Files
asyncio-mongo_tailablecursor/asyncio_mongo/_bson/_cbsonmodule.c
2014-01-27 22:58:22 -07:00

2510 lines
78 KiB
C

/*
* Copyright 2009-2012 10gen, Inc.
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
/*
* This file contains C implementations of some of the functions
* needed by the bson module. If possible, these implementations
* should be used to speed up BSON encoding and decoding.
*/
#include "Python.h"
#include "datetime.h"
#include "buffer.h"
#include "time64.h"
#include "encoding_helpers.h"
#define _CBSON_MODULE
#include "_cbsonmodule.h"
/* New module state and initialization code.
* See the module-initialization-and-state
* section in the following doc:
* http://docs.python.org/release/3.1.3/howto/cporting.html
* which references the following pep:
* http://www.python.org/dev/peps/pep-3121/
* */
struct module_state {
PyObject* Binary;
PyObject* Code;
PyObject* ObjectId;
PyObject* DBRef;
PyObject* RECompile;
PyObject* Regex;
PyObject* UUID;
PyObject* Timestamp;
PyObject* MinKey;
PyObject* MaxKey;
PyObject* UTC;
PyTypeObject* REType;
};
/* The Py_TYPE macro was introduced in CPython 2.6 */
#ifndef Py_TYPE
#define Py_TYPE(ob) (((PyObject*)(ob))->ob_type)
#endif
#if PY_MAJOR_VERSION >= 3
#define GETSTATE(m) ((struct module_state*)PyModule_GetState(m))
#else
#define GETSTATE(m) (&_state)
static struct module_state _state;
#endif
/* Maximum number of regex flags */
#define FLAGS_SIZE 7
#define JAVA_LEGACY 5
#define CSHARP_LEGACY 6
#define BSON_MAX_SIZE 2147483647
/* The smallest possible BSON document, i.e. "{}" */
#define BSON_MIN_SIZE 5
/* Get an error class from the bson.errors module.
*
* Returns a new ref */
static PyObject* _error(char* name) {
PyObject* error;
PyObject* errors = PyImport_ImportModule("bson.errors");
if (!errors) {
return NULL;
}
error = PyObject_GetAttrString(errors, name);
Py_DECREF(errors);
return error;
}
/* Safely downcast from Py_ssize_t to int, setting an
* exception and returning -1 on error. */
static int
_downcast_and_check(Py_ssize_t size, int extra) {
if (size > BSON_MAX_SIZE || ((BSON_MAX_SIZE - extra) < size)) {
PyObject* InvalidStringData = _error("InvalidStringData");
if (InvalidStringData) {
PyErr_SetString(InvalidStringData,
"String length must be <= 2147483647");
Py_DECREF(InvalidStringData);
}
return -1;
}
return (int)size + extra;
}
static PyObject* elements_to_dict(PyObject* self, const char* string,
unsigned max, PyObject* as_class,
unsigned char tz_aware,
unsigned char uuid_subtype,
unsigned char compile_re);
static int _write_element_to_buffer(PyObject* self, buffer_t buffer,
int type_byte, PyObject* value,
unsigned char check_keys,
unsigned char uuid_subtype);
/* Date stuff */
static PyObject* datetime_from_millis(long long millis) {
/* To encode a datetime instance like datetime(9999, 12, 31, 23, 59, 59, 999999)
* we follow these steps:
* 1. Calculate a timestamp in seconds: 253402300799
* 2. Multiply that by 1000: 253402300799000
* 3. Add in microseconds divided by 1000 253402300799999
*
* (Note: BSON doesn't support microsecond accuracy, hence the rounding.)
*
* To decode we could do:
* 1. Get seconds: timestamp / 1000: 253402300799
* 2. Get micros: (timestamp % 1000) * 1000: 999000
* Resulting in datetime(9999, 12, 31, 23, 59, 59, 999000) -- the expected result
*
* Now what if the we encode (1, 1, 1, 1, 1, 1, 111111)?
* 1. and 2. gives: -62135593139000
* 3. Gives us: -62135593138889
*
* Now decode:
* 1. Gives us: -62135593138
* 2. Gives us: -889000
* Resulting in datetime(1, 1, 1, 1, 1, 2, 15888216) -- an invalid result
*
* If instead to decode we do:
* diff = ((millis % 1000) + 1000) % 1000: 111
* seconds = (millis - diff) / 1000: -62135593139
* micros = diff * 1000 111000
* Resulting in datetime(1, 1, 1, 1, 1, 1, 111000) -- the expected result
*/
int diff = (int)(((millis % 1000) + 1000) % 1000);
int microseconds = diff * 1000;
Time64_T seconds = (millis - diff) / 1000;
struct TM timeinfo;
gmtime64_r(&seconds, &timeinfo);
return PyDateTime_FromDateAndTime(timeinfo.tm_year + 1900,
timeinfo.tm_mon + 1,
timeinfo.tm_mday,
timeinfo.tm_hour,
timeinfo.tm_min,
timeinfo.tm_sec,
microseconds);
}
static long long millis_from_datetime(PyObject* datetime) {
struct TM timeinfo;
long long millis;
timeinfo.tm_year = PyDateTime_GET_YEAR(datetime) - 1900;
timeinfo.tm_mon = PyDateTime_GET_MONTH(datetime) - 1;
timeinfo.tm_mday = PyDateTime_GET_DAY(datetime);
timeinfo.tm_hour = PyDateTime_DATE_GET_HOUR(datetime);
timeinfo.tm_min = PyDateTime_DATE_GET_MINUTE(datetime);
timeinfo.tm_sec = PyDateTime_DATE_GET_SECOND(datetime);
millis = timegm64(&timeinfo) * 1000;
millis += PyDateTime_DATE_GET_MICROSECOND(datetime) / 1000;
return millis;
}
/* Just make this compatible w/ the old API. */
int buffer_write_bytes(buffer_t buffer, const char* data, int size) {
if (buffer_write(buffer, data, size)) {
PyErr_NoMemory();
return 0;
}
return 1;
}
static int write_unicode(buffer_t buffer, PyObject* py_string) {
int size;
const char* data;
PyObject* encoded = PyUnicode_AsUTF8String(py_string);
if (!encoded) {
return 0;
}
#if PY_MAJOR_VERSION >= 3
data = PyBytes_AS_STRING(encoded);
#else
data = PyString_AS_STRING(encoded);
#endif
if (!data)
goto unicodefail;
#if PY_MAJOR_VERSION >= 3
if ((size = _downcast_and_check(PyBytes_GET_SIZE(encoded), 1)) == -1)
#else
if ((size = _downcast_and_check(PyString_GET_SIZE(encoded), 1)) == -1)
#endif
goto unicodefail;
if (!buffer_write_bytes(buffer, (const char*)&size, 4))
goto unicodefail;
if (!buffer_write_bytes(buffer, data, size))
goto unicodefail;
Py_DECREF(encoded);
return 1;
unicodefail:
Py_DECREF(encoded);
return 0;
}
/* returns 0 on failure */
static int write_string(buffer_t buffer, PyObject* py_string) {
int size;
const char* data;
#if PY_MAJOR_VERSION >= 3
if (PyUnicode_Check(py_string)){
return write_unicode(buffer, py_string);
}
data = PyBytes_AsString(py_string);
#else
data = PyString_AsString(py_string);
#endif
if (!data) {
return 0;
}
#if PY_MAJOR_VERSION >= 3
if ((size = _downcast_and_check(PyBytes_Size(py_string), 1)) == -1)
#else
if ((size = _downcast_and_check(PyString_Size(py_string), 1)) == -1)
#endif
return 0;
if (!buffer_write_bytes(buffer, (const char*)&size, 4)) {
return 0;
}
if (!buffer_write_bytes(buffer, data, size)) {
return 0;
}
return 1;
}
/*
* Are we in the main interpreter or a sub-interpreter?
* Useful for deciding if we can use cached pure python
* types in mod_wsgi.
*/
static int
_in_main_interpreter(void) {
static PyInterpreterState* main_interpreter = NULL;
PyInterpreterState* interpreter;
if (main_interpreter == NULL) {
interpreter = PyInterpreterState_Head();
while (PyInterpreterState_Next(interpreter))
interpreter = PyInterpreterState_Next(interpreter);
main_interpreter = interpreter;
}
return (main_interpreter == PyThreadState_Get()->interp);
}
/*
* Get a reference to a pure python type. If we are in the
* main interpreter return the cached object, otherwise import
* the object we need and return it instead.
*/
static PyObject*
_get_object(PyObject* object, char* module_name, char* object_name) {
if (_in_main_interpreter()) {
Py_XINCREF(object);
return object;
} else {
PyObject* imported = NULL;
PyObject* module = PyImport_ImportModule(module_name);
if (!module)
return NULL;
imported = PyObject_GetAttrString(module, object_name);
Py_DECREF(module);
return imported;
}
}
/* Load a Python object to cache.
