microsoft / microsoft/TypeScript
Proposal: Easier Migration with Types in JavaScript Files
Nadie ha tomado este issue todavía.
- Lenguaje dominante
- Go
- Estrellas
- 111k
- Forks
- 14.3k
- Merge medio
- 2 d 4 h
- PR fusionados (30 d)
- 132
Descripción
This issue is the dual of #22665.
Background
In last week's design meeting, we discussed #22665 and it seemed like we had some positive sentiment around the proposal. Specifically, issues where users hit an expressivity wall or mismatch in .js when migrating to .ts files provide a compelling case.
However, there are certain issues with the proposal:
- It still requires users to make a file extension change from
.jsto.ts, which usually implies some change in build steps. - It provides no easy migration path from other type systems with annotation syntax used in
.jsfiles (see https://github.com/Microsoft/TypeScript-Babel-Starter/issues/1). - It introduces the concept of a "diluted/loose" TypeScript experience which can make things more confusing.
Proposal
We can introduce a new mode that enables TypeScript-specific constructs in .js files. This new mode would allow things like type annotations, interfaces, class property annotations, etc. to be declared in .js files. Users would have to explicitly opt-in to this mode for transitionary purposes, with the end goal of migrating to TypeScript.
Supported constructs
The following constructs from TypeScript would be supported:
- Type annotations (
: Foo) - Interfaces (
interface Foo { /*...*/ }) - Type alias (
type Foo = "hello" | "world") as-style type assertions (xyz as SomeType)- Non-nullable type assertions (
x!.bar) - Definite assignment assertions (
let x!: number) - Enums (
enum E { /*...*/ }) - Import aliases and export aliases:
- `import
Unsupported constructs
The following would not be supported to encourage users to write ECMAScript modules and standard enums that utilize late-binding.
- Namespaces
constenums
Existing typed .js file semantics
A JavaScript file still has the same specialized semantics including understanding things like CommonJS modules, JSDoc type annotations, ES5-style constructor functions, object literals potentially being open-ended, etc. However, in the presence of an analogous TypeScript construct, the TypeScript construct always wins. For example, the following declares x as type number, and potentially issues an error on the JSDoc.
/**
* @type {string}
*/
var x: number;
Drawbacks
This proposal avoids the aforementioned issues with #22665, but has some new drawbacks.
For one, it seems like there's little-to-no advantage to moving from a .js file to a .ts file, even though TypeScript is easier to reason about and make assumptions on (whereas much of our JavaScript semantics are more "best-effort").
The other issue, which is likely much bigger, is that this conflates where JavaScript ends and TypeScript begins. While the intent here is to make migration easier, enabling this functionality could be perceived poorly as an attempt to extend the language inappropriately. While we hope users understand this is not the case, we are extremely open to feedback from the community on this issue.
If this flag was called something like --jsMigrationMode, perhaps we could be explicit on the intent and also signal to users that having this flag on is not ideal.
Existing Precedent
Common usage of JSX and Flow in .js files in the Babel ecosystem (as opposed to using the .jsx file extensions) provides a reasonable precedent for supporting TypeScript constructs in .js files.
Guía de contribución
Primeros pasos
- Lee el issue completo y luego la guía de contribución del proyecto.
- Comenta en el issue que vas a ocuparte — evita que dos personas hagan lo mismo.
- Haz un fork del repositorio y trabaja en una rama.
- Abre un pull request que haga referencia al número del issue.
Línea de trabajo
Comienza leyendo las construcciones compatibles y no compatibles de la propuesta, y después revisa el issue relacionado #22665 y la discusión enlazada de Babel Starter. No se mencionan archivos de implementación, pruebas ni puntos de entrada del compilador, por lo que la primera tarea es establecer el diseño y las áreas afectadas antes de definir qué significaría completar el trabajo.
Escrito por el modelo de indexación a partir del texto del issue.
Evaluación
- Stack tecnológico
- javascript, typescript
- Área
- compilers
- Tipo de issue
- Nueva funcionalidad
- Dificultad
- 5/5
- Tiempo estimado
- Más de una semana
- Estado de actividad
- Estancado
- Claridad
- Bastante claro
- Aptitud para principiantes
- 30/100