Support string indexing to yield substrings

[xiaq]

If s=hello, s[0] is h, s[-1] is o, and s[1:4] is ell.

For strings containing multi-byte codepoints, we can copy Julia, i.e.

• Indexing yields codepoints instead of bytes.
• Indices are byte indices.

For instance, if s=你好 (Chinese codepoints occupy 3 bytes in UTF-8), s[0] is 你, s[3] is 好. Indexing at other indices throws an error.

We need to give up the current behavior of string indexing. Right now s[a b] is equivalent to (s a b). I suppose that one syntax for command substitution is enough.

[xiaq]

In addition to indexing, iteration is supported as well, meaning that splicing is also supported:

s=你好
put $@s | each echo
# 你
# 好