We recently added a sanity check that ensures that `python` is
available for Bazel. The check relies on `/usr/bin/python` to check
if python is available w/ the correct version. The problem is
that on Windows there is no guarantee that the `/usr/bin` folder
exists, so the `ng-dev release publish` command always fails.
We fix this by just accessing the `env` binary that is also consulted
by scripts executed within Bazel actions. The `env` binary can be
assumed exist in the shell's `$PATH` and can point us to Python
as if it would be executed within Bazel.
PR Close#41943
Currently we have a common utility method for running commands
in a child process. This method pipes all stdout and stderr, but sets
the `stdin` to `inherited`. This seemed to work as expected in terms of
allowing interactive commands being executed, but it messes with the
TTY in Windows (and potentially other platforms) so that colors and
prompts no longer work properly. See attached screenshot.
We fix this by not inheriting the stdin by default; but exposing
a dedicated method for interactive commands. This results in more
readable and obvious code too, so it's worth making this change
regardless of the TTY issues.
PR Close#41948
Currently if a major release-train in the `release-candidate`/`feature-freeze`
phase becomes `latest`, we intend to set the NPM LTS dist tag for all packages
of the previous major (as the old release train in `latest` moves into LTS phase).
The logic for this exists but the release tool sets the NPM dist tag for
all packages of the new major. This means that the script might error if
a new package is part of the new major; or it could cause a deleted
package to not receive the LTS tag properly.
PR Close#41946
Some localization workflows want to use the extracted source translation
files directy back in the project as a target translation file.
The extraction process generates files that only contain "source" messages
and not "target" messages. This is actually valid for most translation formats
but currently the Angular localization process expects target translation files
to always contain target messages and will stop with an error in this case.
Now, instead of an error, the translation file loader will log a warning,
and then try to falback to a source message, only erroring if this is also
missing.
Fixes#21690
PR Close#41944
Seems like the tagline has changed on the [Angular Twitter account](https://twitter.com/angular) and angular.io. This change makes this doc in line with those two resources.
PR Close#41922
When there are multiple attributes that are marked for i18n translation,
which contain expression bindings, we were generating i18n update op-codes
that did not accurately map to the correct value to be bound in the lView.
Each attribute's bindings were relative to the start of the lView first
attributes values rather than to their own.
This fix passes the current binding index to `generateBindingUpdateOpCodes()`
when processing i18n attributes to account for this.
Fixes#41869
PR Close#41912
When including a component in a template, the component's host element
is immediately appended as child of the parent node upon creation.
Hence, `hostElement.parentNode` will be a valid reference. However, if
the parent component is being inserted as an embedded view—through
`ngIf` for example—then the parent's node itself will not have been
inserted yet. This means that we cannot properly determine the position
of the transition namespace, as any `containsElement` check will return
false given that the partial DOM tree has not been inserted yet, even
though it will be contained within an existing transition namespace once
the partial tree is attached.
This commit fixes the issue by not just looking at the existence of a
parent node, but rather a more extensive check using the driver's
`containsElement` method.
PR Close#19854
Previously example for handling client-side errors in networking didn't work when there was no Internet connection.
Caught error is ProgressEvent in such case, not ErrorEvent and error.error.message is undefined.
PR Close#36608
Currently when the change is made to the code inside the `packages/common/http` folder, both `fw-http` and `fw-common` group members are assigned for review. However only `fw-http` is required and should be sufficient. This commit updates the PullApprove config to exclude `packages/common/http` path from the list of folders that `fw-common` "owns".
PR Close#41887
* `SelectivePreloadingStrategyService` is already provided in the root
module, so it should not be added to `AppRoutingModule` providers
array
PR Close#41859
Previously, the `components-repo-unit-tests` CI job was temporarily
disabled due to a version mismatch between the `rules_nodejs`
dependency version on the two repos (angular/angular and
angular/components).
Now that both repos have been updated to a `rules_nodejs` version
>=2.0.0, we can re-enable the job and have `@angular/components` unit
tests run on every build.
PR Close#41816
Previously, the fallback key used for the CircleCI cache could match a
cache indefinitely (as long as `.bazelversion` didn't change). This
would allow the cache to grow quite large, which in turn would lead to
slow-down in CI jobs. See, also, angular/angular-cli#17533 for more
details of the impact of a growing CircleCI cache.
Unfortunately, using something like the lockfile checksum in the
fallback cache key would cause too many cache misses (esp. with
automatic updates via Renovate), again slowing CI down.
(The problem was originally discussed [here][2].)
This commit uses the technique described in [this blogpost][1] to
invalidate the cache monthly. This keeps the extra cache misses low
(essentially once per month per fork), while also preventing the cache
from growing indefinitely.
[1]: https://support.circleci.com/hc/en-us/articles/360012618473-Creating-a-daily-cache
[2]: https://github.com/angular/angular/pull/41467#discussion_r607818494
PR Close#41814
PR Close#41855
In #41788, logic was added to disambiguate case-insensitively equal docs
paths/URLs. This process includes appending a `-\d+` suffix to some
paths/URLs (for example, `/.../inject-1`). Unfortunately, some of the
Firebase redirects configured in `firebase.json` would match these URLs
and redirect them to non-existing paths.
