This commit updates the `NgOptimizedImage` directive to add a logic to detect whether an image, marked with the "priority" attribute has a corresponding `<link rel="preconnect">` tag in the `document.head` for better performance.
PR Close#47082
This commit adds a built-in Imgix loader for the
NgOptimizedImage directive. If you provide the
desired Imgix hostname, an ImageLoader will be
generated with the correct options.
Usage looks like this:
```ts
providers: [
provideImgixLoader('https://some.imgix.net')
]
```
It sets the "auto=format" flag by default, which
ensures that the smallest image format supported
by the browser is served.
This change also moves the IMAGE_LOADER, ImageLoader,
and ImageLoaderConfig into a new directory that will
be shared by all built-in image loaders.
PR Close#47082
This commit updates the error codes used in the `NgOptimizedImage` tests. The error codes got u[dated recently in a PR that got merged earlier.
PR Close#47082
This commit adds a `rawSrcset` attribute as a replacement for the
`srcset` attribute. The `srcset` attribute cannot be set statically
on the image directive because it would cause images to start
downloading before the "loading" attribute could be set to "lazy".
Changing the name to `rawSrcset` allows the directive to control
the timing of image loading. It also makes it possible to support
custom loaders for `srcset` file names. Rather than having to repeat
the image origin for each image, the existing `rawSrc` value and
image loader can be composed to generate each full URL string. The
developer must only provide the necessary widths for the `srcset`.
For example, the developer might write:
```markup
<img rawSrc="hermes.jpg" rawSrcset="100w, 200w" ... />
```
with a loader like:
```js
const loader = (config) => `path/${config.src}?w=${config.width}`;
```
and the img tag will ultimately be set up as something like:
```markup
<img src="path/hermes.jpg?w=100" srcset="path/hermes.jpg?w=100 100w, path/hermes.jpg?w=200 200w" .../>
```
PR Close#47082
Currently if you use a static `srcset` with the image directive,
lazy loading no longer works because the image would start to
load before the loading attribute could be set to "lazy". This
commit throws an error if you try to use `srcset` this way.
In a follow-up commit, a new attribute will be added to support
responsive images in a lazy-loading-friendly way.
PR Close#47082
This commit updates the `NgOptimizedImage` directive to drop the `standalone` flag and create a new NgModule which declares and exports it, so that the directive can be used in apps that use pre-v14 version of Angular.
PR Close#47082
The CI failed to run on the last PR, so we didn't catch
that there were web tests failing due to absolute/relative
URL issues. This commit should fix the issue.
PR Close#47082
Previously NgOptimizedImage would default to requesting
an image at the width matching the width attribute in
the image HTML. While this works for width attrs that
match the intrinsic size of the image (e.g. responsive
images or images sized with CSS), this can be a sharp
edge for developers who use the width/height attributes
to set rendered size (i.e. instead of CSS, which one can
do for a fixed size image). In this case, defaulting to
the width attribute for the requested image would mean
requesting always at 1x quality, so screens with a DPR
of 2+ get an image that is too small. Without a default
request width, the image served by the CDN would likely
be at intrinsic size, so 2x images would look correct.
This PR also updates the NgOptimizedImage sandbox and tests.
PR Close#47082
This commit optimizes the logic that monitors whether a give image is an LCP element. If an image has the `priority` attribute set, there is no need to include it into monitoring. Also, if we already warned about a particular image (via a `console.warn`) - there is no need to warn again later (to avoid spamming a console).
PR Close#47082
This commit adds e2e tests for the LCP check logic. Those tests are needed to verify the behavior in a real browser (vs relying on a Node environment).
PR Close#47082
This commit adds extra logic into the `NgOptimizedImage` experimental directive to detect an LCP image and assert whether the `priority` attribute is applied.
PR Close#47082
Currently, the logic that sets the `src` on the host `<img>` element is located in the `ngOnChanges` lifecycle hook and guarded by the dev-mode checks that the inputs do not change. However, those checks would be tree-shaken in prod mode and the `src` would be set each time the `ngOnChanges` hook is invoked. This is undesirable and may potentially lead to performance issues.
This commit moves the `src`-related logic to the `ngOnInit` hook instead, which would have the same effect (executed only once, after all inputs are set) and would behave consistently in dev and prod modes.
PR Close#47082
This commit updates the `NgOptimizeImage` directive to add asserts to make sure the `width` and `height` inputs are set and have correct values (numbers only).
PR Close#47082
This commit updates the `NgOptimizedImage` directive as standalone, so it's easier to import it in
an app (without importing any NgModules). The `NgOptimizedImageModule` is removed as no longer needed.
PR Close#47082
This commit updates the `NgOptimizedImage` directive to drop the `loader` input. Component-specific loaders can still be configured via `IMAGE_LOADER` token and the `loader` input was only useful in case different loaders have to be present in a single template, which doesn't seem to be a common case. We'll be able to re-introduce the input later if needed.
PR Close#47082
We want it to be clear what benefits the image directive
confers over a normal <img> tag, and the `NgImage` name
didn't provide much information. `NgOptimizedImage` makes
it obvious that the new directive is intended to improve
performance.
PR Close#47082
This commit exports experimental `NgImage` directive via a private API of the `@angular/common` package, so that it can be used acorss other Angular packages for testing purposes.
PR Close#47082
This brings the typing of ngStyle into parity with ngClass since commit
e2ab99b95e. Should help with some strict template checking edge cases.
PR Close#47069
When using the Angular Router, one of `APP_BASE_HREF` or a `<base>` in
the header must be provided. When _not_ using the `RouterModule`,
injecting the `LocationStrategy` will result in the
`PathLocationStrategy` being provided with a default value used in place
of `APP_BASE_HREF` that is `document?.location?.origin ?? ''`.
