Previously, the compiler performed an incremental build by analyzing and
resolving all classes in the program (even unchanged ones) and then using
the dependency graph information to determine which .js files were stale and
needed to be re-emitted. This algorithm produced "correct" rebuilds, but the
cost of re-analyzing the entire program turned out to be higher than
anticipated, especially for component-heavy compilations.
To achieve performant rebuilds, it is necessary to reuse previous analysis
results if possible. Doing this safely requires knowing when prior work is
viable and when it is stale and needs to be re-done.
The new algorithm implemented by this commit is such:
1) Each incremental build starts with knowledge of the last known good
dependency graph and analysis results from the last successful build,
plus of course information about the set of files changed.
2) The previous dependency graph's information is used to determine the
set of source files which have "logically" changed. A source file is
considered logically changed if it or any of its dependencies have
physically changed (on disk) since the last successful compilation. Any
logically unchanged dependencies have their dependency information copied
over to the new dependency graph.
3) During the `TraitCompiler`'s loop to consider all source files in the
program, if a source file is logically unchanged then its previous
analyses are "adopted" (and their 'register' steps are run). If the file
is logically changed, then it is re-analyzed as usual.
4) Then, incremental build proceeds as before, with the new dependency graph
being used to determine the set of files which require re-emitting.
This analysis reuse avoids template parsing operations in many circumstances
and significantly reduces the time it takes ngtsc to rebuild a large
application.
Future work will increase performance even more, by tackling a variety of
other opportunities to reuse or avoid work.
PR Close#34288
With Angular CLI version 9 RC 3 we can run a single migration for a package using the name of the migration schematic.
We need to pass the schematic name as a value to the `migrate-only` option.
Ex:
```
ng update @angular/core --migrate-only migration-v9-undecorated-classes-with-di
```
See: https://github.com/angular/angular-cli/pull/16174
PR Close#33958
In View Engine, providers which neither used `useValue`, `useClass`,
`useFactory` or `useExisting`, were interpreted differently.
e.g.
```
{provide: X} -> {provide: X, useValue: undefined}, // this is how it works in View Engine
{provide: X} -> {provide: X, useClass: X}, // this is how it works in Ivy
```
The missing-injectable migration should migrate such providers to the
explicit `useValue` provider. This ensures that there is no unexpected
behavioral change when updating to v9.
PR Close#33709
The `renderer-to-renderer2` migration currently does not work
properly in v9 because the migration relies on the type checker
for detecting references to `Renderer` from `@angular/core`.
This is contradictory since the `Renderer` is no longer
exported in v9 `@angular/core`. In order to make sure that
the migration still works in v9, and that we can rely on the
type checker for the best possible detection, we take advantage
of module augmentation and in-memory add the `Renderer` export
to the `@angular/core` module.
PR Close#33571
Currently TypeScript projects with an invalid tsconfig configuration,
cause the undecorated-classes-with-di migration to throw. Instead we
should gracefully exit the migration (like we do for syntactical
diagnostics), but report that there are configuration issues.
This issue surfaced when testing this migration in combination
with the Angular CLI migrations. One of the CLI migrations currently
causes invalid tsconfig files which then cause this issue in the
undecorated-classes-with-di migration.
PR Close#33567
Currently if one of the project targets could not be analyzed
due to AOT compiler program failures, we gracefully proceed
with the migration. This is expected, but we should not
print a message at the end of the migration that the migration
was _successful_. The migration was only done partially, hence
it's potentially incomplete and we should make it clear that once
the failures are resolved, the migration should be re-run.
PR Close#33315
With the next version of the CLI we don't need to add logging for the description of the schematic as part of the schematic itself.
This is because now, the CLI will print the description defined in the `migrations.json` file.
See: https://github.com/angular/angular-cli/pull/15951
PR Close#33440
Currently the `missing-injectable` migration seems to add
`@Injectable()` to third-party classes in type definitions.
This not an issue in general since we do not generate broken code
by inserting a decorator into a type definition file. Though, we can
avoid adding the decorator since it won't have any effect and in
general we should not write to non source files of the compilation unit.
PR Close#33286
We should not migrate the reference from `useExisting`. This is because
developers can only use the `useExisting` value as a token. e.g.
```ts
@NgModule({
providers: [
{provide: AppRippleConfig, useValue: rippleOptions},
{provide: MAT_RIPPLE_OPTIONS, useExisting: AppRippleConfig},
]
})
export class AppModule {}
```
In the case above, nothing should be decorated with `@Injectable`. The
`AppRippleConfig` class is just used as a token for injection.
PR Close#33286
Currently the migration is unable to migrate instances where
the provider definition uses `forwardRef`. Since this is a
common pattern, we should support that from within the migration.