*
* Returns non-zero on failure. */
static int _load_object(PyObject** object, char* module_name, char* object_name) {
PyObject* module;
module = PyImport_ImportModule(module_name);
if (!module) {
return 1;
}
*object = PyObject_GetAttrString(module, object_name);
Py_DECREF(module);
return (*object) ? 0 : 2;
}
/* Load all Python objects to cache.
*
* Returns non-zero on failure. */
static int _load_python_objects(PyObject* module) {
PyObject* empty_string;
PyObject* compiled;
struct module_state *state = GETSTATE(module);
if (_load_object(&state->Binary, "bson.binary", "Binary") ||
_load_object(&state->Code, "bson.code", "Code") ||
_load_object(&state->ObjectId, "bson.objectid", "ObjectId") ||
_load_object(&state->DBRef, "bson.dbref", "DBRef") ||
_load_object(&state->Timestamp, "bson.timestamp", "Timestamp") ||
_load_object(&state->MinKey, "bson.min_key", "MinKey") ||
_load_object(&state->MaxKey, "bson.max_key", "MaxKey") ||
_load_object(&state->UTC, "bson.tz_util", "utc") ||
_load_object(&state->RECompile, "re", "compile") ||
_load_object(&state->Regex, "bson.regex", "Regex")) {
return 1;
}
/* If we couldn't import uuid then we must be on 2.4. Just ignore. */
if (_load_object(&state->UUID, "uuid", "UUID") == 1) {
state->UUID = NULL;
PyErr_Clear();
}
/* Reload our REType hack too. */
#if PY_MAJOR_VERSION >= 3
empty_string = PyBytes_FromString("");
#else
empty_string = PyString_FromString("");
#endif
if (empty_string == NULL) {
state->REType = NULL;
return 1;
}
compiled = PyObject_CallFunction(state->RECompile, "O", empty_string);
if (compiled == NULL) {
state->REType = NULL;
Py_DECREF(empty_string);
return 1;
}
Py_INCREF(Py_TYPE(compiled));
state->REType = Py_TYPE(compiled);
Py_DECREF(empty_string);
Py_DECREF(compiled);
return 0;
}
static int write_element_to_buffer(PyObject* self, buffer_t buffer,
int type_byte, PyObject* value,
unsigned char check_keys,
unsigned char uuid_subtype) {
int result;
if(Py_EnterRecursiveCall(" while encoding an object to BSON "))
return 0;
result = _write_element_to_buffer(self, buffer, type_byte,
value, check_keys, uuid_subtype);
Py_LeaveRecursiveCall();
return result;
}
static void
_fix_java(const char* in, char* out) {
int i, j;
for (i = 0, j = 7; i < j; i++, j--) {
out[i] = in[j];
out[j] = in[i];
}
for (i = 8, j = 15; i < j; i++, j--) {
out[i] = in[j];
out[j] = in[i];
}
}
static void
_set_cannot_encode(PyObject* value) {
PyObject* InvalidDocument = _error("InvalidDocument");
if (InvalidDocument) {
PyObject* repr = PyObject_Repr(value);
if (repr) {
#if PY_MAJOR_VERSION >= 3
PyObject* errmsg = PyUnicode_FromString("Cannot encode object: ");
#else
PyObject* errmsg = PyString_FromString("Cannot encode object: ");
#endif
if (errmsg) {
#if PY_MAJOR_VERSION >= 3
PyObject* error = PyUnicode_Concat(errmsg, repr);
if (error) {
PyErr_SetObject(InvalidDocument, error);
Py_DECREF(error);
}
Py_DECREF(errmsg);
Py_DECREF(repr);
#else
PyString_ConcatAndDel(&errmsg, repr);
if (errmsg) {
PyErr_SetObject(InvalidDocument, errmsg);
Py_DECREF(errmsg);
}
#endif
} else {
Py_DECREF(repr);
}
}
Py_DECREF(InvalidDocument);
}
}
/*
* Encode a builtin Python regular expression or our custom Regex class.
*
* Sets exception and returns 0 on failure.
*/
static int _write_regex_to_buffer(
buffer_t buffer, int type_byte, PyObject* value) {
PyObject* py_flags;
PyObject* py_pattern;
PyObject* encoded_pattern;
long int_flags;
char flags[FLAGS_SIZE];
char check_utf8 = 0;
const char* pattern_data;
int pattern_length, flags_length;
result_t status;
/*
* Both the builtin re type and our Regex class have attributes
* "flags" and "pattern".
*/
py_flags = PyObject_GetAttrString(value, "flags");
if (!py_flags) {
return 0;
}
#if PY_MAJOR_VERSION >= 3
int_flags = PyLong_AsLong(py_flags);
#else
int_flags = PyInt_AsLong(py_flags);
#endif
Py_DECREF(py_flags);
py_pattern = PyObject_GetAttrString(value, "pattern");
if (!py_pattern) {
return 0;
}
if (PyUnicode_Check(py_pattern)) {
encoded_pattern = PyUnicode_AsUTF8String(py_pattern);
Py_DECREF(py_pattern);
if (!encoded_pattern) {
return 0;
}
} else {
encoded_pattern = py_pattern;
check_utf8 = 1;
}
#if PY_MAJOR_VERSION >= 3
if (!(pattern_data = PyBytes_AsString(encoded_pattern))) {
Py_DECREF(encoded_pattern);
return 0;
}
if ((pattern_length = _downcast_and_check(PyBytes_Size(encoded_pattern), 0)) == -1) {
Py_DECREF(encoded_pattern);
return 0;
}
#else
if (!(pattern_data = PyString_AsString(encoded_pattern))) {
Py_DECREF(encoded_pattern);
return 0;
}
if ((pattern_length = _downcast_and_check(PyString_Size(encoded_pattern), 0)) == -1) {
Py_DECREF(encoded_pattern);
return 0;
}
#endif
status = check_string((const unsigned char*)pattern_data,
pattern_length, check_utf8, 1);
if (status == NOT_UTF_8) {
PyObject* InvalidStringData = _error("InvalidStringData");
if (InvalidStringData) {
PyErr_SetString(InvalidStringData,
"regex patterns must be valid UTF-8");
Py_DECREF(InvalidStringData);
}
Py_DECREF(encoded_pattern);
return 0;
} else if (status == HAS_NULL) {
PyObject* InvalidDocument = _error("InvalidDocument");
if (InvalidDocument) {
PyErr_SetString(InvalidDocument,
"regex patterns must not contain the NULL byte");
Py_DECREF(InvalidDocument);
}
Py_DECREF(encoded_pattern);
return 0;
}
if (!buffer_write_bytes(buffer, pattern_data, pattern_length + 1)) {
Py_DECREF(encoded_pattern);
return 0;
}
Py_DECREF(encoded_pattern);
flags[0] = 0;
if (int_flags & 2) {
STRCAT(flags, FLAGS_SIZE, "i");
}
if (int_flags & 4) {
STRCAT(flags, FLAGS_SIZE, "l");
}
if (int_flags & 8) {
STRCAT(flags, FLAGS_SIZE, "m");
}
if (int_flags & 16) {
STRCAT(flags, FLAGS_SIZE, "s");
}
if (int_flags & 32) {
STRCAT(flags, FLAGS_SIZE, "u");
}
if (int_flags & 64) {
STRCAT(flags, FLAGS_SIZE, "x");
}
flags_length = (int)strlen(flags) + 1;
if (!buffer_write_bytes(buffer, flags, flags_length)) {
return 0;
}
*(buffer_get_buffer(buffer) + type_byte) = 0x0B;
return 1;
}
/* TODO our platform better be little-endian w/ 4-byte ints! */
/* Write a single value to the buffer (also write its type_byte, for which
* space has already been reserved.
*
* returns 0 on failure */
static int _write_element_to_buffer(PyObject* self, buffer_t buffer,
int type_byte, PyObject* value,
unsigned char check_keys,
unsigned char uuid_subtype) {
struct module_state *state = GETSTATE(self);
/*
* Don't use PyObject_IsInstance for our custom types. It causes
* problems with python sub interpreters. Our custom types should
* have a _type_marker attribute, which we can switch on instead.
*/
if (PyObject_HasAttrString(value, "_type_marker")) {
long type;
PyObject* type_marker = PyObject_GetAttrString(value, "_type_marker");
if (type_marker == NULL)
return 0;
#if PY_MAJOR_VERSION >= 3
type = PyLong_AsLong(type_marker);
#else
type = PyInt_AsLong(type_marker);
#endif
Py_DECREF(type_marker);
/*
* Py(Long|Int)_AsLong returns -1 for error but -1 is a valid value
* so we call PyErr_Occurred to differentiate.
