Change UnicodeDecodeError objects so that the 'object' attribute

is a bytes object.

Add 'y' and 'y#' format specifiers that work like 's' and 's#'
but only accept bytes objects.

5 files changed, 98 insertions, 8 deletions

diff --git a/Doc/api/utilities.tex b/Doc/api/utilities.tex
index 93e3796..fb9c909 100644
--- a/Doc/api/utilities.tex
+++ b/Doc/api/utilities.tex
@@ -424,6 +424,18 @@ whose address should be passed.
   compatible objects pass back a reference to the raw internal data
   representation.
 
+  \item[\samp{y} (bytes object)
+  {[const char *]}]
+  This variant on \samp{s} convert a Python bytes object to a C pointer to a
+  character string. The bytes object must not contain embedded NUL bytes;
+  if it does, a \exception{TypeError} exception is raised.
+
+  \item[\samp{y\#} (bytes object)
+  {[const char *, int]}]
+  This variant on \samp{s#} stores into two C variables, the first one
+  a pointer to a character string, the second one its length.  This only
+  accepts bytes objects.
+
   \item[\samp{z} (string or \code{None}) {[const char *]}]
   Like \samp{s}, but the Python object may also be \code{None}, in
   which case the C pointer is set to \NULL.
diff --git a/Doc/ext/extending.tex b/Doc/ext/extending.tex
index 2af88b5..1f3e2d5 100644
--- a/Doc/ext/extending.tex
+++ b/Doc/ext/extending.tex
@@ -802,8 +802,10 @@ Examples (to the left the call, to the right the resulting Python value):
     Py_BuildValue("i", 123)                  123
     Py_BuildValue("iii", 123, 456, 789)      (123, 456, 789)
     Py_BuildValue("s", "hello")              'hello'
+    Py_BuildValue("y", "hello")              b'hello'
     Py_BuildValue("ss", "hello", "world")    ('hello', 'world')
     Py_BuildValue("s#", "hello", 4)          'hell'
+    Py_BuildValue("y#", "hello", 4)          b'hell'
     Py_BuildValue("()")                      ()
     Py_BuildValue("(i)", 123)                (123,)
     Py_BuildValue("(ii)", 123, 456)          (123, 456)
diff --git a/Objects/exceptions.c b/Objects/exceptions.c
index e30e9df..1096bac 100644
--- a/Objects/exceptions.c
+++ b/Objects/exceptions.c
@@ -1243,6 +1243,22 @@ set_string(PyObject **attr, const char *value)
 
 
 static PyObject *
+get_bytes(PyObject *attr, const char *name)
+{
+    if (!attr) {
+        PyErr_Format(PyExc_TypeError, "%.200s attribute not set", name);
+        return NULL;
+    }
+
+    if (!PyBytes_Check(attr)) {
+        PyErr_Format(PyExc_TypeError, "%.200s attribute must be bytes", name);
+        return NULL;
+    }
+    Py_INCREF(attr);
+    return attr;
+}
+
+static PyObject *
 get_unicode(PyObject *attr, const char *name)
 {
     if (!attr) {
@@ -1280,7 +1296,7 @@ PyUnicodeEncodeError_GetObject(PyObject *exc)
 PyObject *
 PyUnicodeDecodeError_GetObject(PyObject *exc)
 {
-    return get_string(((PyUnicodeErrorObject *)exc)->object, "object");
+    return get_bytes(((PyUnicodeErrorObject *)exc)->object, "object");
 }
 
 PyObject *
@@ -1314,10 +1330,10 @@ PyUnicodeDecodeError_GetStart(PyObject *exc, Py_ssize_t *start)
 {
     if (!get_int(((PyUnicodeErrorObject *)exc)->start, start, "start")) {
         Py_ssize_t size;
-        PyObject *obj = get_string(((PyUnicodeErrorObject *)exc)->object,
+        PyObject *obj = get_bytes(((PyUnicodeErrorObject *)exc)->object,
                                    "object");
         if (!obj) return -1;
-        size = PyString_GET_SIZE(obj);
+        size = PyBytes_GET_SIZE(obj);
         if (*start<0)
             *start = 0;
         if (*start>=size)
@@ -1382,10 +1398,10 @@ PyUnicodeDecodeError_GetEnd(PyObject *exc, Py_ssize_t *end)
 {
     if (!get_int(((PyUnicodeErrorObject *)exc)->end, end, "end")) {
         Py_ssize_t size;
-        PyObject *obj = get_string(((PyUnicodeErrorObject *)exc)->object,
+        PyObject *obj = get_bytes(((PyUnicodeErrorObject *)exc)->object,
                                    "object");
         if (!obj) return -1;
-        size = PyString_GET_SIZE(obj);
+        size = PyBytes_GET_SIZE(obj);
         if (*end<1)
             *end = 1;
         if (*end>size)
@@ -1629,7 +1645,7 @@ UnicodeDecodeError_init(PyObject *self, PyObject *args, PyObject *kwds)
     if (BaseException_init((PyBaseExceptionObject *)self, args, kwds) == -1)
         return -1;
     return UnicodeError_init((PyUnicodeErrorObject *)self, args,
-                             kwds, &PyString_Type);
+                             kwds, &PyBytes_Type);
 }
 
