Also this can be useful for the
caretaker status command, where we could print the active
version branches (i.e. "is there currently a feature-freeze branch").
PR Close#38656
Adds logic for determining active LTS branches for a given
release configuration. The active LTS branches can be determined
by querying NPM and matching dist tags against a specific
pattern. i.e. `v{major}-lts`.
This logic will be useful for the release tool that supports
publishing of active LTS version branches.
PR Close#38656
Cleans up outdated comments in the shared dev-infra Git
utilities. We also export the Graphql client for consistency
as we expose the `GithubClient` and `GitClient` too.
PR Close#38656
We initially added logic for determining active release trains into
the merge script. Given we now build more tools that rely on this
information, we move the logic into a more general "versioning" folder
that can contain common logic following the versioning document for the
Angular organization.
PR Close#38656
Introduces a new configuration for the `ng-dev release` command. This
configuration will be the source of truth for all release packages
and how they can be built.
Additionally, in a temporary manner where each project has its own
way of generating the changelog, the changelog generation can be
configured. This will be removed in the future when there is
canonical changelog generation in the dev-infra shared package.
PR Close#38656
Exposes logic for dealing with LTS branches, so that the release
tool can re-use it for cutting LTS patch releases.
Eventually, we can move all of this logic to a more dedicated
folder instead of having it inside the merge folder.
PR Close#38656
Instead of maintaining multiple interface for grouping
owner name and repo name, we expose a shared interface
describing a Github repository.
One unfortunate downside is that the GraphQL Github
and Rest API diverge slightly with the key for the
repository name. i.e. rest uses `repo` for the name
of a repository, while GraphQL uses `name` for the name.
If that would be consistent, we could use the rest operator
to pass a repository to the Octokit REST or GraphQL API. This
does not work, so we have a small manual overhead as seen
in the `branches.ts` file.
PR Close#38656
The dev-infra package is currently built with Bazel and ts-node.
In Bazel, the shared tsconfig from the `packages/` folder is used.
This means that the code is built in strict mode, but IDEs and
ts-node do not know about the strictness. This is because the tsconfig
is part of the `packages` folder and not accessible from the
dev-infra package. We fix this by adding an IDE and ts-node specific
tsconfig to the dev-infra package.
This helps with spotting compilation failures before building
with Bazel / waiting for CI to check build state.
PR Close#38656
Previously, the logic for determing the active release trains did not
return the resolved version of a release train. With the publish script
being created, we need this information and can just pass it through,
so that we do not need to fetch and parse the package.json of given
branches multiple times.
PR Close#38656
Sets up the NPM `ora` package in the project and in dev-infra,
so that we can show progress spinners when needed. This is useful
in the publish release script when we wait for a pull request to
be merged.
PR Close#38656
The git client respects the `SpawnSyncOptions` when a command
is executed. Currently it does not hide the command info
messages when commands are run in silent mode.
We fix this as part of this commit, so that the command info
is only printed to `debug` if `stdio` is set to `ignore`.
Additonally, the github token is made public so that it can be
used by commands if other repositories like forks are targeted.
PR Close#38656
Instead of repeating the logic for adding the github token to
a repository git url, we add a shared function for automatically
computing the URls with token.
Additionally, URLs for updating/generating tokens have been moved
to a dedicated file in the `utils` folder. Also while being at it,
the yargs github token helper is also moved into the dedicated
Git/Github related util folder.
PR Close#38656
Even in the overloads, state that it can accept `null` and
`undefined`, in order to ensure easy composition with `async`.
Additionally, change the implementation to return `null` on an
`undefined` input, for consistency with other pipes.
BREAKING CHANGE:
The `slice` pipe now returns `null` for the `undefined` input value,
which is consistent with the behavior of most pipes. If you rely on
`undefined` being the result in that case, you now need to check for it
explicitly.
PR Close#37447
As shown in the tests, `KeyValuePipe.transform` can accept
`undefined`, in which case it always returns `null`.
Additionally, the typing for `string` keys can be made generic, so the
comparison function is only required to accept the relevant cases.
Finally, the typing for `number` records now shows that the comparison
function and the result entries will actually receive the string version
of the numeric keys, just as shown in the tests.
BREAKING CHANGE:
The typing of the `keyvalue` pipe has been fixed to report that for
input objects that have `number` keys, the result will contain the
string representation of the keys. This was already the case and the
code has simply been updated to reflect this. Please update the
consumers of the pipe output if they were relying on the incorrect
types. Note that this does not affect use cases where the input values
are `Map`s, so if you need to preserve `number`s, this is an effective
way.
PR Close#37447
I18nPluralPipe can actually accept `null` and `undefined` (which are
convenient for composing it with the async pipe), but it is currently
typed to only accept `number`.
PR Close#37447
Make typing of number pipes stricter to catch some misuses (such as
passing an Observable or an array) at compile time.
BREAKING CHANGE:
The signatures of the number pipes now explicitly state which types are
accepted. This should only cause issues in corner cases, as any other
values would result in runtime exceptions.
PR Close#37447
Make typing of DatePipe stricter to catch some misuses (such as passing
an Observable or an array) at compile time.
BREAKING CHANGE:
The signature of the `date` pipe now explicitly states which types are
accepted. This should only cause issues in corner cases, as any other
values would result in runtime exceptions.
PR Close#37447
`AsyncPipe.transform` will never return `undefined`, even when passed
`undefined` in input, in contrast with what was declared in the
overloads.
Additionally the "actual" method signature can be updated to match the
most generic case, since the implementation does not rely on wrappers
anymore.
BREAKING CHANGE:
The async pipe no longer claims to return `undefined` for an input that
was typed as `undefined`. Note that the code actually returned `null` on
`undefined` inputs. In the unlikely case you were relying on this,
please fix the typing of the consumers of the pipe output.
PR Close#37447
The old implementation of case conversion types can handle several
values which are not strings, but the signature did not reflect this.
The new one reports errors when falsy non-string inputs are given to
the pipe (such as `false` or `0`) and has a new signature which
instead reflects the behaviour on `null` and `undefined`.
Fixes#36259
BREAKING CHANGE:
The case conversion pipes no longer let falsy values through. They now
map both `null` and `undefined` to `null` and raise an exception on
invalid input (`0`, `false`, `NaN`) just like most "common pipes". If
your code required falsy values to pass through, you need to handle them
explicitly.
PR Close#37447
For the following example, the cursor is between `keySpan` and `valueSpan`
of the `BoundAttribute`.
```html
<test-cmp [foo]¦="bar"></test-cmp>
```
Our hybrid visitor will return `Element`in this case, which is the parent
node of the `BoundAttribute`.
This is because we only look at the `keySpan` and `valueSpan`, and not
the source span. The last element in the AST path is `Element`, so it gets
returned.
In this PR, I propose fixing this by adding a sentinel value `undefined`
to the AST path to signal that we've found a source span but the cursor is
neither in the key span nor the value span.
PR Close#38995
JoostK
Kristiyan Kostadinov
George Kalpakas
Joey Perrott
Sonu Kapoor
Andrew Scott
Misko Hevery
Paul Gschwendtner
Andrea Canciani
Keen Yee Liau