*
* One potential reason for an error is the user passing an invalid
* type that overrides __getattr__ (e.g. pymongo.collection.Collection)
*/
if (type == -1 && PyErr_Occurred()) {
PyErr_Clear();
_set_cannot_encode(value);
return 0;
}
switch (type) {
case 5:
{
/* Binary */
PyObject* subtype_object;
long subtype;
const char* data;
int size;
*(buffer_get_buffer(buffer) + type_byte) = 0x05;
subtype_object = PyObject_GetAttrString(value, "subtype");
if (!subtype_object) {
return 0;
}
#if PY_MAJOR_VERSION >= 3
subtype = PyLong_AsLong(subtype_object);
#else
subtype = PyInt_AsLong(subtype_object);
#endif
if (subtype == -1) {
Py_DECREF(subtype_object);
return 0;
}
#if PY_MAJOR_VERSION >= 3
size = _downcast_and_check(PyBytes_Size(value), 0);
#else
size = _downcast_and_check(PyString_Size(value), 0);
#endif
if (size == -1) {
Py_DECREF(subtype_object);
return 0;
}
Py_DECREF(subtype_object);
if (subtype == 2) {
#if PY_MAJOR_VERSION >= 3
int other_size = _downcast_and_check(PyBytes_Size(value), 4);
#else
int other_size = _downcast_and_check(PyString_Size(value), 4);
#endif
if (other_size == -1)
return 0;
if (!buffer_write_bytes(buffer, (const char*)&other_size, 4)) {
return 0;
}
if (!buffer_write_bytes(buffer, (const char*)&subtype, 1)) {
return 0;
}
}
if (!buffer_write_bytes(buffer, (const char*)&size, 4)) {
return 0;
}
if (subtype != 2) {
if (!buffer_write_bytes(buffer, (const char*)&subtype, 1)) {
return 0;
}
}
#if PY_MAJOR_VERSION >= 3
data = PyBytes_AsString(value);
#else
data = PyString_AsString(value);
#endif
if (!data) {
return 0;
}
if (!buffer_write_bytes(buffer, data, size)) {
return 0;
}
return 1;
}
case 7:
{
/* ObjectId */
const char* data;
PyObject* pystring = PyObject_GetAttrString(value, "_ObjectId__id");
if (!pystring) {
return 0;
}
#if PY_MAJOR_VERSION >= 3
data = PyBytes_AsString(pystring);
#else
data = PyString_AsString(pystring);
#endif
if (!data) {
Py_DECREF(pystring);
return 0;
}
if (!buffer_write_bytes(buffer, data, 12)) {
Py_DECREF(pystring);
return 0;
}
Py_DECREF(pystring);
*(buffer_get_buffer(buffer) + type_byte) = 0x07;
return 1;
}
case 11:
{
/* Regex */
return _write_regex_to_buffer(buffer, type_byte, value);
}
case 13:
{
/* Code */
int start_position,
length_location,
length;
PyObject* scope = PyObject_GetAttrString(value, "scope");
if (!scope) {
return 0;
}
if (!PyDict_Size(scope)) {
Py_DECREF(scope);
*(buffer_get_buffer(buffer) + type_byte) = 0x0D;
return write_string(buffer, value);
}
*(buffer_get_buffer(buffer) + type_byte) = 0x0F;
start_position = buffer_get_position(buffer);
/* save space for length */
length_location = buffer_save_space(buffer, 4);
if (length_location == -1) {
PyErr_NoMemory();
Py_DECREF(scope);
return 0;
}
if (!write_string(buffer, value)) {
Py_DECREF(scope);
return 0;
}
if (!write_dict(self, buffer, scope, 0, uuid_subtype, 0)) {
Py_DECREF(scope);
return 0;
}
Py_DECREF(scope);
length = buffer_get_position(buffer) - start_position;
memcpy(buffer_get_buffer(buffer) + length_location, &length, 4);
return 1;
}
case 17:
{
/* Timestamp */
PyObject* obj;
long i;
obj = PyObject_GetAttrString(value, "inc");
if (!obj) {
return 0;
}
#if PY_MAJOR_VERSION >= 3
i = PyLong_AsLong(obj);
#else
i = PyInt_AsLong(obj);
#endif
Py_DECREF(obj);
if (!buffer_write_bytes(buffer, (const char*)&i, 4)) {
return 0;
}
obj = PyObject_GetAttrString(value, "time");
if (!obj) {
return 0;
}
#if PY_MAJOR_VERSION >= 3
i = PyLong_AsLong(obj);
#else
i = PyInt_AsLong(obj);
#endif
Py_DECREF(obj);
if (!buffer_write_bytes(buffer, (const char*)&i, 4)) {
return 0;
}
*(buffer_get_buffer(buffer) + type_byte) = 0x11;
return 1;
}
case 100:
{
/* DBRef */
PyObject* as_doc = PyObject_CallMethod(value, "as_doc", NULL);
if (!as_doc) {
return 0;
}
if (!write_dict(self, buffer, as_doc, 0, uuid_subtype, 0)) {
Py_DECREF(as_doc);
return 0;
}
Py_DECREF(as_doc);
*(buffer_get_buffer(buffer) + type_byte) = 0x03;
return 1;
}
case 255:
{
/* MinKey */
*(buffer_get_buffer(buffer) + type_byte) = 0xFF;
return 1;
}
case 127:
{
/* MaxKey */
*(buffer_get_buffer(buffer) + type_byte) = 0x7F;
return 1;
}
}
}
/* No _type_marker attibute or not one of our types. */
if (PyBool_Check(value)) {
#if PY_MAJOR_VERSION >= 3
const long bool = PyLong_AsLong(value);
#else
const long bool = PyInt_AsLong(value);
#endif
const char c = bool ? 0x01 : 0x00;
*(buffer_get_buffer(buffer) + type_byte) = 0x08;
return buffer_write_bytes(buffer, &c, 1);
}
#if PY_MAJOR_VERSION >= 3
else if (PyLong_Check(value)) {
const long long_value = PyLong_AsLong(value);
#else
else if (PyInt_Check(value)) {
const long long_value = PyInt_AsLong(value);
#endif
const int int_value = (int)long_value;
if (PyErr_Occurred() || long_value != int_value) { /* Overflow */
long long long_long_value;
PyErr_Clear();
long_long_value = PyLong_AsLongLong(value);
if (PyErr_Occurred()) { /* Overflow AGAIN */
PyErr_SetString(PyExc_OverflowError,
"MongoDB can only handle up to 8-byte ints");
return 0;
}
*(buffer_get_buffer(buffer) + type_byte) = 0x12;
return buffer_write_bytes(buffer, (const char*)&long_long_value, 8);
}
*(buffer_get_buffer(buffer) + type_byte) = 0x10;
return buffer_write_bytes(buffer, (const char*)&int_value, 4);
#if PY_MAJOR_VERSION < 3
} else if (PyLong_Check(value)) {
const long long long_long_value = PyLong_AsLongLong(value);
if (PyErr_Occurred()) { /* Overflow */
PyErr_SetString(PyExc_OverflowError,
"MongoDB can only handle up to 8-byte ints");
return 0;
}
*(buffer_get_buffer(buffer) + type_byte) = 0x12;
return buffer_write_bytes(buffer, (const char*)&long_long_value, 8);
#endif
} else if (PyFloat_Check(value)) {
const double d = PyFloat_AsDouble(value);
*(buffer_get_buffer(buffer) + type_byte) = 0x01;
return buffer_write_bytes(buffer, (const char*)&d, 8);
} else if (value == Py_None) {
*(buffer_get_buffer(buffer) + type_byte) = 0x0A;
return 1;
} else if (PyDict_Check(value)) {
*(buffer_get_buffer(buffer) + type_byte) = 0x03;
return write_dict(self, buffer, value, check_keys, uuid_subtype, 0);
} else if (PyList_Check(value) || PyTuple_Check(value)) {
Py_ssize_t items, i;
int start_position,
length_location,
length;
char zero = 0;
*(buffer_get_buffer(buffer) + type_byte) = 0x04;
start_position = buffer_get_position(buffer);
/* save space for length */
length_location = buffer_save_space(buffer, 4);