It can be quite surprising and annoying that once you add `RouterModule`
to the application, suddenly the `APP_BASE_HREF` must be specifically
provide something new when it could use a sensible default instead.
The current behavior (before this commit) is as follows:
* When `RouterModule` is not provided (or the dev doesn't specifically provide
`PathLocationStrategy`): use `DOCUMENT.location?.origin ?? ''`.
Note that the base href in the dom and `APP_BASE_HREF` are not used.
* When `RouterModule` _is_ provided:
1. APP_BASE_HREF if defined
2. Get base href from DOM
3. throw if neither of the two above are defined
This commit updates this behavior to be aligned regardless of `RouterModule`
usage. The order (by default) is now:
1. Developer provided `APP_BASE_HREF`
2. base href from the DOM
3. `location.origin`
4. If none of the above exist, use `''`
This is slightly different than the behavior before. However, I believe
it is more appropriate. For the case without `RouterModule`, it would
likely be surprising that `APP_BASE_HREF` and the base href from the DOM
are ignored by default. For the case with `RouterModule`, we now have a
more sensible fallback/default when neither `APP_BASE_HREF` nor `<base>`
are defined (instead of just throwing an error).
PR Close#46929
With strict template type checking, a null/undefined value will raise an
error. However the implementation is completely fine with it, and it
would be pointless to "fix" it at the callsite and convert to e.g. an
empty string. Allow all of the "supported types" to be passed in
directly to ngClass.
Fixes#39280
PR Close#46906
This commit adds the `createNgModuleRef` function alias to the public API. The alias is called `createNgModule`. The `createNgModule` name is more consistent with the rest of the API surface, where functions that return `*Ref`s don't include the `Ref` into the function name, for example: `createPlatform`, `ViewContainerRef.createComponent`, etc.
DEPRECATED:
The `createNgModuleRef` is deprecated in favor of newly added `createNgModule` one.
PR Close#46789
This commit cleans-up/removes a check we added before we supported CLDR
39. This check was necessary due to a incomplete/invalid list of locales
provided as part of the JSON data.
PR Close#46606
This commit updates the directives presents in the `CommonModule` and annotates them with the `standalone: true` flag. With that flag, the directives can now be imported individually, as well as imported via the `CommonModule`.
PR Close#46469
This commit updates the pipes presents in the `CommonModule` and annotates them with the `standalone: true` flag. With that flag, the pipes can now be imported individually, as well as imported via the `CommonModule`.
PR Close#46401
Simplifying the tree shakable error messages with the new format and removing the errorMessage variables
```ts
throw new RuntimeError(
RuntimeErrorCode.INJECTOR_ALREADY_DESTROYED,
ngDevMode && 'Injector has already been destroyed.');
```
PR Close#46382
`ngComponentOutlet` already handles null/undefined values, but the types don't reflect that. These changes update the types.
Fixes#45716.
PR Close#46280
To mitigate the effect of Observables with memory leak, this change clears
the `ChangeDetectorRef` when the `AsyncPipe` is destroyed. This avoids any
leak within the `Observable` to retain the view data. A test has been added
to verify that this change works correctly with promises that resolve _after_
the pipe has been destroyed.
Note: this is not marked as fix as `AsyncPipe` itself is _not_ leaking any
memory.
Closes#17624
PR Close#46128
The field's type does not account for the `null` value that is assigned
in the field's initializer. This commit includes `null` in the type for
better type-safety. Exising reads are updated with a non-null assertion
operation in places where initialization is known to have occurred.
PR Close#46128
tweak the current structural directives guide (currently mainly targeted
at the creation of custom structural directives) so that is more generic
and a point of reference for structural directives in general
this also includes the re-addition of the one-per-element section
removed in PR #40015resolves#44786
PR Close#44895
The NodeJS Bazel linker does not work well on Windows because there
is no sandboxing and linker processes from different tests will attempt
to modify the same `node_modules`, causing concurrency race conditions
and resulting in flakiness.
PR Close#45872
Before this commit, when initializing `HttpParams` with:
const body = new HttpParams({fromObject: {b: 2}});
then `body.get('b')` returned `2` instead of `'2'` as expected.
This commit makes sure the values are converted to strings in such cases.
Fixes#42641
PR Close#42643
Update CLDR, which is used to generate the locales files, to version 41.
Also, make necessary code changes to account for changes in the CLDR
data.
Fixes#43301
PR Close#45714
While generating locales, two sets of data (both derived from the
[CLDR project][1]) are used:
- JSON data with the [`cldrjs` package][2], which is used for most
operations.
- XML data with the [`cldr` package][3], which is used for generating
plural-related data.
The JSON data is brought in from the [unicode-org/cldr-json][4]
repository. Since we control the version of the repository that we use,
we can control the CLDR version that these correspond to.
Previously, however, we used the XML data that were bundled with the
[`cldr` package[3]. As a result, the two sets of data could correspond
to different CLDR versions, resulting in incorrect/inconsistent locales
files.
This commit addresses the problem by utilizing the `load()` method of
the [`cldr` package][3], which allows passing in a custom path to the
CLDR XML data (instead of using the bundled data. This way, we can
ensure that the data used for all operations correspond to the same CLDR
version.
Related discussion: #43301
[1]: https://github.com/unicode-org/cldr
[2]: https://www.npmjs.com/package/cldrjs
[3]: https://www.npmjs.com/package/cldr
[4]: https://github.com/unicode-org/cldr-json
PR Close#45714
Previously, some functions/targets related to the
generation/verification of the `closure-locale.ts` file included
`closure_locales` instead of `closure_locale` in their name, which was
inconsistent and confusing.
Rename all tooling to use `closure_locale` (which more closely matches
the generated file's name).
PR Close#45714
Andrew Kushnir