Implement unicodedata.combining()

Author: andrykonchin

graalpython/com.oracle.graal.python.test/src/tests/test_unicodedata.py

@@ -88,3 +88,17 @@ def test_east_asian_width(self):
         empty_string = ""
         with self.assertRaisesRegex(TypeError, r"east_asian_width\(\) argument must be a unicode character, not str"):
             unicodedata.east_asian_width(empty_string)
+
+
+    def test_combining(self):
+        list = [1, 2, 3]
+        with self.assertRaisesRegex(TypeError, r"combining\(\) argument must be a unicode character, not list"):
+            unicodedata.combining(list)
+
+        multi_character_string = "abc"
+        with self.assertRaisesRegex(TypeError, r"combining\(\) argument must be a unicode character, not str"):
+            unicodedata.combining(multi_character_string)
+
+        empty_string = ""
+        with self.assertRaisesRegex(TypeError, r"combining\(\) argument must be a unicode character, not str"):
+            unicodedata.combining(empty_string)

graalpython/com.oracle.graal.python/src/com/oracle/graal/python/builtins/modules/UnicodeDataModuleBuiltins.java

@@ -52,6 +52,7 @@
 
 import com.oracle.graal.python.builtins.objects.module.PythonModule;
 import com.oracle.graal.python.nodes.function.builtins.PythonUnaryBuiltinNode;
+import com.oracle.graal.python.nodes.util.CannotCastException;
 import com.oracle.graal.python.nodes.util.CastToTruffleStringNode;
 import com.oracle.truffle.api.strings.TruffleString.CodePointAtByteIndexNode;
 import com.oracle.truffle.api.strings.TruffleString.CodePointLengthNode;
@@ -333,6 +334,32 @@ protected ArgumentClinicProvider getArgumentClinic() {
         }
     }
 
+    // unicodedata.combining(chr)
+    @Builtin(name = "combining", minNumOfPositionalArgs = 1, numOfPositionalOnlyArgs = 1, parameterNames = {"chr"})
+    @GenerateNodeFactory
+    public abstract static class CombiningNode extends PythonUnaryBuiltinNode {
+
+        @Specialization
+        @TruffleBoundary
+        static Object combining(Object object,
+                        @Bind Node inliningTarget) {
+            final TruffleString chr;
+
+            try {
+                chr = CastToTruffleStringNode.getUncached().execute(inliningTarget, object);
+            } catch (CannotCastException e) {
+                throw PRaiseNode.raiseStatic(inliningTarget, TypeError, ErrorMessages.S_ARG_MUST_BE_S_NOT_P, "combining()", "a unicode character", object);
+            }
+
+            if (CodePointLengthNode.getUncached().execute(chr, TS_ENCODING) != 1) {
+                throw PRaiseNode.raiseStatic(inliningTarget, TypeError, ErrorMessages.S_ARG_MUST_BE_S_NOT_P, "combining()", "a unicode character", object);
+            }
+
+            int codepoint = CodePointAtByteIndexNode.getUncached().execute(chr, 0, TS_ENCODING);
+            return UCharacter.getCombiningClass(codepoint);
+        }
+    }
+
     // unicode.east_asia_width(chr)
     @Builtin(name = "east_asian_width", minNumOfPositionalArgs = 1, numOfPositionalOnlyArgs = 1, parameterNames = {"chr"})
     @GenerateNodeFactory

graalpython/lib-graalpython/unicodedata.py

@@ -40,6 +40,6 @@
 __graalpython__.import_current_as_named_module_with_delegate(
     module_name="unicodedata",
     delegate_name="_cpython_unicodedata",
-    delegate_attributes=['ucd_3_2_0', 'combining'],
+    delegate_attributes=['ucd_3_2_0'],
     wrap_methods=False,
     owner_globals=globals())