The solution to the problem is adding a foreign function resolver
to the `PartialEvaluator`. This basically matches the usage of
the static evaluation that is used by the ngtsc annotations.
PR Close#33286
Static methods that return a type of ModuleWithProviders currently
do not have to specify a type because the generic falls back to any.
This is problematic because the type of the actual module being
returned is not present in the type information.
Since Ivy uses d.ts files exclusively for downstream packages
(rather than metadata.json files, for example), we no longer have
the type of the actual module being created.
For this reason, a generic type should be added for
ModuleWithProviders that specifies the module type. This will be
required for all users in v10, but will only be necessary for
users of Ivy in v9.
PR Close#33217
Angular v9 schematics should print out a link to the migration
guide associated with each schematic. This way, users have an
easy way to find more information about the automatic code
transformations they will see with `ng update`.
PR Close#33258
Currenly the `missing-injectable` migration only migrates providers referenced from
`@NgModule` definitions. The schematic currently does not cover the migration for
providers referenced in `@Directive` or `@Component` definitions.
We need to handle the following keys for directives/components:
- `@Directive` -> `providers`
- `@Component` -> `providers` and `viewProviders`.
This commit ensures that the migration handles providers for these
definitions.
PR Close#33011
Re-enables the dynamic queries migration, now that we have all of the necessary framework changes in place.
Also moves the logic that identifies static queries out of the compiler and into the static queries migration, because that's the only place left that's using it.
PR Close#32992
Current we need to create and override certain compiler host methods in every schematic because schematics use a virtual fs. We this change we extract this logic to a common util.
PR Close#32827
The creation of StaticReflector in createMetadataResolver() is a very expensive operation because it involves numerous module resolutions.
To make matter worse, since the API of the Reflector does not provide the ability to invalidate its internal caches, it has to be destroyed and recreated on *every* program change.
This has a HUGE impact on performance.
This PR fixes this problem by carefully invalidating all StaticSymbols in a file that has changed, thereby reducing the overhead of recomputation on program change.
PR Close#32543
Currently the undecorated-classes-with-di migration leverages NGC in order
to work with metadata resolution. Since NGC by default tries to resolve referenced
resources on initialization of the underlying TS program, it can result in unexpected
migration failures due to missing resource files.
This is especially an issue since the CLI wraps the `AngularCompilerProgram` with
special logic (i.e. to support SCSS preprocessing etc.). We don't have all of this since
we instantiate a vanilla NGC program.
The solution to the problem is to simply treat resource requests as valid, and returning
a fake content. The migration is not dependent on templates or stylesheets.. so it's the
simplest and most robust solution.
Fixes#32826
PR Close#32953
ec4381d enabled Ivy by default. This is problematic as migrations
like `static-queries` depend on the `AngularCompilerProgram` (NGC)
in order to perform the migration from version 7 to version 8.
In order to ensure that the migration always runs with NGC
(and doesn't get the `NgtscProgram`), we need to explicitly disable
ivy when creating the `@angular/compiler-cli` program for the migration.
This code is still relevant even though the update from version 7
to version 8 landed. Developers can run `ng update` from version 7
and immediately get to version 9 where Ivy is enabled by default (and in
that case we need to ensure that ngtsc is not accidentally used).
Similar to
e5636a322c.
PR Close#32954
Commit 904a2018e0 introduced a new migration for
undecorated classes with decorated Angular class members. Currently the migration
always calls `tree.beginUpdate` and `tree.commitUpdate` (even if there are no changes).
This causes unnecessary updates to be reported to developers running `ng update`. Once
an update is commited, the CLI will report the update regardless of whether any changes were
made or not.
This behavior can be observed in the `ng_update_migrations` integration test. See:
https://circleci.com/gh/angular/angular/438470#tests/containers/3. Notice how all
source files are denoted as `UPDATED` (even though there are no changes).
PR Close#32391
Currently the undecorated classes migration decorates base classes if no
explicit constructor is defined on all classes in the inheritance chain.
We only want to decorate base classes which define a constructor that is
inherited. Additionally for best practice, all classes in between the class
that inherits the constructor and the one that defines it are also decorated.
PR Close#32319
The `undecorated-classes-with-di` migration currently creates invalid object literals from parsed
NGC metadata files if there are object literal properties with keys that contain special characters.
e.g. consider a decorated base class with a host binding using `[class.X]`. Currently the migration
parses and converts the metadata to TypeScript code but incorrectly uses `[class.X]` unquoted as
identifier.
