Is it really suitable for array diffing?

[kayahr]

I'm currently searching for a fast algorithm to diff large arrays. Just one level, values can be compared with ===, nothing special. So I tried microdiff like this:

import diff from "microdiff";

// Build source array with values 0..9999
const a: number[] = [];
for (let i = 0; i < 10000; i++) {
    a.push(i);
}

// Clone source array into target array and just add one number at the beginning
const b = a.slice();
b.unshift(-1);

console.log(diff(a, b));

I expected a diff with a CREATE change for index 0 but instead I get a diff with 10000 changes. When I add the number at the end then I get a single CREATE change as expected but when inserting or removing a single value somewhere then all subsequent entries are reported as CHANGED. Is this intended? Am I doing something wrong? I don't expect the shortest possible diff (which I know is expensive to generate) but this result seems pretty much useless.