Remove encoding option from Unpacker. (#380)

This commit is contained in:
Inada Naoki 2019-12-03 21:13:05 +09:00 committed by GitHub
parent 83ebb63c44
commit e419cd8e2d
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
4 changed files with 15 additions and 51 deletions

View file

@ -5,7 +5,7 @@ Release Date: TBD
* Remove Python 2 support from the ``msgpack/_cmsgpack``. * Remove Python 2 support from the ``msgpack/_cmsgpack``.
``msgpack/fallback`` still supports Python 2. ``msgpack/fallback`` still supports Python 2.
* Remove ``encoding`` option from the Packer. * Remove ``encoding`` option from the Packer and Unpacker.
0.6.2 0.6.2

View file

@ -31,7 +31,6 @@ cdef extern from "unpack.h":
PyObject* object_hook PyObject* object_hook
PyObject* list_hook PyObject* list_hook
PyObject* ext_hook PyObject* ext_hook
char *encoding
char *unicode_errors char *unicode_errors
Py_ssize_t max_str_len Py_ssize_t max_str_len
Py_ssize_t max_bin_len Py_ssize_t max_bin_len
@ -58,7 +57,7 @@ cdef inline init_ctx(unpack_context *ctx,
object object_hook, object object_pairs_hook, object object_hook, object object_pairs_hook,
object list_hook, object ext_hook, object list_hook, object ext_hook,
bint use_list, bint raw, bint strict_map_key, bint use_list, bint raw, bint strict_map_key,
const char* encoding, const char* unicode_errors, const char* unicode_errors,
Py_ssize_t max_str_len, Py_ssize_t max_bin_len, Py_ssize_t max_str_len, Py_ssize_t max_bin_len,
Py_ssize_t max_array_len, Py_ssize_t max_map_len, Py_ssize_t max_array_len, Py_ssize_t max_map_len,
Py_ssize_t max_ext_len): Py_ssize_t max_ext_len):
@ -99,7 +98,6 @@ cdef inline init_ctx(unpack_context *ctx,
raise TypeError("ext_hook must be a callable.") raise TypeError("ext_hook must be a callable.")
ctx.user.ext_hook = <PyObject*>ext_hook ctx.user.ext_hook = <PyObject*>ext_hook
ctx.user.encoding = encoding
ctx.user.unicode_errors = unicode_errors ctx.user.unicode_errors = unicode_errors
def default_read_extended_type(typecode, data): def default_read_extended_type(typecode, data):
@ -141,9 +139,9 @@ cdef inline int get_data_from_buffer(object obj,
1) 1)
return 1 return 1
def unpackb(object packed, object object_hook=None, object list_hook=None, def unpackb(object packed, *, object object_hook=None, object list_hook=None,
bint use_list=True, bint raw=True, bint strict_map_key=False, bint use_list=True, bint raw=True, bint strict_map_key=False,
encoding=None, unicode_errors=None, unicode_errors=None,
object_pairs_hook=None, ext_hook=ExtType, object_pairs_hook=None, ext_hook=ExtType,
Py_ssize_t max_str_len=-1, Py_ssize_t max_str_len=-1,
Py_ssize_t max_bin_len=-1, Py_ssize_t max_bin_len=-1,
@ -170,14 +168,9 @@ def unpackb(object packed, object object_hook=None, object list_hook=None,
cdef Py_buffer view cdef Py_buffer view
cdef char* buf = NULL cdef char* buf = NULL
cdef Py_ssize_t buf_len cdef Py_ssize_t buf_len
cdef const char* cenc = NULL
cdef const char* cerr = NULL cdef const char* cerr = NULL
cdef int new_protocol = 0 cdef int new_protocol = 0
if encoding is not None:
PyErr_WarnEx(DeprecationWarning, "encoding is deprecated, Use raw=False instead.", 1)
cenc = encoding
if unicode_errors is not None: if unicode_errors is not None:
cerr = unicode_errors cerr = unicode_errors
@ -196,7 +189,7 @@ def unpackb(object packed, object object_hook=None, object list_hook=None,
try: try:
init_ctx(&ctx, object_hook, object_pairs_hook, list_hook, ext_hook, init_ctx(&ctx, object_hook, object_pairs_hook, list_hook, ext_hook,
use_list, raw, strict_map_key, cenc, cerr, use_list, raw, strict_map_key, cerr,
max_str_len, max_bin_len, max_array_len, max_map_len, max_ext_len) max_str_len, max_bin_len, max_array_len, max_map_len, max_ext_len)