Example failures: [stable][1], [next][2]
NOTE:
This was not picked up in the regular CI tests run for PRs, because the
local devserver and the preview server used to test PRs do not support
Firebase-like redirects.
This commit fixes this by ensuring these disambiguated paths/URLs are
not matched by the redirect rules by checking whether the part of the
suffix after the `-` contains any numeric digits. While this check is
not ideal, it should be good enough for our purpose, since the legacy
URLs that we do want to redirect contain suffixes such as `-class`,
`-function` and thus no numeric digits.
[1]: https://circleci.com/gh/angular/angular/974345
[2]: https://circleci.com/gh/angular/angular/974346
PR Close#41842
PR Close#41854
In ##41788, the `disambiguateDocsPathsProcessor` was introduced to fix
an issue with case-insensitively equal paths. This processor may alter
the paths of some docs and thus their final URL in the app.
Previously, both the `disambiguateDocPathsProcessor` and the
`createSitemap` processor (which relies on the docs' computed paths to
generate the sitemap file) were configured to run before the
"rendering-docs" phase. However, this resulted in the
`disambiguateDocPathsProcessor`'s running after `createSitemap`, which
meant that the sitemap did not include the updated doc paths.
This commit fixes it by ensuring that the
`disambiguateDocPathsProcessor` is explicitly run before the
`createSitemap` processor, so that the latter will be able to take into
account any changes made by the former.
PR Close#41842
PR Close#41854
This commit includes the URL under test in the test description in
`testFirebaseRedirection.spec.ts` to make it easier to identify the
affected URL when a test fails.
It also avoids unnecessarily creating multiple `FirebaseRedirector`
instances by sharing instances between tests.
PR Close#41842
PR Close#41854
This commit updates the utilities in `firebase-test-utils/` to also
support testing Firebase redirects that are configured using regular
expressions (via the `regex` property).
See the [Firebase docs][1] for more details.
[1]: https://firebase.google.com/docs/hosting/full-config#redirects
PR Close#41842
PR Close#41854
Currently, the utilities for testing Firebase redirects assume that the
redirects are configured using the glob-based `source` property.
However, Firebase also supports configuring redirects using regular
expressions (via the `regex` property).
See the [Firebase docs][1] for more details.
This commit refactors the utilities in `firebase-test-utils/` to make it
easy to add support for such regex-based redirect configurations.
[1]: https://firebase.google.com/docs/hosting/full-config#redirects
PR Close#41842
PR Close#41854
Previously, the tests in `create-example.spec.ts` made assertions using
some hard-coded absolute paths (something like `/foo/bar`). This caused
the tests to fail on Windows, where the absolute paths are prefixed with
the drive letter (something like `C:/foo/bar`).
This commit uses `path.resolve()` to ensure paths are converted to the
format used on the current OS.
PR Close#41842
PR Close#41854
Previously, the `autoLinkCode` Dgeni post-processor would not apply the
custom filters when matching the whole contents of a `<code>` element.
This meant that custom filters would not be applied to single-word
`<code>` elements.
You can see occurrences of this issue in the following sections of the
"Reactive forms" guide:
- [Creating nested form groups][1]
(look for `street, city, state, and zip controls`)
- [Using the FormBuilder service to generate controls][2]
(look for `group method`)
This commit fixes this by also applying the custom filters when
processing the whole contents of a `<code>` element.
This commit also updates the `filterPipes` custom filter to allow
matching a pipe's name in a single-word `<code>` element (where there is
no preceeding `|` character).
[1]: https://v10.angular.io/guide/reactive-forms#creating-nested-form-groups
[2]: https://v10.angular.io/guide/reactive-forms#using-the-formbuilder-service-to-generate-controls
PR Close#41709
PR Close#41854
While generating the docs, when a `<code>` element is inspected for
auto-linking, the `autoLinkCode` Dgeni post-processor will break its
contents up into words and generate text nodes for those words that
should not be auto-linked.
Previously, our text node visitor would visit these generated text nodes
and try to auto-link them too. As a result, it would unnecessarily
process nodes that had already been checked (and could potentially
generate links that would otherwise be ignored).
You can see an occurrence of this issue in the
[Create the product list][1] section of the
"Getting started with Angular" tutorial (look for `<a>`).
This commit fixes this by ensuring the visitor will skip the current
node and any nodes generated by `autoLinkCode`.
[1]: https://v11.angular.io/start#create-the-product-list
PR Close#41709
PR Close#41854
This section states that "Importing a module with services means that you will have a new instance of that service". This is only true for lazy-loaded `NgModules`. For non-lazy-loaded modules, my understanding is that the providers arrays are flattened into the root injector meaning that importing a module with a service doesn't create a new instance of that service.
PR Close#41835
`@font-face` rules cannot contain nested selectors. Nor can they be
nested under a selector. Normally this would be a syntax error by the
author of the styles. But in some rare cases, such as importing styles
from a library, and applying `:host ::ng-deep` to the imported styles,
we can end up with broken css if the imported styles happen to contain
`@font-face` rules.
This commit works around this problem by sanitizing such cases (erasing
any scoping selectors) during emulated ShadowDOM encapsulation style
processing.
Fixes#41751
PR Close#41815
Misko Hevery