Unexpected behaviour with null-characters in string using '\0' notation

[RickardCardell]

Bug Report

Not sure what the reason is. It seems it handles the token after '\0' a bit odd. However, the '\x00 notation seems to work as expected.

1. Minimal reproduce step (Required)

    assert 'aa' == ("a" + "a")       # works
    assert '\0\0' == ("\0" + "\0")  # fails!

    assert '\0\0' == ("\0" + "0")  # doesn't fail!
    assert '\0A' == '\0' # doesn't fail!  Where did 'A' go?

2. What did you expect to see? (Required)

That each character with the '\0' notation is handled as a separate character. As in e.g. Python.
I.e. '\0\0' is two chars of the same code point.
'\0A' should be the same as "\0"+"A", which it isn't currently.

Python

> [ ord(c) for c in "\0\0"]
[0, 0]

KCL

> [ ord(c) for c in "\0\0"]
[0, 48]

3. What did you see instead (Required)

That '\0\0' equals to "\00".
"\0A" equals to "\0"

4. What is your KCL components version? (Required)

> kcl --version
kcl version 0.11.0

Same with 0.11.1

system:

> uname -a
Darwin 24.3.0 Darwin Kernel Version 24.3.0 ... arm64

[RickardCardell]

Additional test cases to showcase the issue.
All these tests fails currently.

test_null0 = lambda {
    assert 'aa' == ("a" + "a") # Works as expected
    assert '\0\0' == ("\0" + "\0")
}

test_null1 = lambda {
    n = '\0\0'
    res = [ord(i) for i in n]
    assert res == [0, 0]
    # actual [0, 48]
}


test_null2 = lambda {
    n = '\0\0\0\0'
    res = [ord(i) for i in n]
    assert res == [0, 0, 0, 0]
    # actual [0, 48, 0, 48]
}

test_null3 = lambda {
    assert "\0A" != "\0"
}

test_null4 = lambda {
    n = 'A\0A'
    res = [ord(i) for i in n]
    assert res == [65, 0, 65]
    #actual [65, 0]
}

test_null5 = lambda {
    n = '\t\0\t'
    res = [ord(i) for i in n]
    assert res == [9, 0, 9]
    #actual [9, 0, 116]
}