mongo-ruby-driver/lib/mongo/util/bson.rb

411 lines
9.9 KiB
Ruby
Raw Normal View History

2008-12-17 16:49:06 +00:00
# --
2009-01-06 15:51:01 +00:00
# Copyright (C) 2008-2009 10gen Inc.
2008-12-17 16:43:08 +00:00
#
# This program is free software: you can redistribute it and/or modify it
# under the terms of the GNU Affero General Public License, version 3, as
# published by the Free Software Foundation.
#
# This program is distributed in the hope that it will be useful, but WITHOUT
# ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
# FITNESS FOR A PARTICULAR PURPOSE. See the GNU Affero General Public License
# for more details.
#
# You should have received a copy of the GNU Affero General Public License
# along with this program. If not, see <http://www.gnu.org/licenses/>.
2008-12-17 16:49:06 +00:00
# ++
2008-12-17 16:43:08 +00:00
require 'base64'
2008-11-22 01:00:51 +00:00
require 'mongo/util/byte_buffer'
2008-12-08 20:27:20 +00:00
require 'mongo/util/ordered_hash'
require 'mongo/objectid'
2009-01-08 19:43:30 +00:00
require 'mongo/dbref'
2008-11-22 01:00:51 +00:00
2008-12-17 16:43:08 +00:00
# A BSON seralizer/deserializer.
2008-11-22 01:00:51 +00:00
class BSON
MINKEY = -1
EOO = 0
NUMBER = 1
STRING = 2
OBJECT = 3
2008-11-22 01:00:51 +00:00
ARRAY = 4
BINARY = 5
UNDEFINED = 6
OID = 7
BOOLEAN = 8
DATE = 9
NULL = 10
2008-11-22 01:00:51 +00:00
REGEX = 11
REF = 12
CODE = 13
SYMBOL = 14
CODE_W_SCOPE = 15
NUMBER_INT = 16
MAXKEY = 127
2008-11-22 01:00:51 +00:00
def self.serialize_cstr(buf, val)
buf.put_array(val.to_s.unpack("C*") + [0])
end
def initialize(db=nil)
# db is only needed during deserialization when the data contains a DBRef
@db = db
2008-11-22 01:00:51 +00:00
@buf = ByteBuffer.new
end
def to_a
@buf.to_a
end
def serialize(obj)
raise "Document is null" unless obj
@buf.rewind
# put in a placeholder for the total size
@buf.put_int(0)
obj.each {|k, v|
type = bson_type(v, k)
case type
when STRING, CODE
serialize_string_element(@buf, k, v, type)
when NUMBER, NUMBER_INT
serialize_number_element(@buf, k, v, type)
when OBJECT
serialize_object_element(@buf, k, v)
when OID
serialize_oid_element(@buf, k, v)
2008-12-08 20:27:20 +00:00
when ARRAY
serialize_array_element(@buf, k, v)
2008-12-08 20:57:20 +00:00
when REGEX
serialize_regex_element(@buf, k, v)
2008-11-22 01:00:51 +00:00
when BOOLEAN
serialize_boolean_element(@buf, k, v)
when DATE
serialize_date_element(@buf, k, v)
when NULL
serialize_null_element(@buf, k)
2009-01-08 19:43:30 +00:00
when REF
serialize_dbref_element(@buf, k, v)
2009-01-09 23:20:53 +00:00
when SYMBOL
serialize_symbol_element(@buf, k, v)
when BINARY
serialize_binary_element(@buf, k, v)
2009-01-12 13:45:21 +00:00
when UNDEFINED, CODE_W_SCOPE
# UNDEFINED should never happen because Ruby has no UNDEFINED type
# TODO
raise "unimplemented type #{type}"
2008-11-22 01:00:51 +00:00
else
raise "unhandled type #{type}"
2008-11-22 01:00:51 +00:00
end
}
serialize_eoo_element(@buf)
@buf.put_int(@buf.size, 0)
self
end
def deserialize(buf=nil, parent=nil)
# If buf is nil, use @buf, assumed to contain already-serialized BSON.
# This is only true during testing.
@buf = ByteBuffer.new(buf.to_a) if buf
2008-11-22 01:00:51 +00:00
@buf.rewind
@buf.get_int # eat message size
doc = OrderedHash.new
2008-11-22 01:00:51 +00:00
while @buf.more?
