The ViewEngine message extraction would trim the values
of the `equiv-text` attributes. This commit aligns the Ivy
extraction of these attributes.
Fixes#41176
PR Close#41180
Previously, the generated `404.html` page did not include a `<body>`
tag. In some browsers (such as IE 11), this was causing warnings in the
console.
This commit ensures the generated page contains a `<body>` tag. It also
fixes the indentation in the generated page.
PR Close#41163
Previously, the angular.io app was broken on IE 11. In particular, pages
that included Custom Elements would fail to load, because the
`Reflect.construct()` method (which the Custom Elements ES5 shim relies
on) was not available.
This commit fixes this by loading the polyfill for `Reflect.construct()`
on browsers that do not support ES2015 (including IE 11).
PR Close#41162
The custom elements spec is not compatible with ES5 style classes. This
means ES2015 code compiled to ES5 will not work with a native
implementation of Custom Elements. To support browsers that natively
support Custom Elements but not ES2015 modules, we load
`@webcomponents/custom-elements/src/native-shim.js`, which minimally
augments the native implementation to be compatible with ES5 code.
(See [here][1] for more details.)
Previously, the shim was included in `polyfills.ts`, which meant it was
loaded in all browsers (even those supporting ES2015 modules and thus
not needing the shim).
This commit moves the shim from `polyfills.ts` to a `nomodule` script
tag in `index.html`. This will ensure that it is only loaded in browsers
that do not support ES2015 modules and thus do not needed the shim.
NOTE:
This commit also reduces size of the polyfills bundle by ~400B
(52609B --> 52215B).
[1]: https://www.npmjs.com/package/@webcomponents/custom-elements#es5-vs-es2015
PR Close#41162
The `DomAdapter` is present in all Angular apps and its methods aren't tree shakeable.
These changes remove the methods that either aren't being used anymore or were only
used by our own tests. Note that these changes aren't breaking, because the adapter
is an internal API.
The following methods were removed:
* `getProperty` - only used within our own tests.
* `log` - Guaranteed to be defined on `console`.
* `logGroup` and `logGroupEnd` - Only used in one place. It was in the DomAdapter for built-in null checking.
* `logGroupEnd` - Only used in one place. It was placed in the DomAdapter for built in null checking.
* `performanceNow` - Only used in one place that has to be invoked through the browser console.
* `supportsCookies` - Unused.
* `getCookie` - Unused.
* `getLocation` and `getHistory` - Only used in one place which appears to have access to the DOM
already, because it had direct accesses to `window`. Furthermore, even if this was being used
in a non-browser context already, the `DominoAdapter` was set up to throw an error.
The following APIs were changed to be more compact:
* `supportsDOMEvents` - Changed to a readonly property.
* `remove` - No longer returns the removed node.
PR Close#41102
With this change we remove the lines length restriction in commit message body. This was previously in place due to a bug in GitHub UI, which is now fixed.
PR Close#41157
When there are elements in a translated message, the start and end tags
are encoded as placeholders. The names of these placeholders are computed
from the name of the element. For example `<a> will be `START_LINK` and
`</a>` will be `CLOSE_LINK`.
If there are more than one element with the same name, but different attributes,
then the starting placeholder name is made unique.
For example `<a href="a">` would be `START_LINK`, while `<a href="b">` in
the same message would then be called `START_LINK_1`.
But the closing tags will not be made unique since there are no attrbutes;
the always have the same text `</a>`, which will produce, for example,
`CLOSE_LINK`.
Previously, when extracting XLIFF 2 formatted translation files, the closing
tag placeholder names were computed incorrectly from the opening tag
placeholder names. For example `CLOSE_LINK_1`.
This commit strips these `_1` type endings from the start tag placeholder
name when computing the closing tag placeholder name. It also ensures
that the `type` of the placeholder is computed accurately in these cases
too.
Fixes#41142
PR Close#41152
The Angular compiler creates two `ts.Program`s; one for emit and one for
template type-checking. The creation of the type-check program could
benefit from reusing the `ts.ModuleResolutionCache` that was primed
during the creation of the emit program. This requires that the compiler
host implements `resolveModuleNames`, as otherwise TypeScript will setup
a `ts.ModuleResolutionHost` of its own for both programs.
This commit ensures that `resolveModuleNames` is always implemented,
even if the originally provided compiler host does not. This is
beneficial for the `ngc` binary.
PR Close#39693
Previously, the indentation of code snippets in the "Cheat sheet" guide
was done using `<br>` elements (for line breaks) and ` ` HTML
entities (for space). This was laborious and put the onus on the author
to remember to use these symbols (instead of regular whitespace
characters).
(Discussed in
https://github.com/angular/angular/pull/41051#discussion_r585651621.)
