fix

Author: EvgenyVashkevich

mypyc/irbuild/specialize.py

@@ -84,13 +84,17 @@
     join_formatted_strings,
     tokenizer_format_call,
 )
+from mypyc.primitives.bytes_ops import (
+    bytes_decode_ascii_strict,
+    bytes_decode_latin1_strict,
+    bytes_decode_utf8_strict,
+)
 from mypyc.primitives.dict_ops import (
     dict_items_op,
     dict_keys_op,
     dict_setdefault_spec_init_op,
     dict_values_op,
 )
-from mypyc.primitives.bytes_ops import bytes_decode_utf8_strict, bytes_decode_latin1_strict, bytes_decode_ascii_strict
 from mypyc.primitives.list_ops import new_list_set_item_op
 from mypyc.primitives.str_ops import (
     str_encode_ascii_strict,
@@ -748,15 +752,12 @@ def bytes_decode_fast_path(builder: IRBuilder, expr: CallExpr, callee: RefExpr)
         return None
 
     encoding = "utf8"
-    errors = "strict"
 
     # Handle up to 2 arguments: decode([encoding], [errors])
     if len(expr.arg_kinds) > 0 and isinstance(expr.args[0], StrExpr):
         if expr.arg_kinds[0] == ARG_NAMED:
             if expr.arg_names[0] == "encoding":
                 encoding = expr.args[0].value
-            elif expr.arg_names[0] == "errors":
-                errors = expr.args[0].value
         elif expr.arg_kinds[0] == ARG_POS:
             encoding = expr.args[0].value
         else:
@@ -766,16 +767,9 @@ def bytes_decode_fast_path(builder: IRBuilder, expr: CallExpr, callee: RefExpr)
         if expr.arg_kinds[1] == ARG_NAMED:
             if expr.arg_names[1] == "encoding":
                 encoding = expr.args[1].value
-            elif expr.arg_names[1] == "errors":
-                errors = expr.args[1].value
-        elif expr.arg_kinds[1] == ARG_POS:
-            errors = expr.args[1].value
         else:
             return None
 
-    if errors != "strict":
-        return None
-
     normalized = encoding.lower().replace("-", "").replace("_", "")
 
     if normalized in ("utf8", "utf", "u8", "cp65001"):

mypyc/lib-rt/CPy.h

@@ -764,9 +764,9 @@ CPyTagged CPyBytes_GetItem(PyObject *o, CPyTagged index);
 PyObject *CPyBytes_Concat(PyObject *a, PyObject *b);
 PyObject *CPyBytes_Join(PyObject *sep, PyObject *iter);
 CPyTagged CPyBytes_Ord(PyObject *obj);
-PyObject *CPy_DecodeUtf8(PyObject *bytes_obj, const char *errors);
-PyObject *CPy_DecodeLatin1(PyObject *bytes_obj, const char *errors);
-PyObject *CPy_DecodeAscii(PyObject *bytes_obj, const char *errors);
+PyObject *CPy_DecodeUtf8(PyObject *bytes_obj);
+PyObject *CPy_DecodeLatin1(PyObject *bytes_obj);
+PyObject *CPy_DecodeAscii(PyObject *bytes_obj);
 
 
 int CPyBytes_Compare(PyObject *left, PyObject *right);

mypyc/lib-rt/bytes_ops.c

@@ -164,7 +164,7 @@ CPyTagged CPyBytes_Ord(PyObject *obj) {
 }
 
 
-PyObject *CPy_DecodeUtf8(PyObject *bytes_obj, const char *errors) {
+PyObject *CPy_DecodeUtf8(PyObject *bytes_obj) {
     if (!PyBytes_Check(bytes_obj)) {
         PyErr_SetString(PyExc_TypeError, "expected bytes object");
         return NULL;
@@ -173,11 +173,11 @@ PyObject *CPy_DecodeUtf8(PyObject *bytes_obj, const char *errors) {
     char *data = PyBytes_AS_STRING(bytes_obj);
     Py_ssize_t size = PyBytes_GET_SIZE(bytes_obj);
 
-    return PyUnicode_DecodeUTF8(data, size, errors);
+    return PyUnicode_DecodeUTF8(data, size, NULL);
 }
 
 
-PyObject *CPy_DecodeLatin1(PyObject *bytes_obj, const char *errors) {
+PyObject *CPy_DecodeLatin1(PyObject *bytes_obj) {
     if (!PyBytes_Check(bytes_obj)) {
         PyErr_SetString(PyExc_TypeError, "expected bytes object");
         return NULL;
@@ -186,11 +186,11 @@ PyObject *CPy_DecodeLatin1(PyObject *bytes_obj, const char *errors) {
     char *data = PyBytes_AS_STRING(bytes_obj);
     Py_ssize_t size = PyBytes_GET_SIZE(bytes_obj);
 
-    return PyUnicode_DecodeLatin1(data, size, errors);
+    return PyUnicode_DecodeLatin1(data, size, NULL);
 }
 
