feat: support UTF32 ans UTF8 position encoding

Contributes to #346

pygls/workspace/document.py

@@ -169,7 +169,7 @@ def offset_at_position(self, client_position: types.Position) -> int:
             lines, client_position
         )
         row, col = server_position.line, server_position.character
-        return col + sum(self.position.utf16_num_units(line) for line in lines[:row])
+        return col + sum(self.position.client_num_units(line) for line in lines[:row])
 
     @property
     def source(self) -> str:

pygls/workspace/position.py

@@ -47,14 +47,21 @@ def utf16_unit_offset(self, chars: str):
         """
         return sum(self.is_char_beyond_multilingual_plane(ch) for ch in chars)
 
-    def utf16_num_units(self, chars: str):
+    def client_num_units(self, chars: str):
         """
         Calculate the length of `str` in utf-16 code units.
 
         Arguments:
             chars (str): The string to return the length in utf-16 code units for.
         """
-        return len(chars) + self.utf16_unit_offset(chars)
+        utf32_units = len(chars)
+        if self.encoding == types.PositionEncodingKind.Utf32:
+            return utf32_units
+
+        if self.encoding == types.PositionEncodingKind.Utf8:
+            return utf32_units + (self.utf16_unit_offset(chars) * 2)
+
+        return utf32_units + self.utf16_unit_offset(chars)
 
     def position_from_client_units(
         self, lines: List[str], position: types.Position
@@ -88,24 +95,24 @@ def position_from_client_units(
         if len(lines) == 0:
             return types.Position(0, 0)
         if position.line >= len(lines):
-            return types.Position(len(lines) - 1, self.utf16_num_units(lines[-1]))
+            return types.Position(len(lines) - 1, self.client_num_units(lines[-1]))
 
         _line = lines[position.line]
         _line = _line.replace("\r\n", "\n")  # TODO: it's a bit of a hack
-        _utf16_len = self.utf16_num_units(_line)
+        _client_len = self.client_num_units(_line)
         _utf32_len = len(_line)
 
-        if _utf16_len == 0:
+        if _client_len == 0:
             return types.Position(position.line, 0)
 
-        _utf16_end_of_line = self.utf16_num_units(_line)
-        if position.character > _utf16_end_of_line:
-            position.character = _utf16_end_of_line - 1
+        _client_end_of_line = self.client_num_units(_line)
+        if position.character > _client_end_of_line:
+            position.character = _client_end_of_line - 1
 
-        _utf16_index = 0
+        _client_index = 0
         utf32_index = 0
         while True:
-            _is_searching_queried_position = _utf16_index < position.character
+            _is_searching_queried_position = _client_index < position.character
             _is_before_end_of_line = utf32_index < _utf32_len
             _is_searching_for_position = (
                 _is_searching_queried_position and _is_before_end_of_line
@@ -116,9 +123,13 @@ def position_from_client_units(
             _current_char = _line[utf32_index]
             _is_double_width = Position.is_char_beyond_multilingual_plane(_current_char)
             if _is_double_width:
-                _utf16_index += 2
+                if self.encoding == types.PositionEncodingKind.Utf32:
+                    _client_index += 1
+                if self.encoding == types.PositionEncodingKind.Utf8:
+                    _client_index += 4
+                _client_index += 2
             else:
-                _utf16_index += 1
+                _client_index += 1
             utf32_index += 1
 
         position = types.Position(line=position.line, character=utf32_index)
@@ -141,10 +152,12 @@ def position_to_client_unit(
             The position with `character` being converted to UTF-[32|16|8] code units.
         """
         try:
+            character = self.client_num_units(
+                lines[position.line][: position.character]
+            )
             return types.Position(
                 line=position.line,
-                character=position.character
-                + self.utf16_unit_offset(lines[position.line][: position.character]),
+                character=character,
             )
         except IndexError:
             return types.Position(line=len(lines), character=0)

tests/conftest.py

@@ -111,11 +111,6 @@ def server_dir():
 json_server_client = create_client_for_server("json_server.py")
 
 
-@pytest.fixture
-def doc():
-    return TextDocument(DOC_URI, DOC)
-
-
 @pytest.fixture
 def feature_manager():
     """Return a feature manager"""

tests/test_document.py

@@ -105,7 +105,8 @@ def test_document_line_edit():
     assert doc.source == "itsgoodbyeworld"
 
 
-def test_document_lines(doc):
+def test_document_lines():
+    doc = Document(DOC_URI, DOC)
     assert len(doc.lines) == 4
     assert doc.lines[0] == "document\n"
 