This commit changes the way `<code>` elements are styles inside the
"Cheat sheet" guide to allow using regular whitespace characters in code
snippets. It also changes all `<br>`/` ` occurrences to `\n`/` `
respectively to make code snippets more readable in the source code.
PR Close#41051
The "Features" page organizes features in groups/rows of 3 features
each. On wide screens, all 3 paragraphs of a group/row can be shown next
to each other. On narrow screens (between 768px and 1057px), the layout
changes to stack the paragraphs vertically. On medium screens, however,
there is not enough space to show more than two paragraphs next to each
other.
Previously, the 3rd paragraph was wrapped over to the next line.
This commit improves the layout on medium screens by switching to
immediately stacking the paragraphs vertically as soon as there is not
enough space for them to be displayed in one row. Since the total width
is still too much for one paragraph, the paragraphs are limited to 80%
of the total width.
Before (on 1000px width): [features page (on 1000px) before][1]
After (on 1000px width): [features page (on 1000px) after][2]
[1]: https://user-images.githubusercontent.com/8604205/109825316-62128a00-7c42-11eb-8391-650201257274.png
[2]: https://user-images.githubusercontent.com/8604205/109825323-6343b700-7c42-11eb-86c1-e8307c5a727a.png
PR Close#41051
Previously, with the min width of 220px per item, several API list items
were truncated.
This commit increases the min width per item to 330px, which allows
almost all items to have their full text shown. It also increases the
API list page's max content width from 50em (800px) to 62.5em (1000px)
to allow items to be shown on three columns despite their increased
width. This increase in the content width shouldn't negatively affect
UX, since the API list page uses a multi-column layout (i.e. it does not
contain 1000px-lines of text.)
Before: ![api-list before][1]
After: ![api-list after][2]
[1]: https://user-images.githubusercontent.com/8604205/109396457-5f5e1f00-793a-11eb-80cf-1418f409325a.png
[2]: https://user-images.githubusercontent.com/8604205/109396659-499d2980-793b-11eb-95d3-f54250f7fab5.png
PR Close#41051
Previously, each marketing page used a different limit for its content's
width (if it had a limit at all) and implemented the width limiting in a
different way. Besides resulting in an inconsistent UX, this also made
it difficult to apply site-wide layout changes.
This commit makes the limit for most marketing pages consistent and uses
the same CSS class to make it easier to apply site-wide changes in the
future. The chosen limit is slightly larger than that of docs pages
(62.5em/1000px vs 50em/800px), because marketing pages have a different
type of content and layout (i.e. images, multi-column layout, etc.).
Finally, this commit also removes obsolete wrapper elements, CSS classes
and CSS styles, that are no longer necessary after the changes.
Notably, the homepage (`/`) and the "Contributors" page (`/about`) have
remained unchanged, because the former has its own layout that is
different from other marketing pages and the latter would offer a worse
UX with a small content width limit (as the one used on other marketing
pages).
The content widths of the rest of the marketing pages change slightly as
a result of the changes in this commit, but not in a way that would have
a negative impact on UX. More specifically:
| Page (URL) | Size before | Size after |
|:--------------|------------:|-----------:|
| `/contribute` | 880px | 1000px |
| `/events` | unlimited | 1000px |
| `/features` | 996px | 1000px |
| `/presskit` | 800px | 1000px |
| `/resources` | 800px | 1000px |
PR Close#41051
This commit removes an unnecessary wrapper `<div>` from the
"Cheat sheet" guide. The CSS styles that referenced the element's ID
(`#cheatsheet`) have been updated to use `.page-guide-cheatsheet`
instead.
PR Close#41051
Previously, styling of `<code>` elements utilized the `:not()` CSS
pseudo-class with multiple selectors (`:not(h1, h2, ...)`). It turns out
that older browsers (such as IE11) do not support multiple selectors in
a single `:not()` instance.
(See [MDN][1] and [CanIUse][2] for more info.)
This commit fixes `<code>` styling to use multiple separate `:not()`
instances instead (`:not(h1):not(h2)...`), so that they are styled
correctly on older browsers as well.
NOTE:
This change seems to trigger some kind of bug in LightHouse that causes
the a11y score of `/start` to be calculated as 0 (which is clearly
wrong). This happens on Linux (tested on CI and locally using the
Windows Subsystem for Linux (WSL)) - on Windows the score is computed
correctly as 98/100.
([Example failure][3])
The bug seems to be related to the layout of the content and goes away
if we change the viewport size (for example, switching to LightHouse's
`desktop` config) or make another change that affects the content's
layout (for example, reducing the padding of `<code>` elements).
To work around the issue, this commit updates the `test-aio-a11y.js`
script to test `/start-routing` instead of `/start`.
