UTF-8 characters in source seems to throw off source positioning

[thoni56]

Repeat by creating or using a source file with UTF-8 (non-ascii) character in it.

Probably the multiple bytes that that character is stored in creates an offset which the editor (Emacs) adjusts for, so a variable rename breaks since the validation of the position is off by 'n' depending on how many non-single bytes there are in the source.