Merge branch 'master' of https://github.com/antocuni/msgpack-python into newspec

Conflicts:
	msgpack/fallback.py
	msgpack/unpack.h
	msgpack/unpack_define.h
	msgpack/unpack_template.h

README.rst

@@ -140,6 +140,14 @@ It is also possible to pack/unpack custom data types. Here is an example for
 ``object_pairs_hook`` callback may instead be used to receive a list of
 key-value pairs.
 
+Extended types
+^^^^^^^^^^^^^^^
+
+It is also possible to pack/unpack custom data types using the msgpack feature
+of "extended types". For example, msgpack-pypy uses it to provide very fast serialization of int/float lists on top of PyPy (experimental for now):
+
+https://bitbucket.org/antocuni/msgpack-pypy/src/default/msgpack_pypy.py
+
 
 Advanced unpacking control
 ^^^^^^^^^^^^^^^^^^^^^^^^^^

msgpack/_packer.pyx

@@ -5,6 +5,7 @@ from cpython cimport *
 from libc.stdlib cimport *
 from libc.string cimport *
 from libc.limits cimport *
+from libc.stdint cimport int8_t
 
 from msgpack.exceptions import PackValueError
 
@@ -29,6 +30,7 @@ cdef extern from "pack.h":
     int msgpack_pack_raw(msgpack_packer* pk, size_t l)
     int msgpack_pack_bin(msgpack_packer* pk, size_t l)
     int msgpack_pack_raw_body(msgpack_packer* pk, char* body, size_t l)
+    int msgpack_pack_ext(msgpack_packer* pk, int8_t typecode, size_t l)
 
 cdef int DEFAULT_RECURSE_LIMIT=511
 
@@ -183,6 +185,9 @@ cdef class Packer(object):
                 for v in o:
                     ret = self._pack(v, nest_limit-1)
                     if ret != 0: break
+        elif self.handle_unknown_type(o):
+            # it means that obj was succesfully packed, so we are done
+            return 0
         elif self._default:
             o = self._default(o)
             ret = self._pack(o, nest_limit-1)
@@ -202,6 +207,13 @@ cdef class Packer(object):
             self.pk.length = 0
             return buf
 
+    def handle_unknown_type(self, obj):
+        return None
+
+    def pack_extended_type(self, typecode, data):
+        msgpack_pack_ext(&self.pk, typecode, len(data))
+        msgpack_pack_raw_body(&self.pk, data, len(data))
+
     def pack_array_header(self, size_t size):
         cdef int ret = msgpack_pack_array(&self.pk, size)
         if ret == -1:

msgpack/_unpacker.pyx

@@ -24,6 +24,7 @@ cdef extern from "unpack.h":
         PyObject* object_hook
         bint has_pairs_hook # call object_hook with k-v pairs
         PyObject* list_hook
+        PyObject* ext_type_hook
         char *encoding
         char *unicode_errors
 
@@ -45,6 +46,7 @@ cdef extern from "unpack.h":
 
 cdef inline init_ctx(unpack_context *ctx,
                      object object_hook, object object_pairs_hook, object list_hook,
+                     object ext_type_hook,
                      bint use_list, char* encoding, char* unicode_errors):
     unpack_init(ctx)
     ctx.user.use_list = use_list
@@ -71,9 +73,17 @@ cdef inline init_ctx(unpack_context *ctx,
             raise TypeError("list_hook must be a callable.")
         ctx.user.list_hook = <PyObject*>list_hook
 
+    if ext_type_hook is not None:
+        if not PyCallable_Check(ext_type_hook):
+            raise TypeError("ext_type_hook must be a callable.")
+        ctx.user.ext_type_hook = <PyObject*>ext_type_hook
+
     ctx.user.encoding = encoding
     ctx.user.unicode_errors = unicode_errors
 
+def default_read_extended_type(typecode, data):
+    raise NotImplementedError("Cannot decode extended type with typecode=%d" % typecode)
+
 def unpackb(object packed, object object_hook=None, object list_hook=None,
             bint use_list=1, encoding=None, unicode_errors="strict",
             object_pairs_hook=None,
@@ -106,7 +116,8 @@ def unpackb(object packed, object object_hook=None, object list_hook=None,
             unicode_errors = unicode_errors.encode('ascii')
         cerr = PyBytes_AsString(unicode_errors)
 