if (length_location == -1) {
PyErr_NoMemory();
return 0;
}
if ((items = PySequence_Size(value)) > BSON_MAX_SIZE) {
PyObject* BSONError = _error("BSONError");
if (BSONError) {
PyErr_SetString(BSONError,
"Too many items to serialize.");
Py_DECREF(BSONError);
}
return 0;
}
for(i = 0; i < items; i++) {
int list_type_byte = buffer_save_space(buffer, 1);
char name[16];
PyObject* item_value;
if (list_type_byte == -1) {
PyErr_NoMemory();
return 0;
}
INT2STRING(name, (int)i);
if (!buffer_write_bytes(buffer, name, (int)strlen(name) + 1)) {
return 0;
}
if (!(item_value = PySequence_GetItem(value, i)))
return 0;
if (!write_element_to_buffer(self, buffer, list_type_byte,
item_value, check_keys, uuid_subtype)) {
Py_DECREF(item_value);
return 0;
}
Py_DECREF(item_value);
}
/* write null byte and fill in length */
if (!buffer_write_bytes(buffer, &zero, 1)) {
return 0;
}
length = buffer_get_position(buffer) - start_position;
memcpy(buffer_get_buffer(buffer) + length_location, &length, 4);
return 1;
#if PY_MAJOR_VERSION >= 3
/* Python3 special case. Store bytes as BSON binary subtype 0. */
} else if (PyBytes_Check(value)) {
int subtype = 0;
int size;
const char* data = PyBytes_AS_STRING(value);
if (!data)
return 0;
if ((size = _downcast_and_check(PyBytes_GET_SIZE(value), 0)) == -1)
return 0;
*(buffer_get_buffer(buffer) + type_byte) = 0x05;
if (!buffer_write_bytes(buffer, (const char*)&size, 4)) {
return 0;
}
if (!buffer_write_bytes(buffer, (const char*)&subtype, 1)) {
return 0;
}
if (!buffer_write_bytes(buffer, data, size)) {
return 0;
}
return 1;
#else
/* PyString_Check only works in Python 2.x. */
} else if (PyString_Check(value)) {
result_t status;
const char* data;
int size;
if (!(data = PyString_AS_STRING(value)))
return 0;
if ((size = _downcast_and_check(PyString_GET_SIZE(value), 1)) == -1)
return 0;
*(buffer_get_buffer(buffer) + type_byte) = 0x02;
status = check_string((const unsigned char*)data, size - 1, 1, 0);
if (status == NOT_UTF_8) {
PyObject* InvalidStringData = _error("InvalidStringData");
if (InvalidStringData) {
PyObject* repr = PyObject_Repr(value);
char* repr_as_cstr = repr ? PyString_AsString(repr) : NULL;
if (repr_as_cstr) {
PyObject *message = PyString_FromFormat(
"strings in documents must be valid UTF-8: %s",
repr_as_cstr);
if (message) {
PyErr_SetObject(InvalidStringData, message);
Py_DECREF(message);
}
} else {
/* repr(value) failed, use a generic message. */
PyErr_SetString(
InvalidStringData,
"strings in documents must be valid UTF-8");
}
Py_XDECREF(repr);
Py_DECREF(InvalidStringData);
}
return 0;
}
if (!buffer_write_bytes(buffer, (const char*)&size, 4)) {
return 0;
}
if (!buffer_write_bytes(buffer, data, size)) {
return 0;
}
return 1;
#endif
} else if (PyUnicode_Check(value)) {
*(buffer_get_buffer(buffer) + type_byte) = 0x02;
return write_unicode(buffer, value);
} else if (PyDateTime_Check(value)) {
long long millis;
PyObject* utcoffset = PyObject_CallMethod(value, "utcoffset", NULL);
if (utcoffset == NULL)
return 0;
if (utcoffset != Py_None) {
PyObject* result = PyNumber_Subtract(value, utcoffset);
Py_DECREF(utcoffset);
if (!result) {
return 0;
}
millis = millis_from_datetime(result);
Py_DECREF(result);
} else {
millis = millis_from_datetime(value);
}
*(buffer_get_buffer(buffer) + type_byte) = 0x09;
return buffer_write_bytes(buffer, (const char*)&millis, 8);
} else if (PyObject_TypeCheck(value, state->REType)) {
return _write_regex_to_buffer(buffer, type_byte, value);
}
/*
* Try UUID last since we have to import
* it if we're in a sub-interpreter.
*
* If we're running under python 2.4 there likely
* isn't a uuid module.
*/
if (state->UUID) {
PyObject* uuid_type = _get_object(state->UUID, "uuid", "UUID");
if (uuid_type && PyObject_IsInstance(value, uuid_type)) {
/* Just a special case of Binary above, but
* simpler to do as a separate case. */
PyObject* bytes;
/* Could be bytes, bytearray, str... */
const char* data;
/* UUID is always 16 bytes */
int size = 16;
int subtype;
Py_DECREF(uuid_type);
if (uuid_subtype == JAVA_LEGACY || uuid_subtype == CSHARP_LEGACY) {
subtype = 3;
}
else {
subtype = uuid_subtype;
}
*(buffer_get_buffer(buffer) + type_byte) = 0x05;
if (!buffer_write_bytes(buffer, (const char*)&size, 4)) {
return 0;
}
if (!buffer_write_bytes(buffer, (const char*)&subtype, 1)) {
return 0;
}
if (uuid_subtype == CSHARP_LEGACY) {
/* Legacy C# byte order */
bytes = PyObject_GetAttrString(value, "bytes_le");
}
else {
bytes = PyObject_GetAttrString(value, "bytes");
}
if (!bytes) {
return 0;
}
#if PY_MAJOR_VERSION >= 3
/* Work around http://bugs.python.org/issue7380 */
if (PyByteArray_Check(bytes)) {
data = PyByteArray_AsString(bytes);
}
else {
data = PyBytes_AsString(bytes);
}
#else
data = PyString_AsString(bytes);
#endif
if (data == NULL) {
Py_DECREF(bytes);
return 0;
}
if (uuid_subtype == JAVA_LEGACY) {
/* Store in legacy java byte order. */
char as_legacy_java[16];
_fix_java(data, as_legacy_java);
if (!buffer_write_bytes(buffer, as_legacy_java, size)) {
Py_DECREF(bytes);
return 0;
}
}
else {
if (!buffer_write_bytes(buffer, data, size)) {
Py_DECREF(bytes);
return 0;
}
}
Py_DECREF(bytes);
return 1;
} else {
Py_XDECREF(uuid_type);
}
}
/* We can't determine value's type. Fail. */
_set_cannot_encode(value);
return 0;
}
static int check_key_name(const char* name, int name_length) {
if (name_length > 0 && name[0] == '$') {
PyObject* InvalidDocument = _error("InvalidDocument");
if (InvalidDocument) {
#if PY_MAJOR_VERSION >= 3
PyObject* errmsg = PyUnicode_FromFormat(
"key '%s' must not start with '$'", name);
#else
PyObject* errmsg = PyString_FromFormat(
"key '%s' must not start with '$'", name);
#endif
if (errmsg) {
PyErr_SetObject(InvalidDocument, errmsg);
Py_DECREF(errmsg);
}
Py_DECREF(InvalidDocument);
}
return 0;
}
if (strchr(name, '.')) {
PyObject* InvalidDocument = _error("InvalidDocument");
if (InvalidDocument) {
#if PY_MAJOR_VERSION >= 3
PyObject* errmsg = PyUnicode_FromFormat(
"key '%s' must not contain '.'", name);
#else
PyObject* errmsg = PyString_FromFormat(
"key '%s' must not contain '.'", name);
#endif
if (errmsg) {
PyErr_SetObject(InvalidDocument, errmsg);
Py_DECREF(errmsg);
}
Py_DECREF(InvalidDocument);
}
return 0;
}
return 1;
}
/* Write a (key, value) pair to the buffer.
