95efbc8766
[READY] Handle null characters in completion response The `detailed_info` field may contain null characters e.g. `\x00` in Python docstrings (though it doesn't make sense to use such characters in a docstring). Since these characters cannot be evaluated through `pyeval` (and `py3eval`), we replace them with the symbol used by Vim to display them (`^@`). This PR also rewrites the `_ConvertCompletionDataToVimData` function by relying on two things: - [the `insertion_text` field is always defined](600f54de20/ycmd/responses.py (L109)
); - [Vim ignores fields with an empty string](833e5dab14/src/edit.c (L2748-L2755)
). I did some measurements and the performance hit is negligible. Fixes #3207. <!-- Reviewable:start --> --- This change is [<img src="https://reviewable.io/review_button.svg" height="34" align="absmiddle" alt="Reviewable"/>](https://reviewable.io/reviews/valloric/youcompleteme/3208) <!-- Reviewable:end -->
234 lines
8.2 KiB
Python
234 lines
8.2 KiB
Python
# Copyright (C) 2013-2018 YouCompleteMe contributors
|
|
#
|
|
# This file is part of YouCompleteMe.
|
|
#
|
|
# YouCompleteMe is free software: you can redistribute it and/or modify
|
|
# it under the terms of the GNU General Public License as published by
|
|
# the Free Software Foundation, either version 3 of the License, or
|
|
# (at your option) any later version.
|
|
#
|
|
# YouCompleteMe is distributed in the hope that it will be useful,
|
|
# but WITHOUT ANY WARRANTY; without even the implied warranty of
|
|
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
|
# GNU General Public License for more details.
|
|
#
|
|
# You should have received a copy of the GNU General Public License
|
|
# along with YouCompleteMe. If not, see <http://www.gnu.org/licenses/>.
|
|
|
|
from __future__ import unicode_literals
|
|
from __future__ import print_function
|
|
from __future__ import division
|
|
from __future__ import absolute_import
|
|
# Not installing aliases from python-future; it's unreliable and slow.
|
|
from builtins import * # noqa
|
|
|
|
import logging
|
|
from future.utils import iteritems
|
|
from ycmd.utils import ToUnicode
|
|
from ycm.client.base_request import ( BaseRequest, DisplayServerException,
|
|
MakeServerException )
|
|
from ycm import vimsupport
|
|
from ycm.vimsupport import NO_COMPLETIONS
|
|
|
|
_logger = logging.getLogger( __name__ )
|
|
|
|
|
|
class CompletionRequest( BaseRequest ):
|
|
def __init__( self, request_data ):
|
|
super( CompletionRequest, self ).__init__()
|
|
self.request_data = request_data
|
|
self._response_future = None
|
|
self._complete_done_hooks = {
|
|
'cs': self._OnCompleteDone_Csharp,
|
|
'java': self._OnCompleteDone_FixIt,
|
|
'javascript': self._OnCompleteDone_FixIt,
|
|
'typescript': self._OnCompleteDone_FixIt,
|
|
}
|
|
|
|
|
|
def Start( self ):
|
|
self._response_future = self.PostDataToHandlerAsync( self.request_data,
|
|
'completions' )
|
|
|
|
|
|
def Done( self ):
|
|
return bool( self._response_future ) and self._response_future.done()
|
|
|
|
|
|
def RawResponse( self ):
|
|
if not self._response_future:
|
|
return NO_COMPLETIONS
|
|
|
|
response = self.HandleFuture( self._response_future,
|
|
truncate_message = True )
|
|
if not response:
|
|
return NO_COMPLETIONS
|
|
|
|
# Vim may not be able to convert the 'errors' entry to its internal format
|
|
# so we remove it from the response.
|
|
errors = response.pop( 'errors', [] )
|
|
for e in errors:
|
|
exception = MakeServerException( e )
|
|
_logger.error( exception )
|
|
DisplayServerException( exception, truncate_message = True )
|
|
|
|
response[ 'line' ] = self.request_data[ 'line_num' ]
|
|
response[ 'column' ] = self.request_data[ 'column_num' ]
|
|
return response
|
|
|
|
|
|
def Response( self ):
|
|
response = self.RawResponse()
|
|
response[ 'completions' ] = _ConvertCompletionDatasToVimDatas(
|
|
response[ 'completions' ] )
|
|
return response
|
|
|
|
|
|
def OnCompleteDone( self ):
|
|
if not self.Done():
|
|
return
|
|
|
|
complete_done_actions = self._GetCompleteDoneHooks()
|
|
for action in complete_done_actions:
|
|
action()
|
|
|
|
|
|
def _GetCompleteDoneHooks( self ):
|
|
filetypes = vimsupport.CurrentFiletypes()
|
|
for key, value in iteritems( self._complete_done_hooks ):
|
|
if key in filetypes:
|
|
yield value
|
|
|
|
|
|
def _GetCompletionsUserMayHaveCompleted( self ):
|
|
completed_item = vimsupport.GetVariableValue( 'v:completed_item' )