-    init_ctx(&ctx, object_hook, object_pairs_hook, list_hook, use_list, cenc, cerr)
+    init_ctx(&ctx, object_hook, object_pairs_hook, list_hook, default_read_extended_type,
+              use_list, cenc, cerr)
     ret = unpack_construct(&ctx, buf, buf_len, &off)
     if ret == 1:
         obj = unpack_data(&ctx)
@@ -248,7 +259,10 @@ cdef class Unpacker(object):
                 self.unicode_errors = unicode_errors
             cerr = PyBytes_AsString(self.unicode_errors)
 
-        init_ctx(&self.ctx, object_hook, object_pairs_hook, list_hook, use_list, cenc, cerr)
+        ext_type_hook = self.read_extended_type
+        Py_INCREF(ext_type_hook)
+        init_ctx(&self.ctx, object_hook, object_pairs_hook, list_hook,
+                  ext_type_hook, use_list, cenc, cerr)
 
     def feed(self, object next_bytes):
         """Append `next_bytes` to internal buffer."""
@@ -358,6 +372,24 @@ cdef class Unpacker(object):
         """
         return self._unpack(unpack_construct, write_bytes)
 
+    def unpack_one(self, object write_bytes=None):
+        """
+        unpack one object
+
+        If write_bytes is not None, it will be called with parts of the raw
+        message as it is unpacked.
+
+        Raises `UnpackValueError` if there are no more bytes to unpack.
+        Raises ``ExtraData`` if there are still bytes left after the unpacking.
+        """
+        try:
+            result = self.unpack()
+        except OutOfData:
+            raise UnpackValueError("Data is not enough")
+        if self.buf_head < self.buf_tail:
+            raise ExtraData(result, self.buf[self.buf_head:])
+        return result
+
     def skip(self, object write_bytes=None):
         """
         read and ignore one object, returning None
@@ -385,6 +417,9 @@ cdef class Unpacker(object):
         """
         return self._unpack(read_map_header, write_bytes)
 
+    def read_extended_type(self, typecode, data):
+        return default_read_extended_type(typecode, data)
+
     def __iter__(self):
         return self
 

msgpack/pack.h

@@ -71,6 +71,8 @@ static inline int msgpack_pack_raw(msgpack_packer* pk, size_t l);
 static inline int msgpack_pack_bin(msgpack_packer* pk, size_t l);
 static inline int msgpack_pack_raw_body(msgpack_packer* pk, const void* b, size_t l);
 
+static inline int msgpack_pack_ext(msgpack_packer* pk, int8_t typecode, size_t l);
+
 static inline int msgpack_pack_write(msgpack_packer* pk, const char *data, size_t l)
 {
     char* buf = pk->buf;

msgpack/pack_template.h

@@ -708,6 +708,66 @@ static inline int msgpack_pack_raw_body(msgpack_packer* x, const void* b, size_t
     msgpack_pack_append_buffer(x, (const unsigned char*)b, l);
 }
 
+/*
+ * Ext
+ */
+
+static inline int msgpack_pack_ext(msgpack_packer* x, int8_t typecode, size_t l)
+{
+    if (l == 1) {
+        unsigned char buf[2];
+        buf[0] = 0xd4;
+        buf[1] = (unsigned char)typecode;
+        msgpack_pack_append_buffer(x, buf, 2);
+    }
+    else if(l == 2) {
+        unsigned char buf[2];
+        buf[0] = 0xd5;
+        buf[1] = (unsigned char)typecode;
+        msgpack_pack_append_buffer(x, buf, 2);
+    }
+    else if(l == 4) {
+        unsigned char buf[2];
+        buf[0] = 0xd6;
+        buf[1] = (unsigned char)typecode;
+        msgpack_pack_append_buffer(x, buf, 2);
+    }
+    else if(l == 8) {
+        unsigned char buf[2];
+        buf[0] = 0xd7;
+        buf[1] = (unsigned char)typecode;
+        msgpack_pack_append_buffer(x, buf, 2);
+    }
+    else if(l == 16) {
+        unsigned char buf[2];
+        buf[0] = 0xd8;
+        buf[1] = (unsigned char)typecode;
+        msgpack_pack_append_buffer(x, buf, 2);
+    }
+    else if(l < 256) {
+        unsigned char buf[3];
+        buf[0] = 0xc7;
+        buf[1] = l;
+        buf[2] = (unsigned char)typecode;
+        msgpack_pack_append_buffer(x, buf, 3);
+    } else if(l < 65536) {
+        unsigned char buf[4];
+        buf[0] = 0xc8;
+        _msgpack_store16(&buf[1], (uint16_t)l);
+        buf[3] = (unsigned char)typecode;
+        msgpack_pack_append_buffer(x, buf, 4);
+    } else {
+        unsigned char buf[6];
+        buf[0] = 0xc9;
+        _msgpack_store32(&buf[1], (uint32_t)l);
+        buf[5] = (unsigned char)typecode;
+        msgpack_pack_append_buffer(x, buf, 6);
+    }
+
+}
+
+
+
 #undef msgpack_pack_append_buffer
 
