Avoid a buffer copy when decoding utf-8 in node 10

[jridgewell]

No description provided.

[evanw]

These aren't equivalent:

var array = new Uint8Array([51, 52, 53, 54, 55])
var slice = array.subarray(1, 4)
console.log({
  a: Buffer.from(slice).toString(),
  b: Buffer.from(slice.buffer).toString(),
})

This prints { a: '456', b: '34567' }.

[jridgewell]

Ah, I didn't think we called this with anything but a full-view Uint8Array. I can update to use byteOffset and byteLength for decoding.

[Zuckjet]

@jridgewell I don't understand why this change will avoid buffer copy, can you give an explanation ?

[jridgewell]

See https://nodejs.org/api/buffer.html#buffer_static_method_buffer_from_arraybuffer_byteoffset_length:

This creates a view of the ArrayBuffer without copying the underlying memory. For example, when passed a reference to the .buffer property of a TypedArray instance, the newly created Buffer will share the same allocated memory as the TypedArray's underlying ArrayBuffer.

Passing an ArrayBuffer means the new Buffer instance will be backed by the same memory. Passing an array-like forces a the Buffer to allocate a contiguous memory block, and it then has to iterate over the array to copy the data to the new block.