@@ -160,7 +161,9 @@ def test_document_no_edit():
     assert doc.lines == old
 
 
-def test_document_props(doc):
+def test_document_props():
+    doc = Document(DOC_URI, DOC)
+
     assert doc.uri == DOC_URI
     assert doc.source == DOC
 
@@ -180,9 +183,25 @@ def test_position_from_utf16():
         ['x="😋"'], types.Position(line=0, character=5)
     ) == types.Position(line=0, character=4)
 
-    actual = types.Position(line=0, character=5)
-    position.position_from_client_units(['x="😋"'], actual)
-    assert actual == types.Position(line=0, character=5)
+
+def test_position_from_utf32():
+    position = Position(encoding=types.PositionEncodingKind.Utf32)
+    assert position.position_from_client_units(
+        ['x="😋"'], types.Position(line=0, character=3)
+    ) == types.Position(line=0, character=3)
+    assert position.position_from_client_units(
+        ['x="😋"'], types.Position(line=0, character=4)
+    ) == types.Position(line=0, character=4)
+
+
+def test_position_from_utf8():
+    position = Position(encoding=types.PositionEncodingKind.Utf8)
+    assert position.position_from_client_units(
+        ['x="😋"'], types.Position(line=0, character=3)
+    ) == types.Position(line=0, character=3)
+    assert position.position_from_client_units(
+        ['x="😋"'], types.Position(line=0, character=7)
+    ) == types.Position(line=0, character=4)
 
 
 def test_position_to_utf16():
@@ -195,9 +214,27 @@ def test_position_to_utf16():
         ['x="😋"'], types.Position(line=0, character=4)
     ) == types.Position(line=0, character=5)
 
-    actual = types.Position(line=0, character=4)
-    position.position_to_client_unit(['x="😋"'], actual)
-    assert actual == types.Position(line=0, character=4)
+
+def test_position_to_utf32():
+    position = Position(encoding=types.PositionEncodingKind.Utf32)
+    assert position.position_to_client_unit(
+        ['x="😋"'], types.Position(line=0, character=3)
+    ) == types.Position(line=0, character=3)
+
+    assert position.position_to_client_unit(
+        ['x="😋"'], types.Position(line=0, character=4)
+    ) == types.Position(line=0, character=4)
+
+
+def test_position_to_utf8():
+    position = Position(encoding=types.PositionEncodingKind.Utf8)
+    assert position.position_to_client_unit(
+        ['x="😋"'], types.Position(line=0, character=3)
+    ) == types.Position(line=0, character=3)
+
+    assert position.position_to_client_unit(
+        ['x="😋"'], types.Position(line=0, character=4)
+    ) == types.Position(line=0, character=6)
 
 
 def test_range_from_utf16():
@@ -250,7 +287,8 @@ def test_range_to_utf16():
     assert actual == expected
 
 
-def test_offset_at_position(doc):
+def test_offset_at_position_utf16():
+    doc = Document(DOC_URI, DOC)
     assert doc.offset_at_position(types.Position(line=0, character=8)) == 8
     assert doc.offset_at_position(types.Position(line=1, character=5)) == 12
     assert doc.offset_at_position(types.Position(line=2, character=0)) == 13
@@ -262,6 +300,18 @@ def test_offset_at_position(doc):
     assert doc.offset_at_position(types.Position(line=5, character=0)) == 40
 
 
+def test_offset_at_position_utf32():
+    doc = Document(DOC_URI, DOC, position_encoding=types.PositionEncodingKind.Utf32)
+    assert doc.offset_at_position(types.Position(line=0, character=8)) == 8
+    assert doc.offset_at_position(types.Position(line=5, character=0)) == 39
+
+
+def test_offset_at_position_utf8():
+    doc = Document(DOC_URI, DOC, position_encoding=types.PositionEncodingKind.Utf8)
+    assert doc.offset_at_position(types.Position(line=0, character=8)) == 8
+    assert doc.offset_at_position(types.Position(line=5, character=0)) == 41
+
+
 def test_utf16_to_utf32_position_cast():
     position = Position(encoding=types.PositionEncodingKind.Utf16)
     lines = ["", "😋😋", ""]
@@ -305,10 +355,12 @@ def test_position_for_line_endings():
     ) == types.Position(line=1, character=1)
 
 
-def test_word_at_position(doc):
+def test_word_at_position():
     """
     Return word under the cursor (or last in line if past the end)
     """
+    doc = Document(DOC_URI, DOC)
+
     assert doc.word_at_position(types.Position(line=0, character=8)) == "document"
     assert doc.word_at_position(types.Position(line=0, character=1000)) == "document"
     assert doc.word_at_position(types.Position(line=1, character=5)) == "for"