Commit Graph

5988 Commits

Author SHA1 Message Date
Alan Agius 508c555ce2 test: update language service module resolution cache (#36989)
With this change we update the expect of the `module resolution cache` were in the second count with cache the `fileExists` is called less then 700 times.

PR Close #36989
2020-05-14 10:50:30 -07:00
Alan Agius f97d8a9cbd refactor: `EventEmitter` to retain behaviour of pre TypeScript 3.9 (#36989)
TypeScript 3.9 introduced a breaking change where extends `any` no longer acts as `any`, instead it acts as `unknown`.

With this change we retain the behavior we had with TS 3.8 which is;

When using the `EventEmitter` as a type you must always provide a  type;
```ts
let emitter: EventEmitter<string>
```

and when initializing the `EventEmitter` class you can either provide a  type or or use the fallback type which is `any`

```ts
const emitter = new EventEmitter(); // EventEmitter<any>
const emitter = new EventEmitte<string>(); // EventEmitter<string>
``

PR Close #36989
2020-05-14 10:50:30 -07:00
Alan Agius bbd556ee9f test: enable importHelpers for UMD builds (#36989)
This is a workaround for a TS 3.9 regression https://github.com/microsoft/TypeScript/issues/38501 where the emitted `__exportStar` helpers have a missing semi-colon at the end of the unnamed function, when targetting UMD, and causes the following runtime error `Uncaught TypeError: (intermediate value)(…) is not a function`.

This is because the anonymous `__exportStar` function will be invoked with the function on the next like as the parameter which is subsequently invoking whatever was returned.

To get around this TS bug, add `importHelpers: true` in your tsconfig. This also, is recommanded to avoid multiple copies of the same helper being inlined, which might cause increase in bundle size.

PR Close #36989
2020-05-14 10:50:30 -07:00
Pete Bacon Darwin 4e1b5e43fa fix(compiler-cli): compute the correct target output for `$localize` messages (#36989)
In some versions of TypeScript, the transformation of synthetic
`$localize` tagged template literals is broken.
See https://github.com/microsoft/TypeScript/issues/38485

We now compute what the expected final output target of the
compilation will be so that we can generate ES5 compliant
`$localize` calls instead of relying upon TS to do the downleveling
for us.

This is a workaround for the TS compiler bug, which could be removed
when this is fixed. But since it only affects ES5 targeted compilations,
which is now not the norm, it has limited impact on the majority of
Angular projects. So this fix can probably be left in indefinitely.

PR Close #36989
2020-05-14 10:50:30 -07:00
Pete Bacon Darwin 91092f668e fix(ngcc): support `defineProperty()` re-exports in CommonJS and UMD (#36989)
In TypeScript 3.9 some re-export syntaxes have changed to be getter
functions (created by calls to `Object.defineProperty()`) rather than
simple property accessors.

This commit adds support into the CommonJS and UMD reflection hosts
for this style of re-export syntax.

PR Close #36989
2020-05-14 10:50:29 -07:00
Pete Bacon Darwin d268d2ad85 fix(ngcc): `viaModule` should be `null` for local imports (#36989)
In the CommonJS and UMD reflection hosts, the logic for computing the
`viaModule` property of `Declaration` objects was not correct for some
cases when getting the exports of modules.

In these cases it was setting `viaModule` to the path of the local module
rather than `null`.

PR Close #36989
2020-05-14 10:50:29 -07:00
Pete Bacon Darwin 2486db7c2b docs(ngcc): tidy up typos etc in comments (#36989)
This file had a few small typos and other issues that have
now been fixed in this commit.............................

PR Close #36989
2020-05-14 10:50:29 -07:00
Pete Bacon Darwin 0672a0e547 refactor(ngcc): rename `ReexportStatement` to `WildcardReexportStatement` (#36989)
The term `ReexportStatement` is too general for this particular concept.
Here the re-export actually refers to a wildcard where all the module
exports are being re-exported.

When we introduce other re-export statement types later this will be
confusing.

PR Close #36989
2020-05-14 10:50:29 -07:00
Pete Bacon Darwin 9846b19986 test(ngcc): reformat some subject code for tests (#36989)
Using backtick multiline strings leads to confusing layout
that does not fit with the surrounding indentation. Also it
can lead to test fragility due to automated code formatting.

This commit changes just one set of subject code to use
a more resilient string concatenation approach.

PR Close #36989
2020-05-14 10:50:29 -07:00
Pete Bacon Darwin 8b79075497 test(compiler-cli): hande TS 3.9 format in emisson tests (#36989)
In TypeScript 3.9 the emitted JS code has some differences.
This commit updates the tests to be resilient to these changes.

PR Close #36989
2020-05-14 10:50:29 -07:00
Pete Bacon Darwin 2956f21d82 test(ngcc): move specs to correct describe block (#36989)
The recent tests for Enum handling were added to the
incorrect describe block. This commit moves them back
to the correct block.

PR Close #36989
2020-05-14 10:50:29 -07:00
Pete Bacon Darwin c8ee390d23 fix(ngcc): ensure rendering formatters work with IIFE wrapped classes (#36989)
After the refactoring of the reflection hosts to accommodate
ES2015 classes wrapped in IIFEs. The same treatment needs to
be applied to the rendering formatters.

PR Close #36989
2020-05-14 10:50:29 -07:00
Pete Bacon Darwin d7440c452a fix(ngcc): ensure reflection hosts can handle TS 3.9 IIFE wrapped classes (#36989)
In TS 3.9, ES2015 output can contain ES classes that are wrapped in an
IIFE. So now ES2015 class declarations can look like one of:

