Our approach for handling cyclic imports results in code that is
not easy to tree-shake, so it is not suitable for publishing in a
library.
When compiling in partial compilation mode, we are targeting
such library publication, so we now create a fatal diagnostic
error instead of trying to handle the cyclic import situation.
Closes#40678
PR Close#40782
ChromeDriver now supports Apple Silicon ARM processors.
`webdriver-manager` versions 12.1.7 and earlier will, however,
incorrectly download the arm64 ChromeDriver regardless of the
system's architecture. This results in failure to run Protractor tests
on macOS with the error: `spawn Unknown system error -86`
This commit fixes the problem by upgrading `webdriver-manager` to
version 12.1.8, which includes a fix.
See also https://stackoverflow.com/questions/65618558.
PR Close#40885
Previously, the paths to some local dependencies in the `ivy-i18n`
integration project were pointing to non-generic bazel files that would
normally not exist on a user's machine.
This commit changes these paths to relative paths pointing to build
artifacts that are guaranteed to exist (once the necessary build
commands have been executed).
PR Close#40885
Fix router to ensure that a route module is only loaded once especially
in relation to the use of preload strategies with delayed or partial
loading.
Add test to check the interaction of PreloadingStrategy and normal
router navigation under differing scenarios.
Checking:
* Prevention of duplicate loading of modules.
related to #26557
* Prevention of duplicate RouteConfigLoad(Start|End) events
related to #22842
* Ensuring preload strategy remains active for submodules if needed
The selected preload strategy should still decide when to load submodules
* Possibility of memory leak with unfinished preload subscription
related to #26557
* Ensure that the stored loader promise is cleared so that subsequent
load will try the fetch again.
* Add error handle error from loadChildren
* Ensure we handle error from with NgModule create
Fixes#26557#22842#26557
PR Close#40389
The previous commits refactored the `ShadowCss` emulator to support
desirable use-cases of `:host-context()`, but it dropped support
for passing a comma separated list of selectors to the `:host-context()` .
This commit rectifies that omission, despite the use-case not being
valid according to the ShadowDOM spec, to ensure backward compatibility
with the previous implementation.
PR Close#40494
In `ViewEncapsulation.Emulated` mode the compiler converts `:host` and
`:host-context` pseudo classes into new CSS selectors.
Previously, when there was both `:host-context` and `:host` classes in a
selector, the compiler was generating incorrect selectors. There are two
scenarios:
* Both classes are on the same element (i.e. not separated). E.g.
`:host-context(.foo):host(.bar)`. This setup should only match the
host element if it has both `foo` and `bar` classes. So the generated
CSS selector should be: `.foo.bar<hostmarker>`.
* The `:host` class is on a descendant of the `:host-context`. E.g.
`:host-context(.foo) :host(.bar)`. This setup should only match the
`.foo` selector if it is a proper ancestor of the host (and not on the
host itself). So the generated CSS selector should be:
`.foo .bar<hostmarker>`.
This commit fixes the generation to handle these scenarios.
Fixes#14349
PR Close#40494
In `ViewEncapsulation.Emulated` mode, the compiler must generate additional
combinations of selectors to handle the `:host-context()` pseudo-class function.
Previously, when there is was more than one `:host-context()` selector in a
rule, the compiler was generating invalid selectors.
This commit generates all possible combinations of selectors needed to
match the same elements as the native `:host-context()` selector.
Fixes#19199
PR Close#40494
In #40419 we move the code that creates the `heroes` object
above the code that updates the message service. But this moved
that line of code outside of the docregion that is displayed in
the tutorial ealier on, making it non-sensical.
This commit ensures that this line of code is now rendered in the
tutorial be moving it within the correct docregion.
Closes#40788
PR Close#40789
On larger screens the top-bar has a height of 64px. On screens smaller
than 600px, the top-bar has a height of 56px. As a result, the sidenav
should have a top position of 56px on screens smaller than 600px and
64px on other screens.
Previously, the style setting the top position to 56px was tied to the
presence of the `.collapsed` class, which depends on whether the sidenav
is docked or not. The change from docked to collapsed sidenav, however,
happens at 992px. As a result, the sidenav had an incorrect top position
(56px instead of 64px) on screens between 600px and 991px.
This commit fixes this by ensuring the change of the top position for
the sidenav happens at 600px.
PR Close#40802
Both `CodeExampleComponent` and `CodeTabsComponent` components receive
some code via content projection, grab the projected content and pass it
through to a `CodeComponent` instance for formatting and displaying.
Previously, the projected content was kept in the DOM (hidden). This
unnecessarily increased the number of DOM nodes.
This commit fixes this by clearing the projected DOM nodes once their
content has been captured.
PR Close#40802
Update the HTTP guide and associated example to demonstrate
how an interceptor can be used to provide a custom JSON parser.
Resolves#21079
PR Close#40645
After the docs UI redesign `h2` tags got a border top.
`border-top: 1px solid #dbdbdb;`;
In the sections of Getting Started guide in order to separate
`What's next` from above content an `<hr />` tag was used,
that now becomes unnecessary.
This commit removes unnecessary `<hr />` tags.
PR Close#40693
The "Usage Notes" section is often at the bottom of a long page of content
which can make it difficult to discover. This commit adds a link at the
end of the "Description" if there is a "Usage Notes" section on the page.
This link should help the reader to discover that there is more information
available further down the page.
This is basically a marginal fix. The longer term fix should be to
consider what content goes in which sections and how they should
be laid out on the page.
See #40753
PR Close#40835
The CDK has had a service for copying strings to the clipboard. These changes switch
AIO to it, rather than having to maintain a custom solution.
PR Close#40840
In Chrome 83 passing a TrustedScript to eval just returns the
TrustedScript back without evaluating it, causing the
newTrustedFunctionFor{Dev,JIT} functions to fail. This is a browser bug
that has been fixed in Chrome 84, and only affects Angular applications
running with JIT (which includes unit tests).
As a temporary workaround for users still on Chrome 83, detect when this
occurs in the newTrustedFunctionFor* functions and fall back to the
straightforward, non-Trusted Types compatible implementation. The only
combination that is left affected consists of Angular applications
running with JIT, that have explicitly configured Trusted Types in
enforcement mode, with users that are still on Chrome 83.
Also correct docstring for newTrustedFunctionForJIT.
PR Close#40815
Initially raised due to #39509, this change updates our
"Support policy and schedule" with the same disclaimer
from the "Release frequency" section above.
PR Close#40770
This commit implements creating of `ɵɵngDeclarePipe()` calls in partial
compilation, and processing of those calls in the linker and JIT compiler.
See #40677
PR Close#40803