 static PyObject *
@@ -1648,7 +1664,7 @@ UnicodeDecodeError_str(PyObject *self)
         /* FromFormat does not support %02x, so format that separately */
         char byte[4];
         PyOS_snprintf(byte, sizeof(byte), "%02x",
-                      ((int)PyString_AS_STRING(((PyUnicodeErrorObject *)self)->object)[start])&0xff);
+                      ((int)PyBytes_AS_STRING(((PyUnicodeErrorObject *)self)->object)[start])&0xff);
         return PyString_FromFormat(
             "'%.400s' codec can't decode byte 0x%s in position %zd: %.400s",
             PyString_AS_STRING(((PyUnicodeErrorObject *)self)->encoding),
@@ -1689,7 +1705,7 @@ PyUnicodeDecodeError_Create(
     assert(length < INT_MAX);
     assert(start < INT_MAX);
     assert(end < INT_MAX);
-    return PyObject_CallFunction(PyExc_UnicodeDecodeError, "ss#nns",
+    return PyObject_CallFunction(PyExc_UnicodeDecodeError, "sy#nns",
                                  encoding, object, length, start, end, reason);
 }
 
diff --git a/Python/getargs.c b/Python/getargs.c
index f7a6604..8331a18 100644
--- a/Python/getargs.c
+++ b/Python/getargs.c
@@ -819,6 +819,32 @@ convertsimple(PyObject *arg, const char **p_format, va_list *p_va, int flags,
 		break;
 	}
 
+	case 'y': {/* bytes */
+		if (*format == '#') {
+			void **p = (void **)va_arg(*p_va, char **);
+			FETCH_SIZE;
+			
+			if (PyBytes_Check(arg)) {
+				*p = PyBytes_AS_STRING(arg);
+				STORE_SIZE(PyBytes_GET_SIZE(arg));
+			}
+			else
+				return converterr("bytes", arg, msgbuf, bufsize);
+			format++;
+		} else {
+			char **p = va_arg(*p_va, char **);
+			
+			if (PyBytes_Check(arg))
+				*p = PyBytes_AS_STRING(arg);
+			else
+				return converterr("bytes", arg, msgbuf, bufsize);
+			if ((Py_ssize_t)strlen(*p) != PyBytes_Size(arg))
+				return converterr("bytes without null bytes",
+						  arg, msgbuf, bufsize);
+		}
+		break;
+	}
+
 	case 'z': {/* string, may be NULL (None) */
 		if (*format == '#') { /* any buffer-like object */
 			void **p = (void **)va_arg(*p_va, char **);
@@ -1595,6 +1621,7 @@ skipitem(const char **p_format, va_list *p_va, int flags)
 	
 	case 's': /* string */
 	case 'z': /* string or None */
+	case 'y': /* bytes */
 	case 'u': /* unicode string */
 	case 't': /* buffer, read-only */
 	case 'w': /* buffer, read-write */
diff --git a/Python/modsupport.c b/Python/modsupport.c
index af774f0..8f600dc 100644
--- a/Python/modsupport.c
+++ b/Python/modsupport.c
@@ -424,6 +424,39 @@ do_mkvalue(const char **p_format, va_list *p_va, int flags)
 			return v;
 		}
 
+		case 'y':
+		{
+			PyObject *v;
+			char *str = va_arg(*p_va, char *);
+			Py_ssize_t n;
+			if (**p_format == '#') {
+				++*p_format;
+				if (flags & FLAG_SIZE_T)
+					n = va_arg(*p_va, Py_ssize_t);
+				else
+					n = va_arg(*p_va, int);
+			}
+			else
+				n = -1;
+			if (str == NULL) {
+				v = Py_None;
+				Py_INCREF(v);
+			}
+			else {
+				if (n < 0) {
+					size_t m = strlen(str);
+					if (m > PY_SSIZE_T_MAX) {
+						PyErr_SetString(PyExc_OverflowError,
+							"string too long for Python bytes");
+						return NULL;
+					}
+					n = (Py_ssize_t)m;
+				}
+				v = PyBytes_FromStringAndSize(str, n);
+			}
+			return v;
+		}
+
 		case 'N':
 		case 'S':
 		case 'O':