```
class OuterClass1 {}
```

```
let OuterClass = class InnerClass {};
```

```
var AliasClass;
let OuterClass = AliasClass = class InnerClass {};
```

```
let OuterClass = (() => class InnerClass {}};
```

```
var AliasClass;
let OuterClass = AliasClass = (() => class InnerClass {})();
```

```
let OuterClass = (() => {
  let AdjacentClass = class InnerClass {};
  // ... static properties or decorators attached to `AdjacentClass`
  return AdjacentClass;
})();
```

```
var AliasClass;
let OuterClass = AliasClass = (() => {
  let AdjacentClass = class InnerClass {};
  // ... static properties or decorators attached to `AdjacentClass`
  return AdjacentClass;
})();
```

The `Esm5ReflectionHost` already handles slightly different IIFE wrappers
around function-based classes. This can be substantially reused when
fixing `Esm2015ReflectionHost`, since there is a lot of commonality
between the two.

This commit moves code from the `Esm5ReflectionHost` into the `Esm2015ReflectionHost`
and looks to share as much as possible between the two hosts.

PR Close #36989
2020-05-14 10:50:29 -07:00
Pete Bacon Darwin a2b8dc1cfb refactor(ngcc): ensure unlocker process works in mock file-systems (#36989)
Previously the path to the unlocker process was being resolved by the
current file-system. In the case that this was a `MockFileSystemWindows`
on a non-Windows operating system, this resulted in an incorrect path
to the entry-point.

Now the path to the entry-point is hand-crafted to avoid being broken by
whatever FileSystem is in use.

PR Close #36989
2020-05-14 10:50:28 -07:00
Pete Bacon Darwin 7ebf35c5f8 refactor(ngcc): remove unused import (#36989)
The `Import` import from `src/ngtsc/reflection` is not being used.
This commit simply removes this import from the code.

PR Close #36989
2020-05-14 10:50:28 -07:00
Pete Bacon Darwin dbcaf22805 refactor(compiler-cli): delegate `hasBaseClass()` to `getBaseClassExpression()` (#36989)
The previous implementations of `hasBaseClass()` are almost
identical to the implementation of `getBaseClassExpression()`.
There is little benefit in duplicating this code so this refactoring
changes `hasBaseClass()` to just call `getBaseClassExpression()`.
This allows the various hosts that implement this to be simplified.

PR Close #36989
2020-05-14 10:50:28 -07:00
Pete Bacon Darwin 0066a1ae70 refactor(compiler-cli): simplify and clarify `TypeScriptReflectionHost.isClass()` (#36989)
The comment in this function confused me, so I updated it to clarify that
`isClass()` is not true for un-named classes.

Also, I took the opportunity to use a helper method to simplify the function
itself.

PR Close #36989
2020-05-14 10:50:28 -07:00
Pete Bacon Darwin 491da99abe refactor(ngcc): simplify the `detectKnownDeclaration()` signature (#36989)
A number of overloads were added to `detectKnownDeclaration()` to
allow it to support `null` being passed through. In practice this could
easily be avoided, which allows the overloads to be removed and the
method signature and implementations to be simplified.

PR Close #36989
2020-05-14 10:50:28 -07:00
Alan Agius 13ba84731f build: prepare for TypeScript 3.9 (#36989)
- Fix several compilation errors
- Update @microsoft/api-extractor to be compatible with TypeScript 3.9

PR Close #36989
2020-05-14 10:50:28 -07:00
Keen Yee Liau b58bd2bb91 test(language-service): external template update should not invalidate program (#36989)
In TypeScript 3.9, the compiler is able to re-use (i.e. not invalidate)
the previous program if only external templates (i.e. no TS files) have
changed.

PR Close #36989
2020-05-14 10:50:27 -07:00
Keen Yee Liau a32cbed89c fix(language-service): use empty statement as parent of type node (#36989)
In TypeScript 3.9, type nodes need to exist in the context of a statement.
This commit ensures that the synthetic type node has such a parent.

PR Close #36989
2020-05-14 10:50:27 -07:00
Shadi Namrouti 90205079d7 docs(common): English grammar (#36908)
In English grammar, the comma followed by a conjunction is considered incorrect. This kind of mistakes called comma splice.
See: https://www.grammarly.com/blog/comma-splice/
PR Close #36908
2020-05-13 16:03:25 -07:00
Andrew Scott 047642556c refactor(core): remove ActiveIndexFlag from LContainer (#37073)
