Why does splitting on `(?!^)` and `(?<!^)` produce the same answer?

Question

The following two lines of code:

System.out.println(Arrays.toString("test".split("(?<!^)")));
System.out.println(Arrays.toString("test".split("(?!^)")));

each produce the same output:

[t, e, s, t]

I expected the bottom line to produce

[, t, e, s, t]

since it should be willing to split after the ^ and before the t. Can someone point out where my thinking is wrong?

Answer 1

(?!^) matches any position that's not at the start of the string, just as (?<!^). Since the ^ anchor doesn't have any length, it is irrelevant whether you look forward or backwards.

Imagine the string test like this where | denotes the positions between the characters:

|  t  |  e  |  s  |  t  |
^ matches here         ($ matches here)

(?!^) doesn't match at position 0 because the regex engine "sees" the start of string from here when looking forward by 0 characters

(?<!^) doesn't match here either because the regex engine "sees" the start of string from here when looking backwards by 0 characters