*
* Returns 0 on failure */
int write_pair(PyObject* self, buffer_t buffer, const char* name, int name_length,
PyObject* value, unsigned char check_keys,
unsigned char uuid_subtype, unsigned char allow_id) {
int type_byte;
/* Don't write any _id elements unless we're explicitly told to -
* _id has to be written first so we do so, but don't bother
* deleting it from the dictionary being written. */
if (!allow_id && strcmp(name, "_id") == 0) {
return 1;
}
type_byte = buffer_save_space(buffer, 1);
if (type_byte == -1) {
PyErr_NoMemory();
return 0;
}
if (check_keys && !check_key_name(name, name_length)) {
return 0;
}
if (!buffer_write_bytes(buffer, name, name_length + 1)) {
return 0;
}
if (!write_element_to_buffer(self, buffer, type_byte,
value, check_keys, uuid_subtype)) {
return 0;
}
return 1;
}
int decode_and_write_pair(PyObject* self, buffer_t buffer,
PyObject* key, PyObject* value,
unsigned char check_keys,
unsigned char uuid_subtype, unsigned char top_level) {
PyObject* encoded;
const char* data;
int size;
if (PyUnicode_Check(key)) {
encoded = PyUnicode_AsUTF8String(key);
if (!encoded) {
return 0;
}
#if PY_MAJOR_VERSION >= 3
if (!(data = PyBytes_AS_STRING(encoded))) {
Py_DECREF(encoded);
return 0;
}
if ((size = _downcast_and_check(PyBytes_GET_SIZE(encoded), 1)) == -1) {
Py_DECREF(encoded);
return 0;
}
#else
if (!(data = PyString_AS_STRING(encoded))) {
Py_DECREF(encoded);
return 0;
}
if ((size = _downcast_and_check(PyString_GET_SIZE(encoded), 1)) == -1) {
Py_DECREF(encoded);
return 0;
}
#endif
if (strlen(data) != (size_t)(size - 1)) {
PyObject* InvalidDocument = _error("InvalidDocument");
if (InvalidDocument) {
PyErr_SetString(InvalidDocument,
"Key names must not contain the NULL byte");
Py_DECREF(InvalidDocument);
}
Py_DECREF(encoded);
return 0;
}
#if PY_MAJOR_VERSION < 3
} else if (PyString_Check(key)) {
result_t status;
encoded = key;
Py_INCREF(encoded);
if (!(data = PyString_AS_STRING(encoded))) {
Py_DECREF(encoded);
return 0;
}
if ((size = _downcast_and_check(PyString_GET_SIZE(encoded), 1)) == -1) {
Py_DECREF(encoded);
return 0;
}
status = check_string((const unsigned char*)data, size - 1, 1, 1);
if (status == NOT_UTF_8) {
PyObject* InvalidStringData = _error("InvalidStringData");
if (InvalidStringData) {
PyErr_SetString(InvalidStringData,
"strings in documents must be valid UTF-8");
Py_DECREF(InvalidStringData);
}
Py_DECREF(encoded);
return 0;
} else if (status == HAS_NULL) {
PyObject* InvalidDocument = _error("InvalidDocument");
if (InvalidDocument) {
PyErr_SetString(InvalidDocument,
"Key names must not contain the NULL byte");
Py_DECREF(InvalidDocument);
}
Py_DECREF(encoded);
return 0;
}
#endif
} else {
PyObject* InvalidDocument = _error("InvalidDocument");
if (InvalidDocument) {
PyObject* repr = PyObject_Repr(key);
if (repr) {
#if PY_MAJOR_VERSION >= 3
PyObject* errmsg = PyUnicode_FromString(
"documents must have only string keys, key was ");
#else
PyObject* errmsg = PyString_FromString(
"documents must have only string keys, key was ");
#endif
if (errmsg) {
#if PY_MAJOR_VERSION >= 3
PyObject* error = PyUnicode_Concat(errmsg, repr);
if (error) {
PyErr_SetObject(InvalidDocument, error);
Py_DECREF(error);
}
Py_DECREF(errmsg);
Py_DECREF(repr);
#else
PyString_ConcatAndDel(&errmsg, repr);
if (errmsg) {
PyErr_SetObject(InvalidDocument, errmsg);
Py_DECREF(errmsg);
}
#endif
} else {
Py_DECREF(repr);
}
}
Py_DECREF(InvalidDocument);
}
return 0;
}
/* If top_level is True, don't allow writing _id here - it was already written. */
if (!write_pair(self, buffer, data,
size - 1, value, check_keys, uuid_subtype, !top_level)) {
Py_DECREF(encoded);
return 0;
}
Py_DECREF(encoded);
return 1;
}
/* returns 0 on failure */
int write_dict(PyObject* self, buffer_t buffer,
PyObject* dict, unsigned char check_keys,
unsigned char uuid_subtype, unsigned char top_level) {
PyObject* key;
PyObject* iter;
char zero = 0;
int length;
int length_location;
if (!PyDict_Check(dict)) {
PyObject* repr = PyObject_Repr(dict);
if (repr) {
#if PY_MAJOR_VERSION >= 3
PyObject* errmsg = PyUnicode_FromString(
"encoder expected a mapping type but got: ");
if (errmsg) {
PyObject* error = PyUnicode_Concat(errmsg, repr);
if (error) {
PyErr_SetObject(PyExc_TypeError, error);
Py_DECREF(error);
}
Py_DECREF(errmsg);
Py_DECREF(repr);
}
#else
PyObject* errmsg = PyString_FromString(
"encoder expected a mapping type but got: ");
if (errmsg) {
PyString_ConcatAndDel(&errmsg, repr);
if (errmsg) {
PyErr_SetObject(PyExc_TypeError, errmsg);
Py_DECREF(errmsg);
}
}
#endif
else {
Py_DECREF(repr);
}
} else {
PyErr_SetString(PyExc_TypeError,
"encoder expected a mapping type");
}
return 0;
}
length_location = buffer_save_space(buffer, 4);
if (length_location == -1) {
PyErr_NoMemory();
return 0;
}
/* Write _id first if this is a top level doc. */
if (top_level) {
PyObject* _id = PyDict_GetItemString(dict, "_id");
if (_id) {
if (!write_pair(self, buffer, "_id", 3,
_id, check_keys, uuid_subtype, 1)) {
return 0;
}
}
}
iter = PyObject_GetIter(dict);
if (iter == NULL) {
return 0;
}
while ((key = PyIter_Next(iter)) != NULL) {
PyObject* value = PyDict_GetItem(dict, key);
if (!value) {
PyErr_SetObject(PyExc_KeyError, key);
Py_DECREF(key);
Py_DECREF(iter);
return 0;
}
if (!decode_and_write_pair(self, buffer, key, value,
check_keys, uuid_subtype, top_level)) {
Py_DECREF(key);
Py_DECREF(iter);
return 0;
}
Py_DECREF(key);
}
Py_DECREF(iter);
/* write null byte and fill in length */
if (!buffer_write_bytes(buffer, &zero, 1)) {
return 0;
}
length = buffer_get_position(buffer) - length_location;
memcpy(buffer_get_buffer(buffer) + length_location, &length, 4);
return 1;
}
static PyObject* _cbson_dict_to_bson(PyObject* self, PyObject* args) {
PyObject* dict;
PyObject* result;
unsigned char check_keys;
unsigned char uuid_subtype;
unsigned char top_level = 1;
buffer_t buffer;
if (!PyArg_ParseTuple(args, "Obb|b", &dict,
&check_keys, &uuid_subtype, &top_level)) {
return NULL;
}
buffer = buffer_new();
if (!buffer) {
PyErr_NoMemory();
return NULL;
}
if (!write_dict(self, buffer, dict, check_keys, uuid_subtype, top_level)) {
buffer_free(buffer);
return NULL;
}
/* objectify buffer */
#if PY_MAJOR_VERSION >= 3
result = Py_BuildValue("y#", buffer_get_buffer(buffer),
buffer_get_position(buffer));
#else
result = Py_BuildValue("s#", buffer_get_buffer(buffer),
buffer_get_position(buffer));
#endif
buffer_free(buffer);
return result;
}
static PyObject* get_value(PyObject* self, const char* buffer, unsigned* position,
unsigned char type, unsigned max, PyObject* as_class,
unsigned char tz_aware, unsigned char uuid_subtype,
unsigned char compile_re) {
struct module_state *state = GETSTATE(self);
PyObject* value = NULL;
switch (type) {
case 1:
{
double d;
if (max < 8) {
goto invalid;
}
memcpy(&d, buffer + *position, 8);
value = PyFloat_FromDouble(d);
*position += 8;
break;
}
case 2:
case 14:
{
unsigned value_length;
if (max < 4) {
goto invalid;
}
memcpy(&value_length, buffer + *position, 4);
/* Encoded string length + string */
if (!value_length || max < value_length || max < 4 + value_length) {
goto invalid;
}
*position += 4;
/* Strings must end in \0 */
if (buffer[*position + value_length - 1]) {
goto invalid;
}
value = PyUnicode_DecodeUTF8(buffer + *position, value_length - 1, "strict");
if (!value) {
goto invalid;
}
*position += value_length;
break;
}
case 3:
{
PyObject* collection;
unsigned size;
if (max < 4) {
goto invalid;
}
memcpy(&size, buffer + *position, 4);
if (size < BSON_MIN_SIZE || max < size) {
goto invalid;
}
/* Check for bad eoo */
if (buffer[*position + size - 1]) {
goto invalid;
}
value = elements_to_dict(self, buffer + *position + 4,
size - 5, as_class, tz_aware, uuid_subtype,
compile_re);
if (!value) {
goto invalid;
}
/* Decoding for DBRefs */
collection = PyDict_GetItemString(value, "$ref");
if (collection) { /* DBRef */
PyObject* dbref = NULL;
PyObject* dbref_type;
PyObject* id;
PyObject* database;
Py_INCREF(collection);
PyDict_DelItemString(value, "$ref");
id = PyDict_GetItemString(value, "$id");
if (id == NULL) {
id = Py_None;
Py_INCREF(id);
} else {
Py_INCREF(id);
PyDict_DelItemString(value, "$id");
}
database = PyDict_GetItemString(value, "$db");
if (database == NULL) {
database = Py_None;
Py_INCREF(database);
} else {
Py_INCREF(database);
PyDict_DelItemString(value, "$db");
}
if ((dbref_type = _get_object(state->DBRef, "bson.dbref", "DBRef"))) {
dbref = PyObject_CallFunctionObjArgs(dbref_type, collection, id, database, value, NULL);
Py_DECREF(dbref_type);
}
Py_DECREF(value);
value = dbref;
Py_DECREF(id);
Py_DECREF(collection);
Py_DECREF(database);
}
*position += size;
break;
}
case 4:
{
unsigned size, end;
if (max < 4) {
goto invalid;
}
memcpy(&size, buffer + *position, 4);
if (size < BSON_MIN_SIZE || max < size) {
goto invalid;
}
end = *position + size - 1;
/* Check for bad eoo */
if (buffer[end]) {
goto invalid;
}
*position += 4;
value = PyList_New(0);
if (!value) {
goto invalid;
}
while (*position < end) {
PyObject* to_append;
unsigned char bson_type = (unsigned char)buffer[(*position)++];
size_t key_size = strlen(buffer + *position);
if (max < key_size) {
Py_DECREF(value);
goto invalid;
}
/* just skip the key, they're in order. */
*position += (unsigned)key_size + 1;
if (Py_EnterRecursiveCall(" while decoding a list value")) {
Py_DECREF(value);
goto invalid;
}
to_append = get_value(self, buffer, position, bson_type,
max - (unsigned)key_size,
as_class, tz_aware, uuid_subtype,
compile_re);
Py_LeaveRecursiveCall();
if (!to_append) {
Py_DECREF(value);
goto invalid;
}
PyList_Append(value, to_append);
Py_DECREF(to_append);
}
(*position)++;
break;
}
case 5:
{
PyObject* data;
PyObject* st;
PyObject* type_to_create;
unsigned length;
unsigned char subtype;
if (max < 5) {
goto invalid;
}
memcpy(&length, buffer + *position, 4);
if (max < length) {
goto invalid;
}
subtype = (unsigned char)buffer[*position + 4];
*position += 5;
if (subtype == 2 && length < 4) {
goto invalid;
}
#if PY_MAJOR_VERSION >= 3
/* Python3 special case. Decode BSON binary subtype 0 to bytes. */
if (subtype == 0) {
value = PyBytes_FromStringAndSize(buffer + *position, length);
*position += length;
break;
}
if (subtype == 2) {
data = PyBytes_FromStringAndSize(buffer + *position + 4, length - 4);
} else {
data = PyBytes_FromStringAndSize(buffer + *position, length);
}
#else
if (subtype == 2) {
data = PyString_FromStringAndSize(buffer + *position + 4, length - 4);
} else {
data = PyString_FromStringAndSize(buffer + *position, length);
}
#endif
if (!data) {
goto invalid;
}
/* Encode as UUID, not Binary */
if ((subtype == 3 || subtype == 4) && state->UUID) {
PyObject* kwargs;
PyObject* args = PyTuple_New(0);
/* UUID should always be 16 bytes */
if (!args || length != 16) {
Py_DECREF(data);
goto invalid;
}
kwargs = PyDict_New();
if (!kwargs) {
Py_DECREF(data);
Py_DECREF(args);
goto invalid;
}
/*
* From this point, we hold refs to args, kwargs, and data.