The ActiveIndexFlag is no longer needed because we no longer have "inline embedded views".
There is only one type of embedded view so we do not need complex tracking for
inline embedded views.

HAS_TRANSPLANTED_VIEWS now takes the place of the ACTIVE_INDEX slot as a
simple boolean rather than being a shifted flag inside the ACTIVE_INDEX bits.

PR Close #37073
2020-05-13 16:00:41 -07:00
Adam Plumer 9edea0bb75 feat(platform-server): use absolute URLs from Location for HTTP (#37071)
Currently, requests from the server that do not use absolute URLs
fail because the server does not have the same fallback mechanism
that browser XHR does. This adds that mechanism by pulling the
full URL out of the document.location object, if available.

PR Close #37071
2020-05-13 16:00:02 -07:00
Matias Niemelä 45f4a47286 refactor(core): remove style sanitization code for `[style]`/`[style.prop]` bindings (#36965)
In 420b9be1c1 all style-based sanitization code was
disabled because modern browsers no longer allow for javascript expressions within
CSS. This patch is a follow-up patch which removes all traces of style sanitization
code (both instructions and runtime logic) for the `[style]` and `[style.prop]` bindings.

PR Close #36965
2020-05-13 15:56:12 -07:00
Pete Bacon Darwin 141fcb95a4 refactor(localize): use the `FileSystem` from ngtsc (#36843)
This commit makes the leap from its own custom baked `FileUtils`
solution to the fully formed `FileSystem` that is used in the compiler-cli.

This makes testing more straightforward and helps to ensure that the tool
will work across operatings systems.

Also, going forward, it will allow the localize project access to other useful
code from the compiler-cli, such as source-map handling.

PR Close #36843
2020-05-13 15:52:48 -07:00
Pete Bacon Darwin dbf1f6b233 refactor(compiler-cli): expose `loadTestDirectory()` test helper (#36843)
This helper can be useful in other packages to load files from the
real disk into a mock file system.

PR Close #36843
2020-05-13 15:52:48 -07:00
Pete Bacon Darwin 5d12c19ce9 refactor(compiler-cli): support Buffer files in `FileSystem` (#36843)
Adding `readFileBuffer()` method and allowing `writeFile()` to accept a
Buffer object will be useful when reading and writing non-text files,
such as is done in the `@angular/localize` package.

PR Close #36843
2020-05-13 15:52:48 -07:00
cexbrayat 4f3ac1d983 docs(common): fix CLDR website link in FormStyle doc (#37069)
The `FormStyle` enum offers two options, and the explanation of the difference between the two can be found on the CLDR official website. Sadly, the link changed and the one currently referenced is a dead-end. This commit fixes the link.

PR Close #37069
2020-05-12 14:41:20 -07:00
Kara Erickson d7c70caa4c docs(platform-webworker): remove mention of version 10 (#37052)
The deprecation notice for platform-webworker
APIs is too prescriptive and notes that we will
remove the package in version 10. Since we are
not planning to do this for version 10, this
commit updates the notice to read "a future
version of Angular".

PR Close #37052
2020-05-12 14:40:56 -07:00
Misko Hevery cb6ddfc215 Revert "fix(common): `locales/global/*.js` are not ES5 compliant (#36342)" (#37074)
This reverts commit 078b0be4dc.

The original commit was a work around for a bug in CLI. That bug was fixed in the CLI, as a result this change is no longer needed and is being reverted.

PR Close #37074
2020-05-12 13:11:57 -07:00
Jan-Niklas W 198fc6f108 docs(forms): remove rxjs of alias import in form async validator example (#36856)
In the code example of the AsyncValidator example there was an aliased
import for the rxjs operator `of`. To align with the RxJS docs it should
just use a plain import of `of`

PR Close #36856
2020-05-12 10:50:39 -07:00
Michiel Kikkert 8d8e419664 fix(core): correct "development mode" console message (#36571)
The message can be improved by removing the unneeded ‘the’ (x2).

Before:
Angular is running in the development mode. Call enableProdMode() to enable the production mode.

After:
Angular is running in development mode. Call enableProdMode() to enable production mode.

Closes #36570

PR Close #36571
2020-05-12 10:50:12 -07:00
Keen Yee Liau 40025f55ad refactor(language-service): move TS utils to separate file (#36984)
