assert_valid_name: share character classes with lexer

Replicates graphql/graphql-js@22b9504

Author: Cito

src/graphql/language/character_classes.py

@@ -0,0 +1,33 @@
+__all__ = ["is_digit", "is_letter", "is_name_start", "is_name_continue"]
+
+
+def is_digit(char: str) -> bool:
+    """Check whether char is a digit
+
+    For internal use by the lexer only.
+    """
+    return "0" <= char <= "9"
+
+
+def is_letter(char: str) -> bool:
+    """Check whether char is a plain ASCII letter
+
+    For internal use by the lexer only.
+    """
+    return "A" <= char <= "Z" or "a" <= char <= "z"
+
+
+def is_name_start(char: str) -> bool:
+    """Check whether char is allowed at the beginning of a GraphQL name
+
+    For internal use by the lexer only.
+    """
+    return is_letter(char) or char == "_"
+
+
+def is_name_continue(char: str) -> bool:
+    """Check whether char is allowed in the continuation of a GraphQL name
+
+    For internal use by the lexer only.
+    """
+    return is_letter(char) or is_digit(char) or char == "_"

src/graphql/language/lexer.py

@@ -3,6 +3,7 @@
 from ..error import GraphQLSyntaxError
 from .ast import Token
 from .block_string import dedent_block_string_value
+from .character_classes import is_digit, is_name_start, is_name_continue
 from .source import Source
 from .token_kind import TokenKind
 
@@ -129,10 +130,10 @@ def read_next_token(self, start: int) -> Token:
             if kind:
                 return self.create_token(kind, position, position + 1)
 
-            if "0" <= char <= "9" or char == "-":
+            if is_digit(char) or char == "-":
                 return self.read_number(position, char)
 
-            if "A" <= char <= "Z" or "a" <= char <= "z" or char == "_":
+            if is_name_start(char):
                 return self.read_name(position)
 
             if char == ".":
@@ -196,7 +197,7 @@ def read_number(self, start: int, first_char: str) -> Token:
         if char == "0":
             position += 1
             char = body[position : position + 1]
-            if "0" <= char <= "9":
+            if is_digit(char):
                 raise GraphQLSyntaxError(
                     self.source,
                     position,
@@ -240,7 +241,7 @@ def read_number(self, start: int, first_char: str) -> Token:
 
     def read_digits(self, start: int, first_char: str) -> int:
         """Return the new position in the source after reading one or more digits."""
-        if not "0" <= first_char <= "9":
+        if not is_digit(first_char):
             raise GraphQLSyntaxError(
                 self.source,
                 start,
@@ -251,7 +252,7 @@ def read_digits(self, start: int, first_char: str) -> int:
         body = self.source.body
         body_length = len(body)
         position = start + 1
-        while position < body_length and "0" <= body[position] <= "9":
+        while position < body_length and is_digit(body[position]):
             position += 1
         return position
 
@@ -427,12 +428,7 @@ def read_name(self, start: int) -> Token:
 
         while position < body_length:
             char = body[position]
-            if not (
-                "A" <= char <= "Z"
-                or "a" <= char <= "z"
-                or "0" <= char <= "9"
-                or char == "_"
-            ):
+            if not is_name_continue(char):
                 break
             position += 1
 
@@ -558,8 +554,3 @@ def is_supplementary_code_point(body: str, location: int) -> bool:
 
 def decode_surrogate_pair(leading: int, trailing: int) -> int:
     return 0x10000 + (((leading & 0x03FF) << 10) | (trailing & 0x03FF))
-
-
-def is_name_start(char: str) -> bool:
-    """Check whether char is an underscore or a plain ASCII letter"""
-    return char == "_" or "A" <= char <= "Z" or "a" <= char <= "z"

src/graphql/utilities/assert_valid_name.py

@@ -1,6 +1,7 @@
 import re
 from typing import Optional
 
+from ..language.character_classes import is_name_start, is_name_continue
 from ..error import GraphQLError
 
