为什么在编译Typescript文件后应该执行TSlint？

Question

According to the official TSLint, TSLint should executed only after the Typescript code is compiled. What is the reason behind this?

The official TSLint doc: https://palantir.github.io/tslint/usage/cli/

Answer 1

TSLint runs against a parsed AST (Abstract Syntax Tree) that it receives from TypeScript. If the AST it receives from TypeScript has errors, its rules' logic might make some incorrect assumptions.

Defining a few of those terms:

• AST: Tree-like representation of your code that the linter will analyze. You can play around with astexplorer.net to see what this looks like; it's pretty nifty! 😊
• Syntax errors: if you write code such as led foo = "bar"; ( led instead of let ) . TypeScript will create a different representation of your code than you might expect.
• Semantic errors: if you write code that has no syntax errors but the type checker finds issues with, such as let foo: number = "bar"; .

TSLint has an issue open to discuss whether it shouldn't bother checking code that has syntax or semantic errors: https://github.com/palantir/tslint/issues/3808

Two additional points in favor of not running TSLint on invalid source files:

• A lot of people have trouble differentiating between TS and TSLint errors, so it can be confusing when an invalid file results in a lot of lint errors
• TS code that doesn't compile tends to have different linting issues after it gets fixed, so it's a bit of a waste to ask people to fix them along with the TS errors