ESLint Study Portal

How it works

The mental model is two layers, not one. Layer one is consistency — var vs const, quote style, semicolons, import order. There is no judgment here, only a house style, so you delegate it entirely to eslint --fix and never argue about it in review. Layer two is correctness and risk — no-explicit-any, no-floating-promises, no-unused-vars, complexity. A machine can flag these but can't resolve them, because the fix depends on intent: the right type, the missing await, the refactor.

A strict config is the contract that says "every rule is error, but layer one is auto-corrected on save and on commit." That is what makes a zero-warning policy humane instead of punishing — the mechanical noise is gone before a human ever sees it, so the only thing left in the lint output is signal that genuinely needs a decision. Adoption follows the same split: turn the rules on as warn first, gate new code with --max-warnings 0, then promote to error once the back-catalogue is clean.

Auto-Fix Examples

What eslint --fix Can Fix

// Before
var x = 1;
const y = "hello";
if(x == 1){ console.log(y) }
const fn = function() { return 1; };

// After --fix
const x = 1;
const y = 'hello';
if (x === 1) {
  console.log(y);
}
const fn = () => 1;

Rules with Auto-Fix

Rules That Cannot Auto-Fix

• no-explicit-any - Requires manual typing
• no-unused-vars - Requires removing or using
• complexity - Requires refactoring
• no-floating-promises - Requires await/catch/void