From 43b2dd44f299a68ba313e65d8c93b3db6c47e43f Mon Sep 17 00:00:00 2001
From: Val Markovic <val@markovic.io>
Date: Sun, 28 Feb 2016 14:42:18 -0800
Subject: [PATCH] Ensuring vimsupport only returns unicode text

Python 3 is much stricter around mixing bytes with unicode (and by
"stricter," I mean it doesn't allow it at all) so we're making
vimsupport only return `unicode` objects (`str` on py3). The idea is
that YCM (and ycmd) internals only ever deal with unicode.
---
 python/ycm/tests/vimsupport_test.py | 40 ++++++++++++++++++++++++++++-
 python/ycm/vimsupport.py            |  8 +++---
 third_party/ycmd                    |  2 +-
 3 files changed, 44 insertions(+), 6 deletions(-)

diff --git a/python/ycm/tests/vimsupport_test.py b/python/ycm/tests/vimsupport_test.py
index 2cea87cd..120cb662 100644
--- a/python/ycm/tests/vimsupport_test.py
+++ b/python/ycm/tests/vimsupport_test.py
@@ -1,3 +1,5 @@
+# -*- coding: utf-8 -*-
+#
 # Copyright (C) 2015 YouCompleteMe contributors
 #
 # This file is part of YouCompleteMe.
@@ -28,8 +30,9 @@ MockVimModule()
 
 from ycm import vimsupport
 from nose.tools import eq_
-from hamcrest import assert_that, calling, raises, none
+from hamcrest import assert_that, calling, raises, none, has_entry
 from mock import MagicMock, call, patch
+from ycmd.utils import ToBytes
 import os
 import json
 
@@ -1217,3 +1220,38 @@ def OpenFilename_test( vim_current, vim_command ):
   vim_current.window.options.__setitem__.assert_has_exact_calls( [
     call( 'winfixheight', True )
   ] )
+
+
+@patch( 'ycm.vimsupport.BufferModified', side_effect = [ True ] )
+@patch( 'ycm.vimsupport.FiletypesForBuffer', side_effect = [ [ 'cpp' ] ] )
+def GetUnsavedAndCurrentBufferData_EncodedUnicodeCharsInBuffers_test( *args ):
+  mock_buffer = MagicMock()
+  mock_buffer.name = os.path.realpath( 'filename' )
+  mock_buffer.number = 1
+  mock_buffer.__iter__.return_value = [ u'abc', ToBytes( u'fДa' ) ]
+
+  with patch( 'vim.buffers', [ mock_buffer ] ):
+    assert_that( vimsupport.GetUnsavedAndCurrentBufferData(),
+                 has_entry( mock_buffer.name,
+                            has_entry( u'contents', u'abc\nfДa\n' ) ) )
+
+
+# NOTE: Vim returns byte offsets for columns, not actual character columns. This
+# makes 'ДД' have 4 columns: column 0, column 2 and column 4.
+@patch( 'vim.current.line', ToBytes( 'ДДaa' ) )
+@patch( 'ycm.vimsupport.CurrentColumn', side_effect = [ 4 ] )
+def TextBeforeCursor_EncodedUnicode_test( *args ):
+  eq_( vimsupport.TextBeforeCursor(), u'ДД' )
+
+
+# NOTE: Vim returns byte offsets for columns, not actual character columns. This
+# makes 'ДД' have 4 columns: column 0, column 2 and column 4.
+@patch( 'vim.current.line', ToBytes( 'aaДД' ) )
+@patch( 'ycm.vimsupport.CurrentColumn', side_effect = [ 2 ] )
+def TextAfterCursor_EncodedUnicode_test( *args ):
+  eq_( vimsupport.TextAfterCursor(), u'ДД' )
+
+
+@patch( 'vim.current.line', ToBytes( 'fДa' ) )
+def CurrentLineContents_EncodedUnicode_test( *args ):
+  eq_( vimsupport.CurrentLineContents(), u'fДa' )
diff --git a/python/ycm/vimsupport.py b/python/ycm/vimsupport.py
index 2a3ad5fa..4921ed23 100644
--- a/python/ycm/vimsupport.py
+++ b/python/ycm/vimsupport.py
@@ -68,17 +68,17 @@ def CurrentColumn():
 
 
 def CurrentLineContents():
-  return vim.current.line
+  return ToUnicode( vim.current.line )
 
 
 def TextAfterCursor():
   """Returns the text after CurrentColumn."""
-  return vim.current.line[ CurrentColumn(): ]
+  return ToUnicode( vim.current.line[ CurrentColumn(): ] )
 
 
 def TextBeforeCursor():
   """Returns the text before CurrentColumn."""
-  return vim.current.line[ :CurrentColumn() ]
+  return ToUnicode( vim.current.line[ :CurrentColumn() ] )
 
 
 # Expects version_string in 'MAJOR.MINOR.PATCH' format, e.g. '7.4.301'
@@ -123,7 +123,7 @@ def GetUnsavedAndCurrentBufferData():
 
     buffers_data[ GetBufferFilepath( buffer_object ) ] = {
       # Add a newline to match what gets saved to disk. See #1455 for details.
-      'contents': '\n'.join( buffer_object ) + '\n',
+      'contents': '\n'.join( ToUnicode( x ) for x in buffer_object ) + '\n',
       'filetypes': FiletypesForBuffer( buffer_object )
     }
 
diff --git a/third_party/ycmd b/third_party/ycmd
index 7ea44df6..abe2e1c1 160000
--- a/third_party/ycmd
+++ b/third_party/ycmd
@@ -1 +1 @@
-Subproject commit 7ea44df61e524338db66521fd2d5a917a7c472fb
+Subproject commit abe2e1c1145bcf865a5d0c265214f871576644ec