 #undef TAKE8_8

msgpack/unpack.h

@@ -24,6 +24,7 @@ typedef struct unpack_user {
     PyObject *object_hook;
     bool has_pairs_hook;
     PyObject *list_hook;
+    PyObject *ext_type_hook;
     const char *encoding;
     const char *unicode_errors;
 } unpack_user;
@@ -235,4 +236,21 @@ static inline int unpack_callback_bin(unpack_user* u, const char* b, const char*
     return 0;
 }
 
+static inline int unpack_callback_ext(unpack_user* u, const char* base, const char* pos,
+                                      unsigned int lenght, msgpack_unpack_object* o)
+{
+    PyObject *py;
+    int8_t typecode = (int8_t)*pos++;
+    if (!u->ext_type_hook) {
+        PyErr_SetString(PyExc_AssertionError, "u->ext_type_hook cannot be NULL");
+        return -1;
+    }
+    // lenght also includes the typecode, so the actual data is lenght-1
+    py = PyEval_CallFunction(u->ext_type_hook, "(is#)", typecode, pos, lenght-1);
+    if (!py)
+        return -1;
+    *o = py;
+    return 0;
+}
+
 #include "unpack_template.h"

msgpack/unpack_define.h

@@ -45,10 +45,11 @@ typedef enum {
     CS_BIN_8             = 0x04,
     CS_BIN_16            = 0x05,
     CS_BIN_32            = 0x06,
-    //CS_                = 0x07,
 
-    //CS_                = 0x08,
+    CS_EXT_8             = 0x07,
-    //CS_                = 0x09,
+    CS_EXT_16            = 0x08,
+    CS_EXT_32            = 0x09,
+
     CS_FLOAT             = 0x0a,
     CS_DOUBLE            = 0x0b,
     CS_UINT_8            = 0x0c,
@@ -60,6 +61,12 @@ typedef enum {
     CS_INT_32            = 0x12,
     CS_INT_64            = 0x13,
 
+    //CS_FIXEXT1           = 0x14,
+    //CS_FIXEXT2           = 0x15,
+    //CS_FIXEXT4           = 0x16,
+    //CS_FIXEXT8           = 0x17,
+    //CS_FIXEXT16          = 0x18,
+
     CS_RAW_8             = 0x19,
     CS_RAW_16            = 0x1a,
     CS_RAW_32            = 0x1b,
@@ -70,6 +77,7 @@ typedef enum {
 
     ACS_RAW_VALUE,
     ACS_BIN_VALUE,
+    ACS_EXT_VALUE,
 } msgpack_unpack_state;
 
 