type = @buf.get
case type
when STRING, CODE
2009-01-09 16:56:30 +00:00
key = deserialize_cstr(@buf)
2008-11-22 01:00:51 +00:00
doc[key] = deserialize_string_data(@buf)
when NUMBER
2009-01-09 16:56:30 +00:00
key = deserialize_cstr(@buf)
2008-11-22 01:00:51 +00:00
doc[key] = deserialize_number_data(@buf)
when NUMBER_INT
2009-01-09 16:56:30 +00:00
key = deserialize_cstr(@buf)
2008-11-22 01:00:51 +00:00
doc[key] = deserialize_number_int_data(@buf)
when OID
2009-01-09 16:56:30 +00:00
key = deserialize_cstr(@buf)
2008-11-22 01:00:51 +00:00
doc[key] = deserialize_oid_data(@buf)
2008-12-08 20:27:20 +00:00
when ARRAY
2009-01-09 16:56:30 +00:00
key = deserialize_cstr(@buf)
doc[key] = deserialize_array_data(@buf, doc)
2008-12-08 20:57:20 +00:00
when REGEX
2009-01-09 16:56:30 +00:00
key = deserialize_cstr(@buf)
2008-12-08 20:57:20 +00:00
doc[key] = deserialize_regex_data(@buf)
2008-11-22 01:00:51 +00:00
when OBJECT
2009-01-09 16:56:30 +00:00
key = deserialize_cstr(@buf)
doc[key] = deserialize_object_data(@buf, doc)
2008-11-22 01:00:51 +00:00
when BOOLEAN
2009-01-09 16:56:30 +00:00
key = deserialize_cstr(@buf)
2008-11-22 01:00:51 +00:00
doc[key] = deserialize_boolean_data(@buf)
when DATE
2009-01-09 16:56:30 +00:00
key = deserialize_cstr(@buf)
2008-11-22 01:00:51 +00:00
doc[key] = deserialize_date_data(@buf)
when NULL, UNDEFINED
2009-01-09 16:56:30 +00:00
key = deserialize_cstr(@buf)
2008-11-22 01:00:51 +00:00
doc[key] = nil
2009-01-08 19:43:30 +00:00
when REF
2009-01-09 16:56:30 +00:00
key = deserialize_cstr(@buf)
doc[key] = deserialize_dbref_data(@buf, key, parent)
2009-01-09 23:20:53 +00:00
when SYMBOL
key = deserialize_cstr(@buf)
doc[key] = deserialize_symbol_data(@buf, key)
when BINARY
key = deserialize_cstr(@buf)
doc[key] = deserialize_binary_data(@buf, key)
when CODE_W_SCOPE
# TODO
raise "unimplemented type #{type}"
2008-11-22 01:00:51 +00:00
when EOO
break
else
raise "Unknown type #{type}, key = #{key}"
end
end
@buf.rewind
doc
end
def hex_dump
str = ''
@buf.to_a.each_with_index { |b,i|
if (i % 8) == 0
str << "\n" if i > 0
str << '%4d: ' % i
else
str << ' '
end
str << '%02X' % b
}
str
end
def deserialize_date_data(buf)
2009-01-07 20:56:40 +00:00
millisecs = buf.get_long()
Time.at(millisecs.to_f / 1000.0) # at() takes fractional seconds
2008-11-22 01:00:51 +00:00
end
def deserialize_boolean_data(buf)
buf.get == 1
end
def deserialize_number_data(buf)
buf.get_double
end
def deserialize_number_int_data(buf)
buf.get_int
end
def deserialize_object_data(buf, parent)
2008-11-22 01:00:51 +00:00
size = buf.get_int
buf.position -= 4
BSON.new(@db).deserialize(buf.get(size), parent)
2008-11-22 01:00:51 +00:00
end
def deserialize_array_data(buf, parent)
h = deserialize_object_data(buf, parent)
2008-12-08 20:27:20 +00:00
a = []
h.each { |k, v| a[k.to_i] = v }
a
end
2008-12-08 20:57:20 +00:00
def deserialize_regex_data(buf)
2009-01-09 16:56:30 +00:00
str = deserialize_cstr(buf)
options_str = deserialize_cstr(buf)
2008-12-08 20:57:20 +00:00
options = 0
options |= Regexp::IGNORECASE if options_str.include?('i')
options |= Regexp::MULTILINE if options_str.include?('m')
options |= Regexp::EXTENDED if options_str.include?('x')
Regexp.new(str, options)
end
2008-11-22 01:00:51 +00:00
def deserialize_string_data(buf)
len = buf.get_int
bytes = buf.get(len)
bytes[0..-2].pack("C*")
end
def deserialize_oid_data(buf)
XGen::Mongo::Driver::ObjectID.new(buf.get(12))
2008-11-22 01:00:51 +00:00
end
def deserialize_dbref_data(buf, key, parent)
2009-01-09 16:56:30 +00:00
ns = deserialize_cstr(buf)
2009-01-08 19:43:30 +00:00
oid = deserialize_oid_data(buf)
XGen::Mongo::Driver::DBRef.new(parent, key, @db, ns, oid)
2009-01-08 19:43:30 +00:00
end
2009-01-09 23:20:53 +00:00
def deserialize_symbol_data(buf, key)
deserialize_cstr(buf).intern
end
def deserialize_binary_data(buf, key)
len = buf.get_int
bytes = buf.get(len)
str = ''