This commit refactors TS-only utility functions to a separate file so
that they could be shared with Ivy language service.
A separate ts_library rule is created so that there is no dependency on
`compiler` and `compiler-cli` to make the compilation fast and
light-weight.
The method `getPropertyAssignmentFromValue` is modified slightly to
improve the ergonomics of the function.

PR Close #36984
2020-05-11 11:07:50 -07:00
Andrew Kushnir 28f3c1c96a refactor(core): rename `instructions/container.ts` -> `instructions/template.ts` (#34715)
Prior to this change, the `template` instruction logic was located in the `instructions/container.ts` file alongside embedded view instructions. Since unused embedded view instructions are removed in a previous commit, this commit renames `container.ts` -> `template.ts`, since only template-related instructions were retained.

PR Close #34715
2020-05-11 10:52:28 -07:00
Andrew Kushnir 0c8adbc4ec refactor(core): remove unused embedded view instructions (#34715)
This commit performs a cleanup and removes unused embedded view instructions and corresponding tests.

PR Close #34715
2020-05-11 10:52:28 -07:00
Ayaz Hafiz eb34aa551a feat(compiler): add name spans for property reads and method calls (#36826)
ASTs for property read and method calls contain information about
the entire span of the expression, including its receiver. Use cases
like a language service and compile error messages may be more
interested in the span of the direct identifier for which the
expression is constructed (i.e. an accessed property). To support this,
this commit adds a `nameSpan` property on

- `PropertyRead`s
- `SafePropertyRead`s
- `PropertyWrite`s
- `MethodCall`s
- `SafeMethodCall`s

The `nameSpan` property already existed for `BindingPipe`s.

This commit also updates usages of these expressions' `sourceSpan`s in
Ngtsc and the langauge service to use `nameSpan`s where appropriate.

PR Close #36826
2020-05-08 14:42:42 -07:00
Keen Yee Liau 1142c378fd feat(language-service): [ivy] wrap ngtsc to handle typecheck files (#36930)
This commit adds a Compiler interface that wraps the actual ngtsc
compiler. The language-service specific compiler manages multiple
typecheck files using the Project interface, creating and adding
ScriptInfos as necessary.

This commit also adds `overrideInlineTemplate()` method to the mock
service so that we could test the Compiler diagnostics feature.

PR Close #36930
2020-05-08 14:42:11 -07:00
Lars Gyrup Brink Nielsen 14be55c9fa docs(common): remove mention of the deprecated (#36953)
web worker platform. Minor grammar/stylistic changes.

Inline documentation for the PlatformLocation service instead mentions @angular/platform-server.

Typos corrected, minor grammar and stylistic changes.

PR Close #36953
2020-05-08 09:40:35 -07:00
Alex Rickabaugh 42d1091d6a fix(compiler-cli): don't try to tag non-ts files as shims (#36987)
Some projects include .js source files (via the TypeScript allowJs option).
Previously, the compiler would attempt to tag these files for shims, which
caused errors as the regex used to create shim filenames assumes a .ts file.
This commit fixes the bug by filtering out non-ts files during tagging.

PR Close #36987
2020-05-07 14:45:05 -07:00
Paul Gschwendtner d0280a0335 fix(compiler): remove outdated and invalid warning for unresolved DI parameters (#36985)
In past versions of the View Engine compiler, we added a warning that is
printed whenever the compiler comes across an Angular declaration with a
constructor that does not match suitable DI tokens. The warning mentioned
that in `v6.x` it will turn into an actual error.

This actually happened as expected for most cases. e.g. the constructor
of `@NgModule`, `@Component`'s, `@Pipe`'s etc will be checked and an error
will be reported if constructor is not DI compatible.

The warning has never been removed though as it was still relevant for
unprovided injectables, or injectables serialized into summaries of the
Angular compiler.

As of version 10, classes that use Angular features need an Angular decorator.
This includes base classes of services that use the lifecycles Angular feature.
Due to this being a common pattern now, we can remove the warning in