* If anything fails, goto uuiderror to clean them up.
*/
if (uuid_subtype == CSHARP_LEGACY) {
/* Legacy C# byte order */
if ((PyDict_SetItemString(kwargs, "bytes_le", data)) == -1)
goto uuiderror;
}
else {
if (uuid_subtype == JAVA_LEGACY) {
/* Convert from legacy java byte order */
char big_endian[16];
_fix_java(buffer + *position, big_endian);
/* Free the previously created PyString object */
Py_DECREF(data);
#if PY_MAJOR_VERSION >= 3
data = PyBytes_FromStringAndSize(big_endian, length);
#else
data = PyString_FromStringAndSize(big_endian, length);
#endif
if (data == NULL)
goto uuiderror;
}
if ((PyDict_SetItemString(kwargs, "bytes", data)) == -1)
goto uuiderror;
}
if ((type_to_create = _get_object(state->UUID, "uuid", "UUID"))) {
value = PyObject_Call(type_to_create, args, kwargs);
Py_DECREF(type_to_create);
}
Py_DECREF(args);
Py_DECREF(kwargs);
Py_DECREF(data);
if (!value) {
goto invalid;
}
*position += length;
break;
uuiderror:
Py_DECREF(args);
Py_DECREF(kwargs);
Py_XDECREF(data);
goto invalid;
}
#if PY_MAJOR_VERSION >= 3
st = PyLong_FromLong(subtype);
#else
st = PyInt_FromLong(subtype);
#endif
if (!st) {
Py_DECREF(data);
goto invalid;
}
if ((type_to_create = _get_object(state->Binary, "bson.binary", "Binary"))) {
value = PyObject_CallFunctionObjArgs(type_to_create, data, st, NULL);
Py_DECREF(type_to_create);
}
Py_DECREF(st);
Py_DECREF(data);
if (!value) {
goto invalid;
}
*position += length;
break;
}
case 6:
case 10:
{
value = Py_None;
Py_INCREF(value);
break;
}
case 7:
{
PyObject* objectid_type;
if (max < 12) {
goto invalid;
}
if ((objectid_type = _get_object(state->ObjectId, "bson.objectid", "ObjectId"))) {
#if PY_MAJOR_VERSION >= 3
value = PyObject_CallFunction(objectid_type, "y#", buffer + *position, 12);
#else
value = PyObject_CallFunction(objectid_type, "s#", buffer + *position, 12);
#endif
Py_DECREF(objectid_type);
}
*position += 12;
break;
}
case 8:
{
value = buffer[(*position)++] ? Py_True : Py_False;
Py_INCREF(value);
break;
}
case 9:
{
PyObject* utc_type;
PyObject* naive;
PyObject* replace;
PyObject* args;
PyObject* kwargs;
long long millis;
if (max < 8) {
goto invalid;
}
memcpy(&millis, buffer + *position, 8);
naive = datetime_from_millis(millis);
*position += 8;
if (!tz_aware) { /* In the naive case, we're done here. */
value = naive;
break;
}
if (!naive) {
goto invalid;
}
replace = PyObject_GetAttrString(naive, "replace");
Py_DECREF(naive);
if (!replace) {
goto invalid;
}
args = PyTuple_New(0);
if (!args) {
Py_DECREF(replace);
goto invalid;
}
kwargs = PyDict_New();
if (!kwargs) {
Py_DECREF(replace);
Py_DECREF(args);
goto invalid;
}
utc_type = _get_object(state->UTC, "bson.tz_util", "UTC");
if (!utc_type || PyDict_SetItemString(kwargs, "tzinfo", utc_type) == -1) {
Py_DECREF(replace);
Py_DECREF(args);
Py_DECREF(kwargs);
Py_XDECREF(utc_type);
goto invalid;
}
Py_XDECREF(utc_type);
value = PyObject_Call(replace, args, kwargs);
Py_DECREF(replace);
Py_DECREF(args);
Py_DECREF(kwargs);
break;
}
case 11:
{
PyObject* compile_func;
PyObject* pattern;
int flags;
size_t flags_length, i;
size_t pattern_length = strlen(buffer + *position);
if (pattern_length > BSON_MAX_SIZE || max < pattern_length) {
goto invalid;
}
pattern = PyUnicode_DecodeUTF8(buffer + *position, pattern_length, "strict");
if (!pattern) {
goto invalid;
}
*position += (unsigned)pattern_length + 1;
flags_length = strlen(buffer + *position);
if (flags_length > BSON_MAX_SIZE ||
(BSON_MAX_SIZE - pattern_length) < flags_length) {
Py_DECREF(pattern);
goto invalid;
}
if (max < pattern_length + flags_length) {
Py_DECREF(pattern);
goto invalid;
}
flags = 0;
for (i = 0; i < flags_length; i++) {
if (buffer[*position + i] == 'i') {
flags |= 2;
} else if (buffer[*position + i] == 'l') {
flags |= 4;
} else if (buffer[*position + i] == 'm') {
flags |= 8;
} else if (buffer[*position + i] == 's') {
flags |= 16;
} else if (buffer[*position + i] == 'u') {
flags |= 32;
} else if (buffer[*position + i] == 'x') {
flags |= 64;
}
}
*position += (unsigned)flags_length + 1;
/*
* Use re.compile() if we're configured to compile regular
* expressions, else create an instance of our Regex class.