[1]: https://developer.mozilla.org/en-US/docs/Web/CSS/:not#description:~:text=Using%20two%20selectors
[2]: https://caniuse.com/css-not-sel-list
[3]: https://circleci.com/gh/angular/angular/931038
PR Close#41051
The "Features" page organizes features in groups/rows of 3 features
each. On wide screens, all 3 paragraphs of a group/row can be shown next
to each other. On narrow screens, the layout changes to stack the
paragraphs vertically. On medium screens, however, the 3rd paragragh is
wrapped over to the next line.
Previously, the wrapped content was left-aligned, which left a lot of
empty space on the right.
This commit improves the layout on medium screens by ensuring the
paragraphs are horizontally centered (with space distributed evenly
around them).
Before: ![features page before][1]
After: ![features page after][2]
[1]: https://user-images.githubusercontent.com/8604205/109344670-b64ef000-7877-11eb-9013-890562ff2f3d.png
[2]: https://user-images.githubusercontent.com/8604205/109344678-b7801d00-7877-11eb-9224-d7715f7d7235.png
PR Close#41051
Previously, in contrast to docs pages, marketing pages had a non-fixed
top-menu, which meant that the top-menu would scroll out of the viewport
along with the rest of the content. This had a couple of downsides:
- The UI was different between pages (i.e. different top-menu behavior
on docs vs marketing pages).
- Since some of the marketing pages are long, it was not easy for people
to navigate to a different page (i.e. they had to scroll all the way
back up).
This commit improves the UX by using the same, fixed top-menu on all
pages, which restores consistency and allows the user to navigate around
more easily.
NOTE:
The old behavior (non-fixed top-menu) is kept on the homepage, since its
top-menu design in a little different than other pages (e.g. it uses a
transparent top-menu) and would not play well with a fixed top-menu.
PR Close#41051
One of the main goals of the bundling tests is to verify that unused symbols are tree-shaken away in prod bundles.
Currently both Reactive and Template-driven test apps are merged into one. In order to make these tree-shaking
tests even more useful, this commit splits exiting test app into two, so that we can further optimize sets of
symbols that should be retained in both scenarios.
PR Close#41108
Previously, injector definitions contained a `factory` property that
was used to create a new instance of the associated NgModule class.
Now this factory has been moved to its own `ɵfac` static property on the
NgModule class itself. This is inline with how directives, components and
pipes are created.
There is a small size increase to bundle sizes for each NgModule class,
because the `ɵfac` takes up a bit more space:
Before:
```js
let a = (() => {
class n {}
return n.\u0275mod = c.Cb({type: n}),
n.\u0275inj = c.Bb({factory: function(t) { return new (t || n) }, imports: [[e.a.forChild(s)], e.a]}),
n
})(),
```
After:
```js
let a = (() => {
class n {}
return n.\u0275fac = function(t) { return new (t || n) },
n.\u0275mod = c.Cb({type: n}),
n.\u0275inj = c.Bb({imports: [[r.a.forChild(s)], r.a]}),
n
})(),
```
In other words `n.\u0275fac = ` is longer than `factory: ` (by 5 characters)
and only because the tooling insists on encoding `ɵ` as `\u0275`.
This can be mitigated in a future PR by only generating the `ɵfac` property
if it is actually needed.
PR Close#41022
This commit adds a semi-comprehensive README file which describes the
design goals and implementation of the template type checking engine,
which powers the Angular Language Service as well as the main compiler's
understanding of types in templates.
PR Close#41004
The compiler performs cycle analysis for the used directives and pipes
of a component's template to avoid introducing a cyclic import into the
generated output. The used directives and pipes are represented by their
output expression which would typically be an `ExternalExpr`; those are
responsible for the generation of an `import` statement. Cycle analysis
needs to determine the `ts.SourceFile` that would end up being imported
by these `ExternalExpr`s, as the `ts.SourceFile` is then checked against
the program's `ImportGraph` to determine if the import is allowed, i.e.
does not introduce a cycle. To accomplish this, the `ExternalExpr` was
dissected and ran through module resolution to obtain the imported
`ts.SourceFile`.
This module resolution step is relatively expensive, as it typically
needs to hit the filesystem. Even in the presence of a module resolution
cache would these module resolution requests generally see cache misses,
as the generated import originates from a file for which the cache has
not previously seen the imported module specifier.
This commit removes the need for the module resolution by wrapping the
generated `Expression` in an `EmittedReference` struct. This allows the
reference emitter mechanism that is responsible for generating the
`Expression` to also communicate from which `ts.SourceFile` the
generated `Expression` would be imported, precluding the need for module
resolution down the road.