PR Close#32319
ec4381dd40 enabled Ivy by default. This is
problematic as migrations like `undecorated-classes-with-di` depend on the
`AngularCompilerProgram` (NGC) in order to perform the migration from
version 8 to version 9. In order to ensure that the migration always runs
with NGC (and doesn't get the `NgtscProgram`), we need to explicitly disable
ivy when creating the `@angular/compiler-cli` program for the migration.
PR Close#32318
Initially the plan was to have a migration that adds `@Injectable()` to
all pipes in a CLI project so that the pipes can be injected in Ivy
similarly to how it worked in view engine.
Due to the planned refactorings which ensure that `@Directive`, `@Component`
and `@Pipe` also have a factory definition, this migration is no longer
needed for Ivy. Additionally since it is already disabled (due to
572b54967c) and we have a more generic
migration (known as `missing-injectable)` that could do the same as
`injectable-pipe`, we remove the migration from the code-base.
PR Close#32184
Introduces a new migration schematic that follows the given
migration plan: https://hackmd.io/@alx/S1XKqMZeS.
First case: The schematic detects decorated directives which
inherit a constructor. The migration ensures that all base
classes until the class with the explicit constructor are
properly decorated with "@Directive()" or "@Component". In
case one of these classes is not decorated, the schematic
adds the abstract "@Directive()" decorator automatically.
Second case: The schematic detects undecorated declarations
and copies the inherited "@Directive()", "@Component" or
"@Pipe" decorator to the undecorated derived class. This
involves non-trivial import rewriting, identifier aliasing
and AOT metadata serializing
(as decorators are not always part of source files)
PR Close#31650
Follow-up to #30993 where we build all Angular packages with
the TypeScript `--strict` flag. The flag improves overall code
health and also helps us catch issues easier.
PR Close#31967
Moves the `renderer_to_renderer2` migration google3 tslint rule
into the new `google3` directory. This is done for consistency
as we recently moved all google3 migration rules into a new
`google3` folder (see: f69e4e6f77).
PR Close#31817
Creates a separate bazel target for the google3 migration
tests. The benefit is that it's faster to run tests for
public migrations in development. Google3 lint rules are
usually another story/implementation and the tests are quite
slow due to how TSLint applies replacements.
Additionally if something changes in the google3 tslint rules,
the tests which aren't affected re-run unnecessarily.
PR Close#31817
Moves all google3 migration tslint rules into a single directory.
This makes it easier to wire up multiple migration rules in
google3 without having to update the rule directories each time
a new migration is available.
PR Close#30956
Introduces a new migration schematic for adding the "@Injectable()"
decorator to provider classes which are currently not migrated. Previously
in ViewEngine, classes which are declared as providers sometimes don't
require the "@Injectable()" decorator
(e.g. https://stackblitz.com/edit/angular-hpo7gw)
With Ivy, provider classes need to be explicitly decorated with
the "@Injectable()" decorator if they are declared as providers
of a given module. This commit introduces a migration schematic
which automatically adds the explicit decorator to places where
the decorator is currently missing.
The migration logic is designed in a CLI devkit and TSlint agnostic
way so that we can also have this migration run as part of a public
CLI migration w/ `ng update`. This will be handled as part of a follow-up to reiterate on console output etc.
Resolves FW-1371
PR Close#30956
Adds a schematic and tslint rule that automatically migrate the consumer from `Renderer` to `Renderer2`. Supports:
* Renaming imports.
* Renaming property and method argument types.
* Casting to `Renderer`.
* Mapping all of the methods from the `Renderer` to `Renderer2`.
Note that some of the `Renderer` methods don't map cleanly between renderers. In these cases the migration adds a helper function at the bottom of the file which ensures that we generate valid code with the same return value as before. E.g. here's what the migration for `createText` looks like.
Before:
```
class SomeComponent {
createAndAddText() {
const node = this._renderer.createText(this._element.nativeElement, 'hello');
node.textContent += ' world';
}
}
```
After:
```
class SomeComponent {
createAndAddText() {
const node = __rendererCreateTextHelper(this._renderer, this._element.nativeElement, 'hello');
node.textContent += ' world';
}
}
function __rendererCreateTextHelper(renderer: any, parent: any, value: any) {
const node = renderer.createText(value);
if (parent) {
renderer.appendChild(parent, node);
}
return node;
}
```
This PR resolves FW-1344.
PR Close#30936
fix(@schematics/angular): TypeScript related migrations should cater for BOM
In the CLI `UpdateRecorder` methods such as `insertLeft`, `remove` etc.. accepts positions which are not offset by a BOM. This is because when a file has a BOM a different recorder will be used https://github.com/angular/angular-cli/blob/master/packages/angular_devkit/schematics/src/tree/recorder.ts#L72 which caters for an addition offset/delta.
The main reason for this is that when a developer is writing a schematic they shouldn't need to compute the offset based if a file has a BOM or not and is handled out of the box.