*/
if (compile_re) {
compile_func = _get_object(state->RECompile, "re", "compile");
} else {
compile_func = _get_object(state->Regex, "bson.regex", "Regex");
}
if (compile_func) {
value = PyObject_CallFunction(compile_func, "Oi", pattern, flags);
Py_DECREF(compile_func);
}
Py_DECREF(pattern);
break;
}
case 12:
{
unsigned coll_length;
PyObject* collection;
PyObject* id = NULL;
PyObject* objectid_type;
PyObject* dbref_type;
if (max < 4) {
goto invalid;
}
memcpy(&coll_length, buffer + *position, 4);
/* Encoded string length + string + 12 byte ObjectId */
if (!coll_length || max < coll_length || max < 4 + coll_length + 12) {
goto invalid;
}
*position += 4;
/* Strings must end in \0 */
if (buffer[*position + coll_length - 1]) {
goto invalid;
}
collection = PyUnicode_DecodeUTF8(buffer + *position,
coll_length - 1, "strict");
if (!collection) {
goto invalid;
}
*position += coll_length;
if ((objectid_type = _get_object(state->ObjectId, "bson.objectid", "ObjectId"))) {
#if PY_MAJOR_VERSION >= 3
id = PyObject_CallFunction(objectid_type, "y#", buffer + *position, 12);
#else
id = PyObject_CallFunction(objectid_type, "s#", buffer + *position, 12);
#endif
Py_DECREF(objectid_type);
}
if (!id) {
Py_DECREF(collection);
goto invalid;
}
*position += 12;
if ((dbref_type = _get_object(state->DBRef, "bson.dbref", "DBRef"))) {
value = PyObject_CallFunctionObjArgs(dbref_type, collection, id, NULL);
Py_DECREF(dbref_type);
}
Py_DECREF(collection);
Py_DECREF(id);
break;
}
case 13:
{
PyObject* code;
PyObject* code_type;
unsigned value_length;
if (max < 4) {
goto invalid;
}
memcpy(&value_length, buffer + *position, 4);
/* Encoded string length + string */
if (!value_length || max < value_length || max < 4 + value_length) {
goto invalid;
}
*position += 4;
/* Strings must end in \0 */
if (buffer[*position + value_length - 1]) {
goto invalid;
}
code = PyUnicode_DecodeUTF8(buffer + *position, value_length - 1, "strict");
if (!code) {
goto invalid;
}
*position += value_length;
if ((code_type = _get_object(state->Code, "bson.code", "Code"))) {
value = PyObject_CallFunctionObjArgs(code_type, code, NULL, NULL);
Py_DECREF(code_type);
}
Py_DECREF(code);
break;
}
case 15:
{
unsigned c_w_s_size;
unsigned code_size;
unsigned scope_size;
PyObject* code;
PyObject* scope;
PyObject* code_type;
if (max < 8) {
goto invalid;
}
memcpy(&c_w_s_size, buffer + *position, 4);
*position += 4;
if (max < c_w_s_size) {
goto invalid;
}
memcpy(&code_size, buffer + *position, 4);
/* code_w_scope length + code length + code + scope length */
if (!code_size || max < code_size || max < 4 + 4 + code_size + 4) {
goto invalid;
}
*position += 4;
/* Strings must end in \0 */
if (buffer[*position + code_size - 1]) {
goto invalid;
}
code = PyUnicode_DecodeUTF8(buffer + *position, code_size - 1, "strict");
if (!code) {
goto invalid;
}
*position += code_size;
memcpy(&scope_size, buffer + *position, 4);
if (scope_size < BSON_MIN_SIZE) {
Py_DECREF(code);
goto invalid;
}
/* code length + code + scope length + scope */
if ((4 + code_size + 4 + scope_size) != c_w_s_size) {
Py_DECREF(code);
goto invalid;
}
/* Check for bad eoo */
if (buffer[*position + scope_size - 1]) {
goto invalid;
}
scope = elements_to_dict(self, buffer + *position + 4,
scope_size - 5, (PyObject*)&PyDict_Type,
tz_aware, uuid_subtype, compile_re);
if (!scope) {
Py_DECREF(code);
goto invalid;
}
*position += scope_size;
if ((code_type = _get_object(state->Code, "bson.code", "Code"))) {
value = PyObject_CallFunctionObjArgs(code_type, code, scope, NULL);
Py_DECREF(code_type);
}
Py_DECREF(code);
Py_DECREF(scope);
break;
}
case 16:
{
int i;
if (max < 4) {
goto invalid;
}
memcpy(&i, buffer + *position, 4);
#if PY_MAJOR_VERSION >= 3
value = PyLong_FromLong(i);
#else
value = PyInt_FromLong(i);
#endif
if (!value) {
goto invalid;
}
*position += 4;
break;
}
case 17:
{
unsigned int time, inc;
PyObject* timestamp_type;
if (max < 8) {
goto invalid;
}
memcpy(&inc, buffer + *position, 4);
memcpy(&time, buffer + *position + 4, 4);
if ((timestamp_type = _get_object(state->Timestamp, "bson.timestamp", "Timestamp"))) {
value = PyObject_CallFunction(timestamp_type, "II", time, inc);
Py_DECREF(timestamp_type);
}
*position += 8;
break;
}
case 18:
{
long long ll;
if (max < 8) {
goto invalid;
}
memcpy(&ll, buffer + *position, 8);
value = PyLong_FromLongLong(ll);
if (!value) {
goto invalid;
}
*position += 8;
break;
}
case 255:
{
PyObject* minkey_type = _get_object(state->MinKey, "bson.min_key", "MinKey");
if (!minkey_type)
goto invalid;
value = PyObject_CallFunctionObjArgs(minkey_type, NULL);
Py_DECREF(minkey_type);
break;
}
case 127:
{
PyObject* maxkey_type = _get_object(state->MaxKey, "bson.max_key", "MaxKey");
if (!maxkey_type)
goto invalid;
value = PyObject_CallFunctionObjArgs(maxkey_type, NULL);
Py_DECREF(maxkey_type);
break;
}
default:
{
PyObject* InvalidDocument = _error("InvalidDocument");
if (InvalidDocument) {
PyErr_SetString(InvalidDocument,
"no c decoder for this type yet");
Py_DECREF(InvalidDocument);
}
goto invalid;
}
}
if (value) {
return value;
}
invalid:
/*
* Wrap any non-InvalidBSON errors in InvalidBSON.
*/
if (PyErr_Occurred()) {
PyObject *etype, *evalue, *etrace;
PyObject *InvalidBSON;
/*
* Calling _error clears the error state, so fetch it first.
*/
PyErr_Fetch(&etype, &evalue, &etrace);
InvalidBSON = _error("InvalidBSON");
if (InvalidBSON) {
if (!PyErr_GivenExceptionMatches(etype, InvalidBSON)) {
/*
* Raise InvalidBSON(str(e)).
*/
Py_DECREF(etype);
etype = InvalidBSON;
if (evalue) {
PyObject *msg = PyObject_Str(evalue);
Py_DECREF(evalue);
evalue = msg;
}
PyErr_NormalizeException(&etype, &evalue, &etrace);
} else {
/*
* The current exception matches InvalidBSON, so we don't need
* this reference after all.
*/
Py_DECREF(InvalidBSON);
}
}
/* Steals references to args. */
PyErr_Restore(etype, evalue, etrace);
} else {
PyObject *InvalidBSON = _error("InvalidBSON");
if (InvalidBSON) {
PyErr_SetString(InvalidBSON, "invalid length or type code");
Py_DECREF(InvalidBSON);
}
}
return NULL;
}
static PyObject* _elements_to_dict(PyObject* self, const char* string,
unsigned max, PyObject* as_class,
unsigned char tz_aware,
unsigned char uuid_subtype,
unsigned char compile_re) {
unsigned position = 0;
PyObject* dict = PyObject_CallObject(as_class, NULL);
if (!dict) {
return NULL;
}
while (position < max) {
PyObject* name;
PyObject* value;
unsigned char type = (unsigned char)string[position++];
size_t name_length = strlen(string + position);
if (name_length > BSON_MAX_SIZE || position + name_length >= max) {
PyObject* InvalidBSON = _error("InvalidBSON");
if (InvalidBSON) {
PyErr_SetNone(InvalidBSON);
Py_DECREF(InvalidBSON);
}
Py_DECREF(dict);
return NULL;
}
name = PyUnicode_DecodeUTF8(string + position, name_length, "strict");
if (!name) {
Py_DECREF(dict);
return NULL;
}
position += (unsigned)name_length + 1;
value = get_value(self, string, &position, type,
max - position, as_class, tz_aware, uuid_subtype,
compile_re);
if (!value) {
Py_DECREF(name);
Py_DECREF(dict);
return NULL;
}
PyObject_SetItem(dict, name, value);
Py_DECREF(name);
Py_DECREF(value);
}
return dict;
}
static PyObject* elements_to_dict(PyObject* self, const char* string,
unsigned max, PyObject* as_class,
unsigned char tz_aware,
unsigned char uuid_subtype,
unsigned char compile_re) {
PyObject* result;
