Correctly wire up hierarchical injectors for downgraded components in
`upgrade/static`: Downgraded components inherit the injector of the first
downgraded component up the DOM tree.
This is similar to (part of) d91a86a, but for `upgrade/static`.
POSSIBLE BREAKING CHANGE:
In order to enable more control over the wiring of downgraded components and
their content (which eventually allows better control over features like
injector setup and content projection), it was necessary to change the
implementation of the directives generated for downgraed components.
The directives are now terminal and manually take care of projecting and
compiling their contents in the post-linking function. This is similar to how
the dynamic version of `upgrade` does it.
This is not expected to affect apps, since the relative order of individual
operations is preserved. Still, it is difficult to predict how every possible
usecase may be affected.
This affects the dynamic version of `upgrade` and makes it more consistent with
the static version, while removing an artificial limitation.
This commit also refactors the file layout and code, in order to share code wrt
to dowgrading components between the dynamic and static versions.
This makes it more consistent with the dynamic version of `upgrade` and makes it
possible to share code between the dynamic and static versions.
This commit also refactors the file layout, moving common and dynamic-specific
files to `common/` and `dynamic/` directories respectively and renaming `aot/`
to `static/`.
Some private keys, used as AngularJS DI tokens, have also been renamed, but this
should not affect apps, since these keys are undocumented and not supposed to
be used externally.
BREAKING CHANGE:
Previously, `upgrade/static/downgradeInjectable` returned an array of the form:
```js
['dep1', 'dep2', ..., function factory(dep1, dep2, ...) { ... }]
```
Now it returns a function with an `$inject` property:
```js
factory.$inject = ['dep1', 'dep2', ...];
function factory(dep1, dep2, ...) { ... }
```
It shouldn't affect the behavior of apps, since both forms are equally suitable
to be used for registering AngularJS injectable services, but it is possible
that type-checking might fail or that current code breaks if it relies on the
returned value being an array.
`params` has been introduced in 4.0.0-beta.0
Before:
http.get(url, new RequestOptions({params: searchParams}))
After:
http.get(url, {params: searchParams})
Fixes#14100
PR Close#14101
Make sure that context (`this`) that is passed to functions generated by test helpers is passed through to the callback functions. Enables usage of Jasmine's variable sharing system to prevent accidental memory leaks during test runs.
ReflectiveInjector previously used two strategies for resolving dependencies. These
were to support the Dart implementation, but are no longer needed. A result of this
PR is there is no longer a 20 dependency limit and the generated code is smaller.
PR Close#14126
Angular 1.x -> AngularJS
Angular 1 -> AngularJS
Angular1 -> AngularJS
Angular 2+ -> Angular
Angular 2.0 -> Angular
Angular2 -> Angular
I have deliberately not touched any of the symbol names as that would cause big merge collisions with Tobias's work.
All the renames are in .md, .json, and inline comments and jsdocs.
PR Close#14132
Also have a new node type for queries.
This leads to less memory usage and better performance.
Deep Tree Benchmark results (depth 11):
- createAndDestroy (view engine vs current codegen):
* pureScriptTime: 78.80+-4% vs 72.34+-4%
* scriptTime: 78.80+-4% vs 90.71+-9%
* gc: 5371.66+-108% vs 9717.53+-174%
* i.e. faster when gc is also considered and about 2x less memory usage!
- update unchanged
Part of #14013
PR Close#14120
This commit effectively reverts 7e0f02f but for `upgrade/static`
as it was an invalid fix for #6385, that created a more significant
bug, which was that changes were not always being detected.
Angular 1 digests should be run inside the ngZone to ensure
that async changes are detected.
We don't know how to fix#6385 without breaking change detection
at this stage. That issue is triggered by async operations, such as
`setTimeout`, being triggered inside scope watcher functions.
One could argue that watcher functions should be pure and not do
work such as triggering async operations. It is possible that the
original use case could be supported by moving the debounce
logic into the watch listener function, which is only called if the
watched value actually changes.
See #13812
PR Close#14039
Chuck Jazdzewski
Georgios Kalpakas
Miško Hevery
Alex Eagle
Victor Berchet
Tobias Bosch
cexbrayat
evan Liu
Dzmitry Shylovich
Toxicable
Dzmitry Shylovich
Dzmitry Shylovich
WilliamKoza
执衡
Dimitri Benin
Nuno Arruda
Jason Aden
Igor Minar
Gion Kunz
Peter Bacon Darwin
Bowen Ni