 __all__ = ["assert_valid_name", "is_valid_name_error"]
@@ -21,13 +22,22 @@ def is_valid_name_error(name: str) -> Optional[GraphQLError]:
     """Return an Error if a name is invalid."""
     if not isinstance(name, str):
         raise TypeError("Expected name to be a string.")
+
     if name.startswith("__"):
         return GraphQLError(
             f"Name {name!r} must not begin with '__',"
             " which is reserved by GraphQL introspection."
         )
-    if not re_name.match(name):
+
+    if not name:
+        return GraphQLError("Expected name to be a non-empty string.")
+
+    if not all(is_name_continue(char) for char in name[1:]):
         return GraphQLError(
-            f"Names must match /^[_a-zA-Z][_a-zA-Z0-9]*$/ but {name!r} does not."
+            f"Names must only contain [_a-zA-Z0-9] but {name!r} does not."
         )
+
+    if not is_name_start(name[0]):
+        return GraphQLError(f"Names must start with [_a-zA-Z] but {name!r} does not.")
+
     return None

tests/type/test_validation.py

@@ -500,7 +500,7 @@ def rejects_an_object_type_with_incorrectly_named_fields():
         )
         msg = validate_schema(schema)[0].message
         assert msg == (
-            "Names must match /^[_a-zA-Z][_a-zA-Z0-9]*$/"
+            "Names must only contain [_a-zA-Z0-9]"
             " but 'bad-name-with-dashes' does not."
         )
 
@@ -534,7 +534,7 @@ def rejects_field_args_with_invalid_names():
 
         msg = validate_schema(schema)[0].message
         assert msg == (
-            "Names must match /^[_a-zA-Z][_a-zA-Z0-9]*$/"
+            "Names must only contain [_a-zA-Z0-9]"
             " but 'bad-name-with-dashes' does not."
         )
 
@@ -1001,20 +1001,16 @@ def schema_with_enum(name: str) -> GraphQLSchema:
 
         schema1 = schema_with_enum("#value")
         msg = validate_schema(schema1)[0].message
-        assert msg == (
-            "Names must match /^[_a-zA-Z][_a-zA-Z0-9]*$/ but '#value' does not."
-        )
+        assert msg == ("Names must start with [_a-zA-Z] but '#value' does not.")
 
         schema2 = schema_with_enum("1value")
         msg = validate_schema(schema2)[0].message
-        assert msg == (
-            "Names must match /^[_a-zA-Z][_a-zA-Z0-9]*$/ but '1value' does not."
-        )
+        assert msg == ("Names must start with [_a-zA-Z] but '1value' does not.")
 
         schema3 = schema_with_enum("KEBAB-CASE")
         msg = validate_schema(schema3)[0].message
         assert msg == (
-            "Names must match /^[_a-zA-Z][_a-zA-Z0-9]*$/ but 'KEBAB-CASE' does not."
+            "Names must only contain [_a-zA-Z0-9] but 'KEBAB-CASE' does not."
         )
 
         schema4 = schema_with_enum("true")

tests/utilities/test_assert_valid_name.py

@@ -24,6 +24,22 @@ def throws_for_non_strings():
         msg = str(exc_info.value)
         assert msg == "Expected name to be a string."
 
+    def throws_on_empty_strings():
+        with raises(GraphQLError) as exc_info:
+            assert_valid_name("")
+        msg = str(exc_info.value)
+        assert msg == "Expected name to be a non-empty string."
+
     def throws_for_names_with_invalid_characters():
-        with raises(GraphQLError, match="Names must match"):
+        with raises(GraphQLError) as exc_info:
             assert_valid_name(">--()-->")
+        msg = str(exc_info.value)
+        assert msg == "Names must only contain [_a-zA-Z0-9] but '>--()-->' does not."
+
+    def throws_for_names_starting_with_invalid_characters():
+        with raises(GraphQLError) as exc_info:
+            assert_valid_name("42MeaningsOfLife")
+        msg = str(exc_info.value)
+        assert msg == (
+            "Names must start with [_a-zA-Z] but '42MeaningsOfLife' does not."
+        )