# unicorn/prefer-math-trunc

> Require `Math.trunc()` instead of bitwise tricks to drop the fraction of a number

`x | 0` and `~~x` drop the fraction of a number as a side effect of converting it to a 32-bit integer. The result is wrong for large numbers. For example, `2147483648.5 | 0` is `-2147483648`. `Math.trunc()` says what it does and works for every number.

The rule reports `~~x`, and `|`, `^`, `>>` and `<<` with `0` on the right, including the assignment forms such as `x |= 0`.

{/* Generated by src/codegen/lint-rule-docs.ts. Do not edit this section. */}

## Configure

Off by default. To turn it on:

```toml bunfig.toml
[lint.rules]
"unicorn/prefer-math-trunc" = "error" # or "warn"
```

{/* End of the generated section. */}

## Examples

Reported:

{/* bad {"count": 3} */}
```js
const seconds = (elapsed / 1000) | 0;
const page = ~~(index / pageSize);
total >>= 0;
```

Not reported:

{/* good */}
```js
const seconds = Math.trunc(elapsed / 1000);
const page = Math.trunc(index / pageSize);
total = Math.trunc(total);
```

`>>> 0` isn't reported, because it converts to an unsigned integer:

{/* good */}
```js
const unsigned = hash >>> 0;
```

## When not to use it

Hash functions and similar code use `| 0` on purpose to wrap a value to 32 bits. `Math.trunc()` doesn't wrap. It also keeps `NaN` and `Infinity`, which `| 0` turns into `0`.