msgpack/unpack_template.h

@@ -184,9 +184,15 @@ static inline int unpack_execute(unpack_context* ctx, const char* data, size_t l
                     push_simple_value(_false);
                 case 0xc3:  // true
                     push_simple_value(_true);
-                //case 0xc7:
+                //case 0xc4:
-                //case 0xc8:
+                //case 0xc5:
-                //case 0xc9:
+                //case 0xc6:
+                case 0xc7:  // ext 8
+                    again_fixed_trail(NEXT_CS(p), 1);
+                case 0xc8:  // ext 16
+                    again_fixed_trail(NEXT_CS(p), 2);
+                case 0xc9:  // ext 32
+                    again_fixed_trail(NEXT_CS(p), 4);
                 case 0xca:  // float
                 case 0xcb:  // double
                 case 0xcc:  // unsigned int  8
@@ -198,15 +204,16 @@ static inline int unpack_execute(unpack_context* ctx, const char* data, size_t l
                 case 0xd2:  // signed int 32
                 case 0xd3:  // signed int 64
                     again_fixed_trail(NEXT_CS(p), 1 << (((unsigned int)*p) & 0x03));
-                case 0xc4:  // bin 8
+                case 0xd4:  // fixext 1
-                case 0xc5:  // bin 16
+                case 0xd5:  // fixext 2
-                case 0xc6:  // bin 32
+                case 0xd6:  // fixext 4
-                //case 0xd4:
+                case 0xd7:  // fixext 8
-                //case 0xd5:
+                    again_fixed_trail_if_zero(ACS_EXT_VALUE, 
-                //case 0xd6:  // big integer 16
+                                              (1 << (((unsigned int)*p) & 0x03))+1,
-                //case 0xd7:  // big integer 32
+                                              _ext_zero);
-                //case 0xd8:  // big float 16
+                case 0xd8:  // fixext 16
-                case 0xd9:  // raw 8
+                    again_fixed_trail_if_zero(ACS_EXT_VALUE, 16+1, _ext_zero);
+                //case 0xd9:
                 case 0xda:  // raw 16
                 case 0xdb:  // raw 32
                 case 0xdc:  // array 16
@@ -237,8 +244,16 @@ static inline int unpack_execute(unpack_context* ctx, const char* data, size_t l
             if((size_t)(pe - p) < trail) { goto _out; }
             n = p;  p += trail - 1;
             switch(cs) {
-            //case CS_
+            case CS_EXT_8:
-            //case CS_
+                again_fixed_trail_if_zero(ACS_EXT_VALUE, *(uint8_t*)n+1, _ext_zero);
+            case CS_EXT_16:
+                again_fixed_trail_if_zero(ACS_EXT_VALUE,
+                                          _msgpack_load16(uint16_t,n)+1,
+                                          _ext_zero);
+            case CS_EXT_32:
+                again_fixed_trail_if_zero(ACS_EXT_VALUE,
+                                          _msgpack_load32(uint32_t,n)+1,
+                                          _ext_zero);
             case CS_FLOAT: {
                     union { uint32_t i; float f; } mem;
                     mem.i = _msgpack_load32(uint32_t,n);
@@ -269,26 +284,6 @@ static inline int unpack_execute(unpack_context* ctx, const char* data, size_t l
             case CS_INT_64:
                 push_fixed_value(_int64, _msgpack_load64(int64_t,n));
 
-            //case CS_
-            //case CS_
-            //case CS_BIG_INT_16:
-            //    again_fixed_trail_if_zero(ACS_BIG_INT_VALUE, _msgpack_load16(uint16_t,n), _big_int_zero);
-            //case CS_BIG_INT_32:
-            //    again_fixed_trail_if_zero(ACS_BIG_INT_VALUE, _msgpack_load32(uint32_t,n), _big_int_zero);
-            //case ACS_BIG_INT_VALUE:
-            //_big_int_zero:
-            //    // FIXME
-            //    push_variable_value(_big_int, data, n, trail);
-
-            //case CS_BIG_FLOAT_16:
-            //    again_fixed_trail_if_zero(ACS_BIG_FLOAT_VALUE, _msgpack_load16(uint16_t,n), _big_float_zero);
-            //case CS_BIG_FLOAT_32:
-            //    again_fixed_trail_if_zero(ACS_BIG_FLOAT_VALUE, _msgpack_load32(uint32_t,n), _big_float_zero);
-            //case ACS_BIG_FLOAT_VALUE:
-            //_big_float_zero:
-            //    // FIXME
-            //    push_variable_value(_big_float, data, n, trail);
-
             case CS_BIN_8:
                 again_fixed_trail_if_zero(ACS_BIN_VALUE, *(uint8_t*)n, _bin_zero);
             case CS_BIN_16:
@@ -309,6 +304,10 @@ static inline int unpack_execute(unpack_context* ctx, const char* data, size_t l
             _raw_zero:
                 push_variable_value(_raw, data, n, trail);
 
