I was experimenting with the python3 interface and I stumbled upon the
following phenomenon. Suppose the cursor is in the string
in a utf-8 encoded file. I run the following commands:
:py3 import vim
:py3 vim.current.line = vim.current.line
Surprisingly last 16½ of 33 characters disappear.
The same effect can be produced e. g. with
:py3 l = vim.current.line
:py3 vim.current.line = l
However, the string l can be correctly printed and sent to vim via
vim.command and setline. Hence I conclude that the transition from vim
to python does not cause any loss of information. On the contrary,
when the string is being passed from python to vim, vim (or python?)
calculates that it is 33 ‘units’ long and copies 33 *bytes*.
I’m running vim 7.3 on Windows 7.
I’d like to find out whether this is a common effect and whether there
exists an elegant way of changing multibyte strings from python3.