wrong docstring output encoding

[DonJayamanne]

From Igor Novozhilov (@IgorNovozhilov) on March 23, 2017 20:49

Environment data

VS Code version: 1.11.0-insider
Python Extension version: 0.6.0
Python Version: 3.3.5
OS and version: win7-pro

Actual behavior

Expected behavior

Steps to reproduce:

• write "ru" docstring
• hovered methods

It's a heisenbug... The text with the wrong encoding appears not on every hover. He occasionally appears when auto-save file, even when editing parcels not related to the method or docstring.

A similar problem: https://github.com/DonJayamanne/pythonVSCode/issues/372

Settings

Your settings.json: no

Logs

no

Output from Console window (Help->Developer Tools menu)
no

Copied from original issue: DonJayamanne/pythonVSCode#844

[DonJayamanne]

From Igor Novozhilov (@IgorNovozhilov) on March 23, 2017 20:58

I found 🐛. Wrong encoding appears only once executed linting file and highlight warnings.

[DonJayamanne]

Please could you provide some sample code and exact steps to replicate this issue.
I'm having trouble replicating this at my end. Thanks for your help.

[DonJayamanne]

From Igor Novozhilov (@IgorNovozhilov) on April 6, 2017 18:5

Occurs only on windows with RU-ru localization, and utf-8 py-file encoding.

""" Модуль """

def test():
    """ Описание функции """
    pass

test()

[DonJayamanne]

From Marcus Schweda (@mdschweda) on July 18, 2017 19:8

Same for any Umlaut (aöü)

# -*- coding: utf-8 -*-
"""Dieses Modul enthält.... <---"""

class Client(AuthenticatedClient):

    async def schema(self, entity_type: str, single: bool = False) -> object:
        """Fragt das Schema eines Typens von Eintrag ab.

        Parameters
        ----------
        entity_type : str
            Der Entitätstyp.... <---
        single : bool, optional
            Ein Wert, der bestimmt, ob das Schema alle referenzierten Entitätstypen.... <---
        """
        ....

[MikhailArkhipov]

Related microsoft/vscode#5388 actually fixed in VSC right around March 2017.
Currently I see that Russian and other languages working when file is saved in UTF-8 (BOM is not required). Things do get broken if file gets forcibly saved in OEM CP (console CP). VSC loads it incorrectly anyway.

If file is saved in UTF-8 (console CP is 866), VSC works properly

[MikhailArkhipov]

May need to check settings

  // The default character set encoding to use when reading and writing files. This setting can be configured per language too.
  "files.encoding": "utf8",

  // When enabled, will attempt to guess the character set encoding when opening files. This setting can be configured per language too.
  "files.autoGuessEncoding": false,