This adds an additional indentation level that matches indentation of
previous line in a multiline string. Then Tab iterates between 0, the
start indentation level and the previous line level. This is useful
when writing docstrings in numpy/google formats that use indentation
to convey meaning.

For example:

0    4    8
     This is the first item
           This is a sub item
>   >    >

Then pressing Tab while in the third line will iterate between 0, 4
and 8, starting in 8, then 4, then 0. Compare with the current
implementation which will jump just between 0 and 4, which is not
convenient when the subitem is more than one line. A cursory
google/reddit search will show that this is a problem for some people.

Best regards
--
Carlos