if (Py_EnterRecursiveCall(" while decoding a BSON document"))
return NULL;
result = _elements_to_dict(self, string, max,
as_class, tz_aware, uuid_subtype, compile_re);
Py_LeaveRecursiveCall();
return result;
}
static PyObject* _cbson_bson_to_dict(PyObject* self, PyObject* args) {
int size;
Py_ssize_t total_size;
const char* string;
PyObject* bson;
PyObject* as_class;
unsigned char tz_aware;
unsigned char uuid_subtype;
unsigned char compile_re;
PyObject* dict;
PyObject* remainder;
PyObject* result;
if (!PyArg_ParseTuple(
args, "OObbb", &bson, &as_class, &tz_aware, &uuid_subtype, &compile_re)) {
return NULL;
}
#if PY_MAJOR_VERSION >= 3
if (!PyBytes_Check(bson)) {
PyErr_SetString(PyExc_TypeError, "argument to _bson_to_dict must be a bytes object");
#else
if (!PyString_Check(bson)) {
PyErr_SetString(PyExc_TypeError, "argument to _bson_to_dict must be a string");
#endif
return NULL;
}
#if PY_MAJOR_VERSION >= 3
total_size = PyBytes_Size(bson);
#else
total_size = PyString_Size(bson);
#endif
if (total_size < BSON_MIN_SIZE) {
PyObject* InvalidBSON = _error("InvalidBSON");
if (InvalidBSON) {
PyErr_SetString(InvalidBSON,
"not enough data for a BSON document");
Py_DECREF(InvalidBSON);
}
return NULL;
}
#if PY_MAJOR_VERSION >= 3
string = PyBytes_AsString(bson);
#else
string = PyString_AsString(bson);
#endif
if (!string) {
return NULL;
}
memcpy(&size, string, 4);
if (size < BSON_MIN_SIZE) {
PyObject* InvalidBSON = _error("InvalidBSON");
if (InvalidBSON) {
PyErr_SetString(InvalidBSON, "invalid message size");
Py_DECREF(InvalidBSON);
}
return NULL;
}
if (total_size < size || total_size > BSON_MAX_SIZE) {
PyObject* InvalidBSON = _error("InvalidBSON");
if (InvalidBSON) {
PyErr_SetString(InvalidBSON, "objsize too large");
Py_DECREF(InvalidBSON);
}
return NULL;
}
if (size != total_size || string[size - 1]) {
PyObject* InvalidBSON = _error("InvalidBSON");
if (InvalidBSON) {
PyErr_SetString(InvalidBSON, "bad eoo");
Py_DECREF(InvalidBSON);
}
return NULL;
}
dict = elements_to_dict(self, string + 4, (unsigned)size - 5,
as_class, tz_aware, uuid_subtype, compile_re);
if (!dict) {
return NULL;
}
#if PY_MAJOR_VERSION >= 3
remainder = PyBytes_FromStringAndSize(string + size, total_size - size);
#else
remainder = PyString_FromStringAndSize(string + size, total_size - size);
#endif
if (!remainder) {
Py_DECREF(dict);
return NULL;
}
result = Py_BuildValue("OO", dict, remainder);
Py_DECREF(dict);
Py_DECREF(remainder);
return result;
}
static PyObject* _cbson_decode_all(PyObject* self, PyObject* args) {
int size;
Py_ssize_t total_size;
const char* string;
PyObject* bson;
PyObject* dict;
PyObject* result;
PyObject* as_class = (PyObject*)&PyDict_Type;
unsigned char tz_aware = 1;
unsigned char uuid_subtype = 3;
unsigned char compile_re = 1;
if (!PyArg_ParseTuple(
args, "O|Obbb",
&bson, &as_class, &tz_aware, &uuid_subtype, &compile_re)) {
return NULL;
}
#if PY_MAJOR_VERSION >= 3
if (!PyBytes_Check(bson)) {
PyErr_SetString(PyExc_TypeError, "argument to decode_all must be a bytes object");
#else
if (!PyString_Check(bson)) {
PyErr_SetString(PyExc_TypeError, "argument to decode_all must be a string");
#endif
return NULL;
}
#if PY_MAJOR_VERSION >= 3
total_size = PyBytes_Size(bson);
string = PyBytes_AsString(bson);
#else
total_size = PyString_Size(bson);
string = PyString_AsString(bson);
#endif
if (!string) {
return NULL;
}
if (!(result = PyList_New(0)))
return NULL;
while (total_size > 0) {
if (total_size < BSON_MIN_SIZE) {
PyObject* InvalidBSON = _error("InvalidBSON");
if (InvalidBSON) {
PyErr_SetString(InvalidBSON,
"not enough data for a BSON document");
Py_DECREF(InvalidBSON);
}
Py_DECREF(result);
return NULL;
}
memcpy(&size, string, 4);
if (size < BSON_MIN_SIZE) {
PyObject* InvalidBSON = _error("InvalidBSON");
if (InvalidBSON) {
PyErr_SetString(InvalidBSON, "invalid message size");
Py_DECREF(InvalidBSON);
}
Py_DECREF(result);
return NULL;
}
if (total_size < size) {
PyObject* InvalidBSON = _error("InvalidBSON");
if (InvalidBSON) {
PyErr_SetString(InvalidBSON, "objsize too large");
Py_DECREF(InvalidBSON);
}
Py_DECREF(result);
return NULL;
}
if (string[size - 1]) {
PyObject* InvalidBSON = _error("InvalidBSON");
if (InvalidBSON) {
PyErr_SetString(InvalidBSON, "bad eoo");
Py_DECREF(InvalidBSON);
}
Py_DECREF(result);
return NULL;
}
dict = elements_to_dict(self, string + 4, (unsigned)size - 5,
as_class, tz_aware, uuid_subtype, compile_re);
if (!dict) {
Py_DECREF(result);
return NULL;
}
PyList_Append(result, dict);
Py_DECREF(dict);
string += size;
total_size -= size;
}
return result;
}
static PyMethodDef _CBSONMethods[] = {
{"_dict_to_bson", _cbson_dict_to_bson, METH_VARARGS,
"convert a dictionary to a string containing its BSON representation."},
{"_bson_to_dict", _cbson_bson_to_dict, METH_VARARGS,
"convert a BSON string to a SON object."},
{"decode_all", _cbson_decode_all, METH_VARARGS,
"convert binary data to a sequence of documents."},
{NULL, NULL, 0, NULL}
};
#if PY_MAJOR_VERSION >= 3
#define INITERROR return NULL
static int _cbson_traverse(PyObject *m, visitproc visit, void *arg) {
Py_VISIT(GETSTATE(m)->Binary);
Py_VISIT(GETSTATE(m)->Code);
Py_VISIT(GETSTATE(m)->ObjectId);
Py_VISIT(GETSTATE(m)->DBRef);
Py_VISIT(GETSTATE(m)->RECompile);
Py_VISIT(GETSTATE(m)->Regex);
Py_VISIT(GETSTATE(m)->UUID);
Py_VISIT(GETSTATE(m)->Timestamp);
Py_VISIT(GETSTATE(m)->MinKey);
Py_VISIT(GETSTATE(m)->MaxKey);
Py_VISIT(GETSTATE(m)->UTC);
Py_VISIT(GETSTATE(m)->REType);
return 0;
}
static int _cbson_clear(PyObject *m) {
Py_CLEAR(GETSTATE(m)->Binary);
Py_CLEAR(GETSTATE(m)->Code);
Py_CLEAR(GETSTATE(m)->ObjectId);
Py_CLEAR(GETSTATE(m)->DBRef);
Py_CLEAR(GETSTATE(m)->RECompile);
Py_CLEAR(GETSTATE(m)->Regex);
Py_CLEAR(GETSTATE(m)->UUID);
Py_CLEAR(GETSTATE(m)->Timestamp);
Py_CLEAR(GETSTATE(m)->MinKey);
Py_CLEAR(GETSTATE(m)->MaxKey);
Py_CLEAR(GETSTATE(m)->UTC);
Py_CLEAR(GETSTATE(m)->REType);
return 0;
}
static struct PyModuleDef moduledef = {
PyModuleDef_HEAD_INIT,
"_cbson",
NULL,
sizeof(struct module_state),
_CBSONMethods,
NULL,
_cbson_traverse,
_cbson_clear,
NULL
};
PyMODINIT_FUNC
PyInit__cbson(void)
#else
#define INITERROR return
PyMODINIT_FUNC
init_cbson(void)
#endif
{
PyObject *m;
PyObject *c_api_object;
static void *_cbson_API[_cbson_API_POINTER_COUNT];
PyDateTime_IMPORT;
if (PyDateTimeAPI == NULL) {
INITERROR;
}
/* Export C API */
_cbson_API[_cbson_buffer_write_bytes_INDEX] = (void *) buffer_write_bytes;
_cbson_API[_cbson_write_dict_INDEX] = (void *) write_dict;
_cbson_API[_cbson_write_pair_INDEX] = (void *) write_pair;
_cbson_API[_cbson_decode_and_write_pair_INDEX] = (void *) decode_and_write_pair;
#if PY_VERSION_HEX >= 0x03010000
/* PyCapsule is new in python 3.1 */
c_api_object = PyCapsule_New((void *) _cbson_API, "_cbson._C_API", NULL);
#else
c_api_object = PyCObject_FromVoidPtr((void *) _cbson_API, NULL);
#endif
if (c_api_object == NULL)
INITERROR;
#if PY_MAJOR_VERSION >= 3
m = PyModule_Create(&moduledef);
#else
m = Py_InitModule("_cbson", _CBSONMethods);
#endif
if (m == NULL) {
Py_DECREF(c_api_object);
INITERROR;
}
/* Import several python objects */
if (_load_python_objects(m)) {
Py_DECREF(c_api_object);
#if PY_MAJOR_VERSION >= 3
Py_DECREF(m);
#endif
INITERROR;
}
if (PyModule_AddObject(m, "_C_API", c_api_object) < 0) {
Py_DECREF(c_api_object);
#if PY_MAJOR_VERSION >= 3
Py_DECREF(m);
#endif
INITERROR;
}
#if PY_MAJOR_VERSION >= 3
return m;
#endif
}