ret = unpack_construct(&ctx, buf, buf_len, &off) ret = unpack_construct(&ctx, buf, buf_len, &off)
finally: finally:
@ -250,8 +243,6 @@ cdef class Unpacker(object):
near future. So you must specify it explicitly for keeping backward near future. So you must specify it explicitly for keeping backward
compatibility. compatibility.
*encoding* option which is deprecated overrides this option.
:param bool strict_map_key: :param bool strict_map_key:
If true, only str or bytes are accepted for map (dict) keys. If true, only str or bytes are accepted for map (dict) keys.
It's False by default for backward-compatibility. It's False by default for backward-compatibility.
@ -290,11 +281,6 @@ cdef class Unpacker(object):
Deprecated, use *max_buffer_size* instead. Deprecated, use *max_buffer_size* instead.
Limits max size of ext type. (default: max_buffer_size or 1024*1024) Limits max size of ext type. (default: max_buffer_size or 1024*1024)
:param str encoding:
Deprecated, use ``raw=False`` instead.
Encoding used for decoding msgpack raw.
If it is None (default), msgpack raw is deserialized to Python bytes.
:param str unicode_errors: :param str unicode_errors:
Error handler used for decoding str type. (default: `'strict'`) Error handler used for decoding str type. (default: `'strict'`)
@ -330,7 +316,7 @@ cdef class Unpacker(object):
cdef Py_ssize_t read_size cdef Py_ssize_t read_size
# To maintain refcnt. # To maintain refcnt.
cdef object object_hook, object_pairs_hook, list_hook, ext_hook cdef object object_hook, object_pairs_hook, list_hook, ext_hook
cdef object encoding, unicode_errors cdef object unicode_errors
cdef Py_ssize_t max_buffer_size cdef Py_ssize_t max_buffer_size
cdef uint64_t stream_offset cdef uint64_t stream_offset
@ -341,17 +327,16 @@ cdef class Unpacker(object):
PyMem_Free(self.buf) PyMem_Free(self.buf)
self.buf = NULL self.buf = NULL
def __init__(self, file_like=None, Py_ssize_t read_size=0, def __init__(self, file_like=None, *, Py_ssize_t read_size=0,
bint use_list=True, bint raw=True, bint strict_map_key=False, bint use_list=True, bint raw=True, bint strict_map_key=False,
object object_hook=None, object object_pairs_hook=None, object list_hook=None, object object_hook=None, object object_pairs_hook=None, object list_hook=None,
encoding=None, unicode_errors=None, Py_ssize_t max_buffer_size=0, unicode_errors=None, Py_ssize_t max_buffer_size=0,
object ext_hook=ExtType, object ext_hook=ExtType,
Py_ssize_t max_str_len=-1, Py_ssize_t max_str_len=-1,
Py_ssize_t max_bin_len=-1, Py_ssize_t max_bin_len=-1,
Py_ssize_t max_array_len=-1, Py_ssize_t max_array_len=-1,
Py_ssize_t max_map_len=-1, Py_ssize_t max_map_len=-1,
Py_ssize_t max_ext_len=-1): Py_ssize_t max_ext_len=-1):
cdef const char *cenc=NULL,
cdef const char *cerr=NULL cdef const char *cerr=NULL
self.object_hook = object_hook self.object_hook = object_hook
@ -392,17 +377,12 @@ cdef class Unpacker(object):
self.buf_tail = 0 self.buf_tail = 0
self.stream_offset = 0 self.stream_offset = 0
if encoding is not None:
PyErr_WarnEx(DeprecationWarning, "encoding is deprecated, Use raw=False instead.", 1)
self.encoding = encoding
cenc = encoding
if unicode_errors is not None: if unicode_errors is not None:
self.unicode_errors = unicode_errors self.unicode_errors = unicode_errors
cerr = unicode_errors cerr = unicode_errors
init_ctx(&self.ctx, object_hook, object_pairs_hook, list_hook, init_ctx(&self.ctx, object_hook, object_pairs_hook, list_hook,
ext_hook, use_list, raw, strict_map_key, cenc, cerr, ext_hook, use_list, raw, strict_map_key, cerr,
max_str_len, max_bin_len, max_array_len, max_str_len, max_bin_len, max_array_len,
max_map_len, max_ext_len) max_map_len, max_ext_len)

