intrinsified bytes

Author: tomasstupka

graalpython/com.oracle.graal.python/src/com/oracle/graal/python/builtins/Python3Core.java

@@ -307,7 +307,6 @@ private static String[] initializeCoreFiles() {
                         "faulthandler",
                         "base_exception",
                         PythonCextBuiltins.PYTHON_CEXT,
-                        "bytes",
                         "bytearray",
                         "unicodedata",
                         "_locale",

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

@@ -1046,8 +1046,8 @@ private static Object get(PythonContext ctx, String name) {
     }
 
     @Builtin(name = ENCODE, minNumOfPositionalArgs = 1, parameterNames = {"obj", "encoding", "errors"})
-    @ArgumentClinic(name = "encoding", conversion = ArgumentClinic.ClinicConversion.String, defaultValue = "\"utf-8\"", useDefaultForNone = true)
-    @ArgumentClinic(name = "errors", conversion = ArgumentClinic.ClinicConversion.String, defaultValue = "\"strict\"", useDefaultForNone = true)
+    @ArgumentClinic(name = "encoding", conversion = ArgumentClinic.ClinicConversion.String, defaultValue = "\"utf-8\"")
+    @ArgumentClinic(name = "errors", conversion = ArgumentClinic.ClinicConversion.String, defaultValue = "\"strict\"")
     @GenerateNodeFactory
     public abstract static class EncodeNode extends PythonTernaryClinicBuiltinNode {
 
@@ -1075,10 +1075,12 @@ Object encode(VirtualFrame frame, Object obj, String encoding, String errors,
     }
 
     @Builtin(name = DECODE, minNumOfPositionalArgs = 1, parameterNames = {"obj", "encoding", "errors"})
-    @ArgumentClinic(name = "encoding", conversion = ArgumentClinic.ClinicConversion.String, defaultValue = "\"utf-8\"", useDefaultForNone = true)
-    @ArgumentClinic(name = "errors", conversion = ArgumentClinic.ClinicConversion.String, defaultValue = "\"strict\"", useDefaultForNone = true)
+    @ArgumentClinic(name = "encoding", conversion = ArgumentClinic.ClinicConversion.String, defaultValue = "\"utf-8\"")
+    @ArgumentClinic(name = "errors", conversion = ArgumentClinic.ClinicConversion.String, defaultValue = "\"strict\"")
     @GenerateNodeFactory
-    abstract static class DecodeNode extends PythonTernaryClinicBuiltinNode {
+    public abstract static class DecodeNode extends PythonTernaryClinicBuiltinNode {
+
+        public abstract Object executeWithStrings(VirtualFrame frame, Object obj, String encoding, String errors);
 
         @Override
         protected ArgumentClinicProvider getArgumentClinic() {

graalpython/com.oracle.graal.python/src/com/oracle/graal/python/builtins/objects/bytes/BytesBuiltins.java

@@ -77,11 +77,15 @@
 import com.oracle.graal.python.builtins.Python3Core;
 import com.oracle.graal.python.builtins.PythonBuiltinClassType;
 import com.oracle.graal.python.builtins.PythonBuiltins;
+import com.oracle.graal.python.builtins.modules.BuiltinFunctions.IsInstanceNode;
+import com.oracle.graal.python.builtins.modules.CodecsModuleBuiltins;
 import com.oracle.graal.python.builtins.modules.SysModuleBuiltins;
 import com.oracle.graal.python.builtins.objects.PNone;
 import com.oracle.graal.python.builtins.objects.PNotImplemented;
 import com.oracle.graal.python.builtins.objects.buffer.PythonBufferAccessLibrary;
 import com.oracle.graal.python.builtins.objects.buffer.PythonBufferAcquireLibrary;
+import com.oracle.graal.python.builtins.objects.bytes.BytesBuiltinsFactory.LStripNodeFactory;
+import com.oracle.graal.python.builtins.objects.bytes.BytesBuiltinsFactory.RStripNodeFactory;
 import com.oracle.graal.python.builtins.objects.common.IndexNodes;
 import com.oracle.graal.python.builtins.objects.common.SequenceNodes;
 import com.oracle.graal.python.builtins.objects.common.SequenceStorageNodes;
@@ -99,7 +103,9 @@
 import com.oracle.graal.python.lib.PyNumberAsSizeNode;
 import com.oracle.graal.python.lib.PyNumberIndexNode;
 import com.oracle.graal.python.lib.PyObjectGetItem;
+import static com.oracle.graal.python.nodes.BuiltinNames.DECODE;
 import com.oracle.graal.python.nodes.ErrorMessages;
+import static com.oracle.graal.python.nodes.ErrorMessages.DECODER_RETURNED_P_INSTEAD_OF_BYTES;
 import com.oracle.graal.python.nodes.PRaiseNode;
 import com.oracle.graal.python.nodes.SpecialAttributeNames;
 import com.oracle.graal.python.nodes.SpecialMethodNames;
@@ -264,6 +270,48 @@ public static Object repr(PBytes self,
         }
     }
 
