Check code point validity

Author: msimacek

graalpython/com.oracle.graal.python.test/src/tests/unittest_tags/test_array.txt

@@ -437,6 +437,7 @@
 *graalpython.lib-python.3.test.test_array.UnicodeTest.test_index
 *graalpython.lib-python.3.test.test_array.UnicodeTest.test_initialize_with_unicode
 *graalpython.lib-python.3.test.test_array.UnicodeTest.test_insert
+*graalpython.lib-python.3.test.test_array.UnicodeTest.test_issue17223
 *graalpython.lib-python.3.test.test_array.UnicodeTest.test_iterator_pickle
 *graalpython.lib-python.3.test.test_array.UnicodeTest.test_len
 *graalpython.lib-python.3.test.test_array.UnicodeTest.test_mul

graalpython/com.oracle.graal.python/src/com/oracle/graal/python/builtins/objects/array/ArrayBuiltins.java

@@ -554,14 +554,14 @@ static Object delitem(PArray self, PSlice slice,
             int step = sliceInfo.step;
             int sliceLength = sliceInfo.sliceLength;
             int itemsize = self.getFormat().bytesize;
-            if (simpleStepProfile.profile(step == 1)) {
-                self.delSlice(start, sliceLength);
-            } else {
-                if (step < 0) {
-                    start += 1 + step * (sliceLength - 1) - 1;
-                    step = -step;
-                }
-                if (sliceLength > 0) {
+            if (sliceLength > 0) {
+                if (simpleStepProfile.profile(step == 1)) {
+                    self.delSlice(start, sliceLength);
+                } else {
+                    if (step < 0) {
+                        start += 1 + step * (sliceLength - 1) - 1;
+                        step = -step;
+                    }
                     int cur, offset;
                     for (cur = start, offset = 0; offset < sliceLength - 1; cur += step, offset++) {
                         PythonUtils.arraycopy(self.getBuffer(), (cur + 1) * itemsize, self.getBuffer(), (cur - offset) * itemsize, (step - 1) * itemsize);

graalpython/com.oracle.graal.python/src/com/oracle/graal/python/builtins/objects/array/PArray.java

@@ -119,7 +119,7 @@ public void shift(int from, int by) throws OverflowException {
     }
 
     public void delSlice(int at, int count) {
-        assert count > 0;
+        assert count >= 0;
         assert at + count <= lenght;
         int newLength = lenght - count;
         assert newLength >= 0;

graalpython/com.oracle.graal.python/src/com/oracle/graal/python/builtins/objects/common/BufferStorageNodes.java

@@ -42,13 +42,15 @@
 
 import static com.oracle.graal.python.builtins.PythonBuiltinClassType.OverflowError;
 import static com.oracle.graal.python.builtins.PythonBuiltinClassType.TypeError;
+import static com.oracle.graal.python.builtins.PythonBuiltinClassType.ValueError;
 
 import com.oracle.graal.python.builtins.PythonBuiltinClassType;
 import com.oracle.graal.python.builtins.objects.bytes.PBytes;
 import com.oracle.graal.python.builtins.objects.ints.PInt;
 import com.oracle.graal.python.builtins.objects.object.PythonObjectLibrary;
 import com.oracle.graal.python.builtins.objects.str.PString;
 import com.oracle.graal.python.builtins.objects.str.StringNodes;
+import com.oracle.graal.python.nodes.ErrorMessages;
 import com.oracle.graal.python.nodes.PGuards;
 import com.oracle.graal.python.nodes.PRaiseNode;
 import com.oracle.graal.python.nodes.util.CastToJavaUnsignedLongNode;
@@ -142,8 +144,12 @@ static Object unpackChar(@SuppressWarnings("unused") BufferFormat format, byte[]
         }
 
         @Specialization(guards = "format == UNICODE")
-        static Object unpackUnicode(@SuppressWarnings("unused") BufferFormat format, byte[] bytes, int offset) {
+        static Object unpackUnicode(@SuppressWarnings("unused") BufferFormat format, byte[] bytes, int offset,
+                        @Cached PRaiseNode raiseNode) {
             int codePoint = PythonUtils.arrayAccessor.getInt(bytes, offset);
+            if (!Character.isValidCodePoint(codePoint)) {
+                throw raiseNode.raise(ValueError, ErrorMessages.UNMAPPABLE_CHARACTER);
+            }
             return codePointToString(codePoint);
         }