View file

@ -176,8 +176,6 @@ class Unpacker(object):
near future. So you must specify it explicitly for keeping backward near future. So you must specify it explicitly for keeping backward
compatibility. compatibility.
*encoding* option which is deprecated overrides this option.
:param bool strict_map_key: :param bool strict_map_key:
If true, only str or bytes are accepted for map (dict) keys. If true, only str or bytes are accepted for map (dict) keys.
It's False by default for backward-compatibility. It's False by default for backward-compatibility.
@ -193,13 +191,10 @@ class Unpacker(object):
Unpacker calls it with a list of key-value pairs after unpacking msgpack map. Unpacker calls it with a list of key-value pairs after unpacking msgpack map.
(See also simplejson) (See also simplejson)
:param str encoding:
Encoding used for decoding msgpack raw.
If it is None (default), msgpack raw is deserialized to Python bytes.
:param str unicode_errors: :param str unicode_errors:
(deprecated) Used for decoding msgpack raw with *encoding*. The error handler for decoding unicode. (default: 'strict')
(default: `'strict'`) This option should be used only when you have msgpack data which
contains invalid UTF-8 string.
:param int max_buffer_size: :param int max_buffer_size:
Limits size of data waiting unpacked. 0 means system's INT_MAX (default). Limits size of data waiting unpacked. 0 means system's INT_MAX (default).
@ -252,18 +247,13 @@ class Unpacker(object):
def __init__(self, file_like=None, read_size=0, use_list=True, raw=True, strict_map_key=False, def __init__(self, file_like=None, read_size=0, use_list=True, raw=True, strict_map_key=False,
object_hook=None, object_pairs_hook=None, list_hook=None, object_hook=None, object_pairs_hook=None, list_hook=None,
encoding=None, unicode_errors=None, max_buffer_size=0, unicode_errors=None, max_buffer_size=0,
ext_hook=ExtType, ext_hook=ExtType,
max_str_len=-1, max_str_len=-1,
max_bin_len=-1, max_bin_len=-1,
max_array_len=-1, max_array_len=-1,
max_map_len=-1, max_map_len=-1,
max_ext_len=-1): max_ext_len=-1):
if encoding is not None:
warnings.warn(
"encoding is deprecated, Use raw=False instead.",
DeprecationWarning, stacklevel=2)
if unicode_errors is None: if unicode_errors is None:
unicode_errors = 'strict' unicode_errors = 'strict'
@ -306,7 +296,6 @@ class Unpacker(object):
self._read_size = read_size or min(self._max_buffer_size, 16*1024) self._read_size = read_size or min(self._max_buffer_size, 16*1024)
self._raw = bool(raw) self._raw = bool(raw)
self._strict_map_key = bool(strict_map_key) self._strict_map_key = bool(strict_map_key)
self._encoding = encoding
self._unicode_errors = unicode_errors self._unicode_errors = unicode_errors
self._use_list = use_list self._use_list = use_list
self._list_hook = list_hook self._list_hook = list_hook
@ -662,9 +651,7 @@ class Unpacker(object):
if execute == EX_SKIP: if execute == EX_SKIP:
return return
if typ == TYPE_RAW: if typ == TYPE_RAW:
if self._encoding is not None: if self._raw:
obj = obj.decode(self._encoding, self._unicode_errors)
elif self._raw:
obj = bytes(obj) obj = bytes(obj)
else: else:
obj = obj.decode('utf_8', self._unicode_errors) obj = obj.decode('utf_8', self._unicode_errors)

View file

@ -27,7 +27,6 @@ typedef struct unpack_user {
PyObject *object_hook; PyObject *object_hook;
PyObject *list_hook; PyObject *list_hook;
PyObject *ext_hook; PyObject *ext_hook;
const char *encoding;
const char *unicode_errors; const char *unicode_errors;
Py_ssize_t max_str_len, max_bin_len, max_array_len, max_map_len, max_ext_len; Py_ssize_t max_str_len, max_bin_len, max_array_len, max_map_len, max_ext_len;
} unpack_user; } unpack_user;
@ -232,9 +231,7 @@ static inline int unpack_callback_raw(unpack_user* u, const char* b, const char*
PyObject *py; PyObject *py;
if (u->encoding) { if (u->raw) {
py = PyUnicode_Decode(p, l, u->encoding, u->unicode_errors);
} else if (u->raw) {
py = PyBytes_FromStringAndSize(p, l); py = PyBytes_FromStringAndSize(p, l);
} else { } else {
py = PyUnicode_DecodeUTF8(p, l, u->unicode_errors); py = PyUnicode_DecodeUTF8(p, l, u->unicode_errors);