+            case ACS_EXT_VALUE:
+            _ext_zero:
+                push_variable_value(_ext, data, n, trail);
+
             case CS_ARRAY_16:
                 start_container(_array, _msgpack_load16(uint16_t,n), CT_ARRAY_ITEM);
             case CS_ARRAY_32:

setup.py

@@ -92,6 +92,7 @@ if not hasattr(sys, 'pypy_version_info'):
                                  libraries=libraries,
                                  include_dirs=['.'],
                                  define_macros=macros,
+                                 extra_compile_args=['-O0'],
                                  ))
 del libraries, macros
 

test/test_extension.py

@@ -0,0 +1,64 @@
+import py
+import array
+import struct
+import msgpack
+
+def test_pack_extended_type():
+    def p(s):
+        packer = msgpack.Packer()
+        packer.pack_extended_type(0x42, s)
+        return packer.bytes()
+    assert p('A')        == '\xd4\x42A'          # fixext 1
+    assert p('AB')       == '\xd5\x42AB'         # fixext 2
+    assert p('ABCD')     == '\xd6\x42ABCD'       # fixext 4
+    assert p('ABCDEFGH') == '\xd7\x42ABCDEFGH'   # fixext 8
+    assert p('A'*16)     == '\xd8\x42' + 'A'*16  # fixext 16
+    assert p('ABC')      == '\xc7\x03\x42ABC'        # ext 8
+    assert p('A'*0x0123)        == '\xc8\x01\x23\x42' + 'A'*0x0123 # ext 16
+    assert p('A'*0x00012345)    == '\xc9\x00\x01\x23\x45\x42' + 'A'*0x00012345 # ext 32
+
+def test_unpack_extended_type():
+    class MyUnpacker(msgpack.Unpacker):
+        def read_extended_type(self, typecode, data):
+            return (typecode, data)
+
+    def u(s):
+        unpacker = MyUnpacker()
+        unpacker.feed(s)
+        return unpacker.unpack_one()
+
+    assert u('\xd4\x42A')         == (0x42, 'A')        # fixext 1
+    assert u('\xd5\x42AB')        == (0x42, 'AB')       # fixext 2
+    assert u('\xd6\x42ABCD')      == (0x42, 'ABCD')     # fixext 4
+    assert u('\xd7\x42ABCDEFGH')  == (0x42, 'ABCDEFGH') # fixext 8
+    assert u('\xd8\x42' + 'A'*16) == (0x42, 'A'*16)     # fixext 16
+    assert u('\xc7\x03\x42ABC')   == (0x42, 'ABC')      # ext 8
+    assert (u('\xc8\x01\x23\x42' + 'A'*0x0123) ==
+            (0x42, 'A'*0x0123))                         # ext 16
+    assert (u('\xc9\x00\x01\x23\x45\x42' + 'A'*0x00012345) ==
+            (0x42, 'A'*0x00012345))                     # ext 32
+
+
+def test_extension_type():
+    class MyPacker(msgpack.Packer):
+        def handle_unknown_type(self, obj):
+            if isinstance(obj, array.array):
+                typecode = 123 # application specific typecode
+                data = obj.tostring()
+                self.pack_extended_type(typecode, data)
+                return True
+
+    class MyUnpacker(msgpack.Unpacker):
+        def read_extended_type(self, typecode, data):
+            assert typecode == 123
+            obj = array.array('d')
+            obj.fromstring(data)
+            return obj
+
+    obj = [42, 'hello', array.array('d', [1.1, 2.2, 3.3])]
+    packer = MyPacker()
+    unpacker = MyUnpacker(None)
+    s = packer.pack(obj)
+    unpacker.feed(s)
+    obj2 = unpacker.unpack_one()
+    assert obj == obj2

test/test_sequnpack.py

@@ -1,9 +1,10 @@
 #!/usr/bin/env python
 # coding: utf-8
 
+import py
 import six
 from msgpack import Unpacker, BufferFull
-from msgpack.exceptions import OutOfData
+from msgpack.exceptions import OutOfData, ExtraData, UnpackValueError
 from pytest import raises
 
 
@@ -85,3 +86,15 @@ def test_readbytes():
     assert unpacker.unpack() == ord(b'a')
     assert unpacker.unpack() == ord(b'r')
 
+def test_unpack_one():
+    unpacker = Unpacker()
+    unpacker.feed('\xda\x00\x03abc')
+    assert unpacker.unpack_one() == 'abc'
+    #
+    unpacker = Unpacker()
+    unpacker.feed('\xda\x00\x03abcd')
+    py.test.raises(ExtraData, "unpacker.unpack_one()")
+    #
+    unpacker = Unpacker()
+    unpacker.feed('\xda\x00\x03ab')
+    py.test.raises(UnpackValueError, "unpacker.unpack_one()")