nodejs Buffer as Uint8Array cause bad format

[i2534]

Buffer is extends Uint8Array, so I pass it to Uint8ArrayReader
When Buffer(child) is slice from another Buffer(parent) and offset is not 0,
at zip-reader#seekSignature, line 513, bytes type is Buffer, so after slice's buffer is parent's data....
Buffer's slice only mark offset and length, original is keep,
in it's document: To copy the slice, use`Uint8Array.prototype.slice()

my test code(not complete):
`class Reader implements Readable {

        private readonly buffer: Buffer;

        constructor(buffer: Buffer) {
            this.buffer = buffer;
        }

        public get size(): number {
            return this.buffer.length;
        }

        async init(): Promise<void> {
        }
        async read(index: number, length: number): Promise<Uint8Array> {
            //return Uint8Array.prototype.slice.call(this.buffer, index, index + length);
            return this.buffer.slice(index, index + length);
        }
        close(): void {
        }
    }

    it('Readable', async () => {
        const reader = new Reader(buff);
        const zip = new Zip(reader);
        zip.names().then(n => console.log(n))
    });`

[gildas-lormeau]

This is due to incompatibilities in Node.js, see https://nodejs.org/api/buffer.html#buffer_buffers_and_typedarrays.

Buffer.prototype.slice() creates a view over the existing Buffer without copying.

To circumvent this issue due to Node.js, you have to transform the Buffer object into a Uint8Array object. This kind of issue is why zip.js is not officially supported in Node.js.

For the record, this is a duplicate of #305.