def CompletionStartCodepoint(line_value, column_num, filetype):
"""Returns the 1-based codepoint index where the completion query should
start. So if the user enters:
ƒøø.∫å®^
with the cursor being at the location of the caret (so the character *AFTER*
'®'), then the starting column would be the index of the character '∫'
(i.e. 5, not its byte index)."""
# NOTE: column_num and other numbers on the wire are byte indices, but we need
# to walk codepoints for identifier checks.
codepoint_column_num = ByteOffsetToCodepointOffset(line_value, column_num)
unicode_line_value = ToUnicode(line_value)
# -1 and then +1 to account for difference between 0-based and 1-based
# indices/columns
codepoint_start_column = StartOfLongestIdentifierEndingAtIndex(
unicode_line_value, codepoint_column_num - 1, filetype) + 1
return codepoint_start_column
def CompletionStartColumn(line_value, column_num, filetype):
"""Returns the 1-based index where the completion query should start. So if
the user enters:
foo.bar^
with the cursor being at the location of the caret (so the character *AFTER*
'r'), then the starting column would be the index of the letter 'b'."""
# NOTE: column_num and other numbers on the wire are byte indices, but we need
# to walk codepoints for identifier checks.
utf8_line_value = ToUtf8IfNeeded(line_value)
unicode_line_value = ToUnicodeIfNeeded(line_value)
codepoint_column_num = len(
unicode(utf8_line_value[:column_num - 1], 'utf8')) + 1
# -1 and then +1 to account for difference betwen 0-based and 1-based
# indices/columns
codepoint_start_column = StartOfLongestIdentifierEndingAtIndex(
unicode_line_value, codepoint_column_num - 1, filetype) + 1
return len(
unicode_line_value[:codepoint_start_column - 1].encode('utf8')) + 1