PR Close#40948
The import graph scans source files for its import and export statements
to extract the source files that it imports/exports. Such statements
contain a module specifier string and this module specifier used to be
resolved to the actual source file using an explicit module resolution
step. This is especially expensive in incremental rebuilds, as the
module resolution cache has not been primed during program creation
(assuming that the incremental program was able to reuse the module
resolution results from a prior compilation). This meant that all module
resolution requests would have to hit the filesystem, which is
relatively slow.
This commit is able to replace the module resolution with TypeScript's
bound symbol of the module specifier. This symbol corresponds with the
`ts.SourceFile` that is being imported/exported, which is exactly what
the import graph was interested in. As a result, no filesystem accesses
are done anymore.
PR Close#40948
This is necessary for closure compiler in order to support cross chunk code motion.
I'm intentionally not annotating these call sites with @__PURE__ at the moment because
build-optimizer does it within the Angular CLI build. In the future we might want to consider
having both annotations here in case we change how build-optimizer works.
PR Close#41096
This change marks all relevant define* callsites as pure, causing the compiler to
emmit either @__PURE__ or @pureOrBreakMyCode annotation based on whether we are
compiling code annotated for closure or terser.
This change is needed in g3 where we don't run build optimizer but we
need the code to be annotated for the closure compiler.
Additionally this change allows for simplification of CLI and build optimizer as they
will no longer need to rewrite the generated code (there are still other places where
a build optimizer rewrite will be necessary so we can't remove it, we can only simplify it).
PR Close#41096
Adds a new flag to `localize-extract` called `--migrateMapFile` which will generate a JSON file
that can be used to map legacy message IDs to cannonical ones.
Also includes a new script called `localize-migrate` that can take the mapping file which was
generated by `localize-extract` and migrate all of the IDs in the files that were passed in.
PR Close#41026
Previously, when a payload size check failed, the error message prompted
the user to update the size limits using a CI-specific file path, which
was confusing (esp. for external contributors). See, for example,
https://circleci.com/gh/angular/angular/932733.
This commit improves the error message by printing the file path
relative to the repository root instead.
PR Close#41116
In Angular programs, changing a file may require other files to be
emitted as well due to implicit NgModule dependencies. For example, if
the selector of a directive is changed then all components that have
that directive in their compilation scope need to be recompiled, as the
change of selector may affect the directive matching results.
Until now, the compiler solved this problem using a single dependency
graph. The implicit NgModule dependencies were represented in this
graph, such that a changed file would correctly also cause other files
to be re-emitted. This approach is limited in a few ways:
1. The file dependency graph is used to determine whether it is safe to
reuse the analysis data of an Angular decorated class. This analysis
data is invariant to unrelated changes to the NgModule scope, but
because the single dependency graph also tracked the implicit
NgModule dependencies the compiler had to consider analysis data as
stale far more often than necessary.
2. It is typical for a change to e.g. a directive to not affect its
public API—its selector, inputs, outputs, or exportAs clause—in which
case there is no need to re-emit all declarations in scope, as their
compilation output wouldn't have changed.
This commit implements a mechanism by which the compiler is able to
determine the impact of a change by comparing it to the prior
compilation. To achieve this, a new graph is maintained that tracks all
public API information of all Angular decorated symbols. During an
incremental compilation this information is compared to the information
that was captured in the most recently succeeded compilation. This
determines the exact impact of the changes to the public API, which
is then used to determine which files need to be re-emitted.
Note that the file dependency graph remains, as it is still used to
track the dependencies of analysis data. This graph does no longer track
the implicit NgModule dependencies, which allows for better reuse of
analysis data.
These changes also fix a bug where template type-checking would fail to
incorporate changes made to a transitive base class of a
directive/component. This used to be a problem because transitive base
classes were not recorded as a transitive dependency in the file
dependency graph, such that prior type-check blocks would erroneously
be reused.
This commit also fixes an incorrectness where a change to a declaration
in NgModule `A` would not cause the declarations in NgModules that
import from NgModule `A` to be re-emitted. This was intentionally
incorrect as otherwise the performance of incremental rebuilds would
have been far worse. This is no longer a concern, as the compiler is now
able to only re-emit when actually necessary.
Fixes#34867Fixes#40635Closes#40728
PR Close#40947
This commit refactors Ivy runtime code to move `readPatchedData` and `attachPatchedData` functions to a single
location for better maintainability and to make it easier to do further changes if needed. The `readPatchedLView`
function was also moved to the same location (since it's a layer on top of the `readPatchedData` function).
PR Close#41097
Pete Bacon Darwin
Stepan Suvorov
George Kalpakas
Kristiyan Kostadinov
pniedermeyer
Alan Agius
Andrew Kushnir
Andrew Kushnir
Joey Perrott
JoostK
Theoklitos Bampouris
Aristeidis Bampakos
Alex Rickabaugh
Igor Minar
Jason Bedard
Sagar Pandita
abarghoud
LordMsz