|
|
|
|
# If Vim supports user_data (8.0.1493 or later), we actually know the
|
|
# _exact_ element that was selected, having put its index in the
|
|
# user_data field. Otherwise, we have to guess by matching the values in the
|
|
# completed item and the list of completions. Sometimes this returns
|
|
# multiple possibilities, which is essentially unresolvable.
|
|
if 'user_data' not in completed_item:
|
|
completions = self.RawResponse()[ 'completions' ]
|
|
return _FilterToMatchingCompletions( completed_item, completions )
|
|
|
|
if completed_item[ 'user_data' ]:
|
|
completions = self.RawResponse()[ 'completions' ]
|
|
return [ completions[ int( completed_item[ 'user_data' ] ) ] ]
|
|
|
|
return []
|
|
|
|
|
|
def _OnCompleteDone_Csharp( self ):
|
|
completions = self._GetCompletionsUserMayHaveCompleted()
|
|
namespaces = [ _GetRequiredNamespaceImport( c ) for c in completions ]
|
|
namespaces = [ n for n in namespaces if n ]
|
|
if not namespaces:
|
|
return
|
|
|
|
if len( namespaces ) > 1:
|
|
choices = [ "{0} {1}".format( i + 1, n )
|
|
for i, n in enumerate( namespaces ) ]
|
|
choice = vimsupport.PresentDialog( "Insert which namespace:", choices )
|
|
if choice < 0:
|
|
return
|
|
namespace = namespaces[ choice ]
|
|
else:
|
|
namespace = namespaces[ 0 ]
|
|
|
|
vimsupport.InsertNamespace( namespace )
|
|
|
|
|
|
def _OnCompleteDone_FixIt( self ):
|
|
completions = self._GetCompletionsUserMayHaveCompleted()
|
|
fixit_completions = [ _GetFixItCompletion( c ) for c in completions ]
|
|
fixit_completions = [ f for f in fixit_completions if f ]
|
|
if not fixit_completions:
|
|
return
|
|
|
|
# If we have user_data in completions (8.0.1493 or later), then we would
|
|
# only ever return max. 1 completion here. However, if we had to guess, it
|
|
# is possible that we matched multiple completion items (e.g. for overloads,
|
|
# or similar classes in multiple packages). In any case, rather than
|
|
# prompting the user and disturbing her workflow, we just apply the first
|
|
# one. This might be wrong, but the solution is to use a (very) new version
|
|
# of Vim which supports user_data on completion items
|
|
fixit_completion = fixit_completions[ 0 ]
|
|
|
|
for fixit in fixit_completion:
|
|
vimsupport.ReplaceChunks( fixit[ 'chunks' ], silent=True )
|
|
|
|
|
|
def _GetRequiredNamespaceImport( completion ):
|
|
if ( 'extra_data' not in completion
|
|
or 'required_namespace_import' not in completion[ 'extra_data' ] ):
|
|
return None
|
|
return completion[ 'extra_data' ][ 'required_namespace_import' ]
|
|
|
|
|
|
def _GetFixItCompletion( completion ):
|
|
if ( 'extra_data' not in completion
|
|
or 'fixits' not in completion[ 'extra_data' ] ):
|
|
return None
|
|
|
|
return completion[ 'extra_data' ][ 'fixits' ]
|
|
|
|
|
|
def _FilterToMatchingCompletions( completed_item, completions ):
|
|
"""Filter to completions matching the item Vim said was completed"""
|
|
match_keys = [ 'word', 'abbr', 'menu', 'info' ]
|
|
matched_completions = []
|
|
for index, completion in enumerate( completions ):
|
|
item = _ConvertCompletionDataToVimData( index, completion )
|
|
|
|
def matcher( key ):
|
|
return ( ToUnicode( completed_item.get( key, "" ) ) ==
|
|
ToUnicode( item.get( key, "" ) ) )
|
|
|
|
if all( matcher( i ) for i in match_keys ):
|
|
matched_completions.append( completion )
|
|
return matched_completions
|
|
|
|
|
|
def _GetCompletionInfoField( completion_data ):
|
|
info = completion_data.get( 'detailed_info', '' )
|
|
|
|
if 'extra_data' in completion_data:
|
|
docstring = completion_data[ 'extra_data' ].get( 'doc_string', '' )
|
|
if docstring:
|
|
if info:
|
|
info += '\n' + docstring
|
|
else:
|
|
info = docstring
|
|
|
|
# This field may contain null characters e.g. \x00 in Python docstrings. Vim
|
|
# cannot evaluate such characters so they are removed.
|
|
return info.replace( '\x00', '' )
|
|
|
|
|
|
def _ConvertCompletionDataToVimData( completion_identifier, completion_data ):
|
|
# See :h complete-items for a description of the dictionary fields.
|
|
return {
|
|
'word' : completion_data[ 'insertion_text' ],
|
|
'abbr' : completion_data.get( 'menu_text', '' ),
|
|
'menu' : completion_data.get( 'extra_menu_info', '' ),
|
|
'info' : _GetCompletionInfoField( completion_data ),
|
|
'kind' : ToUnicode( completion_data.get( 'kind', '' ) )[ :1 ].lower(),
|
|
'dup' : 1,
|
|
'empty' : 1,
|
|
# We store the completion item index as a string in the completion
|
|
# user_data. This allows us to identify the _exact_ item that was completed
|
|
# in the CompleteDone handler, by inspecting this item from v:completed_item
|
|
#
|
|
# We convert to string because completion user data items must be strings.
|
|
#
|
|
# Note: Not all versions of Vim support this (added in 8.0.1483), but adding
|
|
# the item to the dictionary is harmless in earlier Vims.
|
|
'user_data': str( completion_identifier )
|
|
}
|
|
|
|
|
|
def _ConvertCompletionDatasToVimDatas( response_data ):
|
|
return [ _ConvertCompletionDataToVimData( i, x )
|
|
for i, x in enumerate( response_data ) ]
|