View Engine. The warning is not correct, and also quite confusing as it
mentions the planned removal in `v6.x`.

Resolves FW-2147.

PR Close #36985
2020-05-07 14:44:42 -07:00
Tidi 97d6d909e7 fix(forms): number input fires valueChanges twice. (#36087)
Prior to this commit, number input fields would to fire valueChanges twice: once for `input` events when typing and second for the `change` event when the field lost focus (both events happen at once when using the increment and decrement buttons on the number field).

Fixes #12540

BREAKING CHANGE: Number inputs no longer listen to the `change` event.
* Tests which trigger `change` events need to be updated to trigger `input` events instead.
* The `change` event was in place to support IE9, as we found that `input` events were not fired with backspace or cut actions. If you need to maintain IE9 support, you will need to add a change event listener to number inputs and call the `onChange` method of `NumberValueAccessor` manually.
* Lastly, old versions of WebDriver would synthetically trigger the `change` event on `WebElement.clear` and `WebElement.sendKeys`. If you are using an old version of WebDriver, you may need to update tests to ensure `input` events are triggered. For example, you could use `element.sendKeys(Keys.chord(Keys.CONTROL, "a"), Keys.BACK_SPACE);` in place of `element.clear()`.

PR Close #12540

PR Close #36087
2020-05-07 10:20:26 -07:00
Pete Bacon Darwin f7815cf96d test(compiler-cli): ensure reflection tests are not brittle to case-sensitivity (#36859)
These tests were matching file-paths against what is retrieved from the
TS compiler. But the TS compiler paths have been canonicalised, so the
tests were brittle on case-insensitive file-systems.

PR Close #36859
2020-05-06 15:23:16 -07:00
Pete Bacon Darwin 9e43e4900e test(compiler-cli): ensure partial-evaluator tests are not brittle to case-sensitivity (#36859)
These tests were matching file-paths against what is retrieved from the
TS compiler. But the TS compiler paths have been canonicalised, so the
tests were brittle on case-insensitive file-systems.

PR Close #36859
2020-05-06 15:23:16 -07:00
Pete Bacon Darwin 8ce38cac0d test(compiler-cli): ensure indexer tests are not brittle to case-sensitivity (#36859)
These tests were matching file-paths against what is retrieved from the
TS compiler. But the TS compiler paths have been canonicalised, so the
tests were brittle on case-insensitive file-systems.

PR Close #36859
2020-05-06 15:23:16 -07:00
Pete Bacon Darwin a10c126692 fix(compiler-cli): use CompilerHost to ensure canonical file paths (#36859)
The type checking infrastrure uses file-paths that may come from the
TS compiler. Such paths will have been canonicalized, and so the type
checking classes must also canonicalize paths when matching.

PR Close #36859
2020-05-06 15:23:16 -07:00
Pete Bacon Darwin b682bd1916 fix(compiler-cli): normalize mock Windows file paths correctly (#36859)
Since the `MockFileSystemWindows` is case-insensitive, any
drive path that must be added to a normalized path should be lower
case to make the path canonical.

PR Close #36859
2020-05-06 15:23:16 -07:00
Pete Bacon Darwin 26eacd4fcb fix(compiler-cli): ensure `MockFileSystem` handles case-sensitivity (#36859)
Previously this class used the file passed in directly to look up files in the
in-memory mock file-system. But this doesn't match the behaviour of
case-insensitive file-systems. Now the look up is done on the canonical
file paths.

PR Close #36859
2020-05-06 15:23:16 -07:00
Pete Bacon Darwin fc4741f638 fix(compiler-cli): `isCaseSensitive()` returns correct value (#36859)
Previously this method was returning the exact opposite value
than the correct one.
Also, calling `this.exists()` causes an infinite recursions,
so the actual file-system `fs.existsSync()` method is used
to ascertain the case-sensitivity of the file-system.

PR Close #36859
2020-05-06 15:23:16 -07:00
Pete Bacon Darwin 3f3e9b7555 fix(compiler-cli): ensure `getRootDirs()` handles case-sensitivity (#36859)
Previously the `getRootDirs()` function was not converting
the root directory paths to their canonical form, which can
cause problems on case-insensitive file-systems.

PR Close #36859
2020-05-06 15:23:16 -07:00