On Friday, 17 April 2015 at 00:08:51 UTC, Idan Arye wrote:
What the ycmd API accepts is not the visual column but the "byte column" - how many bytes from the beginning of the line the cursor is placed. This means that if the semantic engine opens the file and reads it, it can calculate the byte offset from the line and "byte column" supplied to it - without caring about multibyte characters tab settings.
The lexer used by DCD does track the byte columns of tokens.