IndexError when current line contains multibyte character?

[timfeirg]

when I try to get completion for function get_tot_click, in a line that contains chinese characters, the following error will arise:

IndexError: string index out of range

and if I move the cursor out of that line, it'll work:

---

Am I posting a duplicate issue? I've read #278, #590, #788, this issue is about utf-8 characters causing autocomplete failure for ASCII names.

[simnalamburt]

👍

[vheon]

@timfeirg can you post a little test case? Images are bad for copy and paste :P

[simnalamburt]

@vheon Try below Nevermind, in my case ^C^Ving the texts doesn't reproduce error (It reproduced when I write it manually)

First you should write somthing
And then multibyte char will cause error
이걸복사해서붙여넣어보시오

[timfeirg]

yes, sorry, the test case:

• create test.py
• copy & paste the following python code (using set paste and set nopaste after):

# -*- coding: utf-8 -*-
def one_special_function():
    return 23

a = {}
a['中文'] = 
a['ascii'] = 

• now, I open the above file in vim, try to call our special function in in the 5th row and assign the returned value to a['中文'], YCM didn't offer any suggestions when I entered onespe, I keep typing and entered onespeci, the function name appears as the last candidate (preceded by a whole bunch of mismatch) in YCM autocomplete window.
• and I try to do the same with the 7th row: I type one and one_special_function pops up as the first candidate.

And for @simnalamburt 's information, In my case copy & paste would reproduce such error, I wouldn't know if it's set paste that cause vim or YCM to treat the text differently.

[Valloric]

@timfeirg From your test case, I can repro the problem of getting a whole bunch of mis-matched completions, but I don't get an IndexError.

WRT copy/pasting and IndexError, this might be related to whatever is set as your current file encoding in Vim.

[timfeirg]

I've add the shebang line in my test case, it's utf-8, I think if you do the same you'll reproduce the problem.

So it is because of my file encoding, any suggestions what I can do to make YCM work?

[oblitum]

I'm getting IndexError and others related to unicode characters in buffer too, this should have been introduced somewhat recently since I only noticed this after bumping my fork after some weeks staled.

[timokau]

I'm having the same problems with the german Symbols ä, ü, ß

[wsdjeg]

hi @timfeirg I have test you case ,has no error,also i have test my Jave code, also no error

[timfeirg]

I'm on 0352ed9 and still can reproduce. @wsdjeg can you post your YCM version as well as your vim dot-file?

[wsdjeg]

i just updated to the latest version of YCM,and here is my ycm . you just need to have a look at part of autocomplete

[wsdjeg]

https://github.com/wsdjeg/DotFiles/blob/develop/vimrc

[timfeirg]

thanks @wsdjeg , your vimrc has been educating, but I'm having trouble identifying the lines that fix this particular problem.

I've tried copying the lines that touched YCM config & apply them to my own vimrc, didn't work, and it seems that they weren't related to character encoding anyway.

in case you'd like to take a look, here's my vimrc

[wsdjeg]

i got the same issue

[wsdjeg]

if i delete the chinese words it works well

[wsdjeg]

but if only one chinese char,it also works well