bytes.each { |c| str << c.chr }
str
end
2008-11-22 01:00:51 +00:00
def serialize_eoo_element(buf)
buf.put(EOO)
end
def serialize_null_element(buf, key)
buf.put(NULL)
self.class.serialize_cstr(buf, key)
end
2009-01-08 19:43:30 +00:00
def serialize_dbref_element(buf, key, val)
2009-01-09 16:56:30 +00:00
buf.put(REF)
self.class.serialize_cstr(buf, key)
self.class.serialize_cstr(buf, val.namespace)
2009-01-08 19:43:30 +00:00
buf.put_array(val.object_id.to_a)
end
2009-01-09 23:20:53 +00:00
def serialize_symbol_element(buf, key, val)
buf.put(SYMBOL)
self.class.serialize_cstr(buf, key)
self.class.serialize_cstr(buf, val)
end
def serialize_binary_element(buf, key, val)
buf.put(BINARY)
self.class.serialize_cstr(buf, key)
2009-01-12 15:48:04 +00:00
buf.put_int(val.length)
bytes = if RUBY_VERSION >= '1.9'
val.bytes.to_a
else
a = []
val.each_byte { |byte| a << byte }
a
end
buf.put_array(bytes)
end
2008-11-22 01:00:51 +00:00
def serialize_boolean_element(buf, key, val)
buf.put(BOOLEAN)
self.class.serialize_cstr(buf, key)
buf.put(val ? 1 : 0)
end
def serialize_date_element(buf, key, val)
buf.put(DATE)
self.class.serialize_cstr(buf, key)
2009-01-07 20:56:40 +00:00
millisecs = (val.to_f * 1000).to_i
buf.put_long(millisecs)
2008-11-22 01:00:51 +00:00
end
def serialize_number_element(buf, key, val, type)
buf.put(type)
self.class.serialize_cstr(buf, key)
if type == NUMBER
buf.put_double(val)
else
buf.put_int(val)
end
end
2008-12-08 20:27:20 +00:00
def serialize_object_element(buf, key, val, opcode=OBJECT)
buf.put(opcode)
2008-11-22 01:00:51 +00:00
self.class.serialize_cstr(buf, key)
buf.put_array(BSON.new.serialize(val).to_a)
2008-11-22 01:00:51 +00:00
end
2008-12-08 20:27:20 +00:00
def serialize_array_element(buf, key, val)
# Turn array into hash with integer indices as keys
h = OrderedHash.new
i = 0
val.each { |v| h[i] = v; i += 1 }
serialize_object_element(buf, key, h, ARRAY)
end
2008-12-08 20:57:20 +00:00
def serialize_regex_element(buf, key, val)
buf.put(REGEX)
self.class.serialize_cstr(buf, key)
str = val.to_s.sub(/.*?:/, '')[0..-2] # Turn "(?xxx:yyy)" into "yyy"
self.class.serialize_cstr(buf, str)
options = val.options
options_str = ''
options_str << 'i' if ((options & Regexp::IGNORECASE) != 0)
options_str << 'm' if ((options & Regexp::MULTILINE) != 0)
options_str << 'x' if ((options & Regexp::EXTENDED) != 0)
self.class.serialize_cstr(buf, options_str)
end
2008-11-22 01:00:51 +00:00
def serialize_oid_element(buf, key, val)
buf.put(OID)
self.class.serialize_cstr(buf, key)
buf.put_array(val.to_a)
2008-11-22 01:00:51 +00:00
end
def serialize_string_element(buf, key, val, type)
buf.put(type)
self.class.serialize_cstr(buf, key)
# Make a hole for the length
len_pos = buf.position
buf.put_int(0)
# Save the string
start_pos = buf.position
self.class.serialize_cstr(buf, val)
end_pos = buf.position
# Put the string size in front
buf.put_int(end_pos - start_pos, len_pos)
2008-11-22 01:00:51 +00:00
# Go back to where we were
buf.position = end_pos
end
2009-01-09 16:56:30 +00:00
def deserialize_cstr(buf)
2008-11-22 01:00:51 +00:00
chars = ""
while 1
b = buf.get
break if b == 0
chars << b.chr
end
chars
end
def bson_type(o, key)
case o
when nil
NULL
when Integer
NUMBER_INT
when Numeric
NUMBER
when XGen::Mongo::Driver::Binary # must be before String
BINARY
2008-11-22 01:00:51 +00:00
when String
# magic awful stuff - the DB requires that a where clause is sent as CODE
key == "$where" ? CODE : STRING
when Array
ARRAY
2008-12-08 20:57:20 +00:00
when Regexp
REGEX
when XGen::Mongo::Driver::ObjectID
2008-11-22 01:00:51 +00:00
OID
2009-01-08 19:43:30 +00:00
when XGen::Mongo::Driver::DBRef
REF
2008-11-22 01:00:51 +00:00
when true, false
BOOLEAN
2008-11-22 01:00:51 +00:00
when Time
DATE
when Hash
OBJECT
2009-01-09 23:20:53 +00:00
when Symbol
SYMBOL
2008-11-22 01:00:51 +00:00
else
raise "Unknown type of object: #{o.class.name}"
end
end
end