+    @Builtin(name = DECODE, minNumOfPositionalArgs = 1, parameterNames = {"$self", "encoding", "errors"}, doc = "Decode the bytes using the codec registered for encoding.\n\n" +
+                    "encoding\n" +
+                    "  The encoding with which to decode the bytes.\n" +
+                    "errors\n" +
+                    "  The error handling scheme to use for the handling of decoding errors.\n" +
+                    "  The default is 'strict' meaning that decoding errors raise a\n" +
+                    "  UnicodeDecodeError. Other possible values are 'ignore' and 'replace'\n" +
+                    "  as well as any other name registered with codecs.register_error that\n" +
+                    "  can handle UnicodeDecodeErrors.")
+    @ArgumentClinic(name = "encoding", conversion = ArgumentClinic.ClinicConversion.String, defaultValue = "\"utf-8\"")
+    @ArgumentClinic(name = "errors", conversion = ArgumentClinic.ClinicConversion.String, defaultValue = "\"strict\"")
+    @GenerateNodeFactory
+    abstract static class DecodeNode extends PythonTernaryClinicBuiltinNode {
+
+        @Override
+        protected ArgumentClinicProvider getArgumentClinic() {
+            return BytesBuiltinsClinicProviders.DecodeNodeClinicProviderGen.INSTANCE;
+        }
+
+        @Specialization
+        public Object decode(VirtualFrame frame, PBytesLike self, String encoding, String errors,
+                        @Cached CodecsModuleBuiltins.DecodeNode decodeNode,
+                        @Cached IsInstanceNode isInstanceNode) {
+            Object result = decodeNode.executeWithStrings(frame, self, encoding, errors);
+            if (!isInstanceNode.executeWith(frame, result, PythonBuiltinClassType.PString)) {
+                throw raise(TypeError, DECODER_RETURNED_P_INSTEAD_OF_BYTES, encoding, result);
+            }
+            return result;
+        }
+    }
+
+    @Builtin(name = "strip", minNumOfPositionalArgs = 1, parameterNames = {"$self", "what"})
+    @GenerateNodeFactory
+    abstract static class StripNode extends PythonBinaryBuiltinNode {
+        @Specialization
+        public Object strip(VirtualFrame frame, PBytesLike self, Object what,
+                        @Cached LStripNode lstripNode,
+                        @Cached RStripNode rstripNode) {
+            return rstripNode.execute(frame, lstripNode.execute(frame, self, what), what);
+        }
+    }
+
     // bytes.translate(table, delete=b'')
     @Builtin(name = "translate", minNumOfPositionalArgs = 2, parameterNames = {"self", "table", "delete"})
     @GenerateNodeFactory
@@ -2235,6 +2283,10 @@ protected int findIndex(byte[] bs, byte[] stripBs, int stripBsLen) {
     @GenerateNodeFactory
     abstract static class LStripNode extends AStripNode {
 
+        static LStripNode create() {
+            return LStripNodeFactory.create();
+        }
+
         @Override
         protected byte[] getResultBytes(byte[] bs, int i) {
             byte[] out;
@@ -2268,6 +2320,10 @@ int start(byte[] bs) {
     @GenerateNodeFactory
     abstract static class RStripNode extends AStripNode {
 
+        static RStripNode create() {
+            return RStripNodeFactory.create();
+        }
+
         @Override
         protected byte[] getResultBytes(byte[] bs, int i) {
             byte[] out;

graalpython/com.oracle.graal.python/src/com/oracle/graal/python/nodes/ErrorMessages.java

@@ -116,6 +116,7 @@ public abstract class ErrorMessages {
     public static final String CALLING_ARG_CONVERTER_FAIL_UNEXPECTED_RETURN = "calling argument converter failed; unexpected return value %s";
     public static final String CALLING_NATIVE_FUNC_EXPECTED_ARGS = "Calling native function %s expected %d arguments but got %d.";
     public static final String CALLING_NATIVE_FUNC_FAILED = "Calling native function %s failed: %m";
+    public static final String DECODER_RETURNED_P_INSTEAD_OF_BYTES = "'%s' decoder returned '%p' instead of 'str'; use codecs.decode() to decode to arbitrary types";
     public static final String ENCODER_RETURNED_P_INSTEAD_OF_BYTES = "'%s' encoder returned '%p' instead of 'bytes'; use codecs.encode() to encode to arbitrary types";
     public static final String EXPECTED_D_ARGS = "expected %d arguments";
     public static final String CAN_ONLY_ASSIGN_S_TO_S_S_NOT_P = "can only assign %s to %s.%s, not %p";