 
-PyObject *CPy_DecodeAscii(PyObject *bytes_obj, const char *errors) {
+PyObject *CPy_DecodeAscii(PyObject *bytes_obj) {
     if (!PyBytes_Check(bytes_obj)) {
         PyErr_SetString(PyExc_TypeError, "expected bytes object");
         return NULL;
@@ -199,5 +199,5 @@ PyObject *CPy_DecodeAscii(PyObject *bytes_obj, const char *errors) {
     char *data = PyBytes_AS_STRING(bytes_obj);
     Py_ssize_t size = PyBytes_GET_SIZE(bytes_obj);
 
-    return PyUnicode_DecodeASCII(data, size, errors);
+    return PyUnicode_DecodeASCII(data, size, NULL);
 }

mypyc/primitives/bytes_ops.py

@@ -111,23 +111,23 @@
 
 bytes_decode_utf8_strict = custom_primitive_op(
     name="decode",
-    arg_types=[bytes_rprimitive, str_rprimitive],
+    arg_types=[bytes_rprimitive],
     return_type=str_rprimitive,
     c_function_name="CPy_DecodeUtf8",
     error_kind=ERR_MAGIC,
 )
 
 bytes_decode_latin1_strict = custom_primitive_op(
     name="decode_latin1",
-    arg_types=[bytes_rprimitive, str_rprimitive],
+    arg_types=[bytes_rprimitive],
     return_type=str_rprimitive,
     c_function_name="CPy_DecodeLatin1",
     error_kind=ERR_MAGIC,
 )
 
 bytes_decode_ascii_strict = custom_primitive_op(
     name="decode_ascii",
-    arg_types=[bytes_rprimitive, str_rprimitive],
+    arg_types=[bytes_rprimitive],
     return_type=str_rprimitive,
     c_function_name="CPy_DecodeAscii",
     error_kind=ERR_MAGIC,

mypyc/test-data/irbuild-bytes.test

@@ -186,7 +186,7 @@ L0:
     b4 = r10
     return 1
 
-[case testDecodeBytes]
+[case testDecode]
 def f(b: bytes) -> None:
     b.decode()
     b.decode('utf8')
@@ -195,29 +195,35 @@ def f(b: bytes) -> None:
     b.decode('latin1', 'strict')
     b.decode('ascii')
     b.decode('latin-1')
-    b.decode('utf-8', 'ignore')  
+    b.decode('utf-8', 'ignore')
     b.decode('ascii', 'replace')
     b.decode('latin1', 'ignore')
 [out]
 def f(b):
     b :: bytes
-    r0, r1, r2, r3, r4, r5, r6, r7, r8, r9, r10, r11, r12, r13, r14, r15 :: str
+    r0, r1, r2, r3, r4, r5, r6, r7, r8, r9, r10, r11, r12, r13, r14, r15, r16, r17, r18, r19, r20, r21 :: str
 L0:
     r0 = CPy_DecodeUtf8(b)
     r1 = CPy_DecodeUtf8(b)
-    r2 = CPy_DecodeUtf8(b)
-    r3 = CPy_DecodeUtf8(b)
-    r4 = CPy_DecodeLatin1(b)
-    r5 = CPy_DecodeAscii(b)
-    r6 = CPy_DecodeLatin1(b)
-    r7 = 'utf-8'
-    r8 = 'ignore'
-    r9 = CPy_Decode(b, r7, r8)
-    r10 = 'ascii'
-    r11 = 'replace'
-    r12 = CPy_Decode(b, r10, r11)
-    r13 = 'latin1'
+    r2 = 'utf-8'
+    r3 = 'strict'
+    r4 = CPy_Decode(b, r2, r3)
+    r5 = 'utf-8'
+    r6 = 'strict'
+    r7 = CPy_Decode(b, r5, r6)
+    r8 = 'latin1'
+    r9 = 'strict'
+    r10 = CPy_Decode(b, r8, r9)
+    r11 = CPy_DecodeAscii(b)
+    r12 = CPy_DecodeLatin1(b)
+    r13 = 'utf-8'
     r14 = 'ignore'
     r15 = CPy_Decode(b, r13, r14)
+    r16 = 'ascii'
+    r17 = 'replace'
+    r18 = CPy_Decode(b, r16, r17)
+    r19 = 'latin1'
+    r20 = 'ignore'
+    r21 = CPy_Decode(b, r19, r20)
     return 1
 

mypyc/test-data/irbuild-str.test

@@ -335,14 +335,13 @@ def f(b: bytes) -> None:
 [out]
 def f(b):
     b :: bytes
-    r0, r1, r2, r3, r4, r5 :: str
+    r0, r1, r2, r3, r4 :: str
 L0:
-    r0 = CPy_Decode(b, 0, 0)
-    r1 = 'utf-8'
-    r2 = CPy_Decode(b, r1, 0)
-    r3 = 'utf-8'
-    r4 = 'backslashreplace'
-    r5 = CPy_Decode(b, r3, r4)
+    r0 = CPy_DecodeUtf8(b)
+    r1 = CPy_DecodeUtf8(b)
+    r2 = 'utf-8'
+    r3 = 'backslashreplace'
+    r4 = CPy_Decode(b, r2, r3)
     return 1
 
 [case testEncode_64bit]