Example
```ts
recorder.insertLeft(5, 'true');
```
However this is unfortunate in the case if a ts SourceFile is used and one uses `getWidth` and `getStart` method they will already be offset by 1, which at the end it results in a double offset and hence the problem.
Fixes#30713
PR Close#30719
Instead of linking to a markdown file explaining what the migration warnings
are about, we should link to the deprecation guide which now also contains
an entry for that schematic. This makes the deprecation explanations
consistent and more centralized.
PR Close#30702
8479cb4233 updated the static-query migration
to refer to the new guide on AIO. Unfortunately these URLs are currently not
valid as the guide is only available on `next.angular.io` right now. In order to
make the link work permanently (e.g. if we eventually remove the guide in future
major versions), we use the permalink from the `v8` subdomain.
PR Close#30649
Currently if a project has source-files with syntax failures and the migration
has been started on a broken project, we silently migrate *as much as possible*,
but never notify the developer that the project was in a broken state and that
he can re-run the migration after the failures are fixed.
Additionally the template strategy does not need to exit gracefully if it detects
Angular semantic diagnostics on generated files (template type checking). These
diagnostics are not relevant for the query timing analysis.
PR Close#30628
We are removing the prompt for the `static-query` migration and make the
template strategy the migration strategy for the migration. The usage
strategy is good for best-practices, but for now we want to ensure that
the migration is a seamless as possible and that is only achievable my
re-using the same logic that View Engine uses for determining the
timing of a query.
PR Close#30628
Slightly improves the messages for the static-query migration in order
to make the terminal output less verbose but more helpful. Unfortunately
we are limited in what we can print due to the devkit not providing much
utilities for printing good messages from a migration schematic.
PR Close#30458
Currently if something fails in the selected strategy (e.g. AOT failures),
the migration currently accidentally falls back to the test strategy. This
is not helpful as we want to give developers the possibility to re-run
the migration after fixing potential AOT failures.
PR Close#30458
Apparently the devkit logger is not able to properly print
out error objects, so we need to convert them to a string
before in order to make the error visible to the user.
This is not testable without an e2e test that validates the CLI
terminal output.
PR Close#30458
Preserve compatibility with rollup_bundle rule.
Add missing npm dependencies, which are now enforced by the strict_deps plugin in tsc_wrapped
PR Close#30370
Currently we always just set the timing to `false` if we aren't
able to analyze a given call expression or new expression. e.g.
```ts
ngOnInit() {
thirdPartyCallSync(() => this.query.doSomething())
}
```
In that case the `thirdPartyCallSync` function comes from the `node_modules`
and is only defined through types while there is no code for the
actual function logic that can be analyzed. This makes it impossible
to tell whether the given call expression actually causes the specified
arrow function to be executed synchronously or not. In order to be able
to make this better, we now peek into the passed arrow function and
check for a synchronous query usage. If so, we set the query timing to
static and mark it as ambiguous. This ensures that the usage strategy is
less "magical" and more correct with third-party code.
Additionally since functions like `setTimeout` are not analyzable but known
to be asynchronous, there is a hard-coded list of known functions which
shouldn't be marked as ambiguous.
Resolves FW-1214. As planned within https://hackmd.io/hPiLWpPlQ4uynC1luIBdfQ
PR Close#30215
Currently the static-query migration ignores queries declared on getters
or setters as these are not part of a `PropertyDeclaration`. We need to
handle these queries in order to cover all queries within a given project.
The usage strategy is not able to detect timing for queries on accessors,
so we add a TODO and print a message. The template strategy is able
to detect the proper timing for such queries because it's not dependent
on detecting the usage of the query.
Resolves FW-1215
PR Close#30327
The static-query template strategy leverages the AOT compiler
in order to determine the query timing. Unfortunately the AOT
compiler has open bugs that can cause unexpected failures which
make the template strategy unusable in rare cases. These rare
exceptions need to be handled gracefully in order to avoid confusion
and to provide a more smooth migration.
Additionally migration strategy setup failures are now reported with
stack traces as the `ng update` command does not print stack traces.
This makes it easier to reproduce and report migration issues.
PR Close#30269
In an Angular CLI project scenario where projects only reference
top-level source-files through the `tsconfig` `files` option, we currently
do not migrate referenced source-files. This can be fixed checking all
referenced source-files which aren't coming from an external library.
This is similar to how `tslint` determines project source-files.
PR Close#30269
Alex Rickabaugh
Alan Agius
Paul Gschwendtner
Filipe Silva
Michael Prentice
Kara Erickson
Adam Plumer
Kristiyan Kostadinov
Alan
Keen Yee Liau
Alex Eagle