def _decode_header(self):
magic = self.reader.read_uint32()
if (hex(magic) != const.SRL_MAGIC_STRING_HIGHBIT_UINT_LE):
raise exception.SrlError('bad header: invalid magic string')
doc_version_type = self.reader.read_uint8()
doc_version = (doc_version_type & 15)
doc_type = (doc_version_type >> 4) & 15
if (doc_version not in [3, 4]):
raise exception.SrlError(
'bad header: unsupported protocol version {}'.format(
doc_version))
if (doc_type < 0 or doc_type > 3):
raise exception.SrlError(
'bad header: unsupported document type {}'.format(doc_type))
header_suffix_size = self.reader.read_varint()
if header_suffix_size:
# Skipping the suffix.
self.reader._read_unpack('{0}s'.format(header_suffix_size))
self.header['version'] = doc_version
self.header['type'] = doc_type
def _decode_bytes(self):
tag = self.reader.read_uint8()
track_pos = None
if (tag & const.SRL_TRACK_BIT) != 0:
tag = tag & ~const.SRL_TRACK_BIT
track_pos = self.reader.tell() - self.body_offset
if (tag >= const.SRL_TYPE_POS_0 and tag < const.SRL_TYPE_POS_0 + 16):
return int(tag)
elif tag >= const.SRL_NEG_16 and tag < const.SRL_NEG_16 + 16:
return int(tag) - 32
elif tag == const.SRL_TYPE_FLOAT:
return self.reader.read_float()
elif tag == const.SRL_TYPE_VARINT:
return self.reader.read_varint()
elif tag == const.SRL_TYPE_UNDEF:
return None
elif (tag == const.SRL_TYPE_BINARY):
return self._decode_binary(tag)
elif (tag == const.SRL_TYPE_REFN):
return self._decode_refn(tag)
elif tag == const.SRL_TYPE_REFP:
return self._decode_refp()
elif (tag == const.SRL_TYPE_HASH):
ln = self.reader.read_varint()
return self._decode_hash(ln, track_pos)
elif (tag == const.SRL_TYPE_ARRAY):
ln = self.reader.read_varint()
return self._decode_array(ln, track_pos)
elif tag == const.SRL_TYPE_COPY:
return self._get_copy()
elif (tag >= const.SRL_TYPE_ARRAYREF_0
and tag < const.SRL_TYPE_ARRAYREF_0 + 16):
ln = tag & 15
return self._decode_array(ln, track_pos)
elif (tag >= const.SRL_TYPE_HASHREF_0
and tag < const.SRL_TYPE_HASHREF_0 + 16):
ln = tag & 15
return self._decode_hash(ln, track_pos)
elif (tag >= const.SRL_TYPE_SHORT_BINARY_0
and tag < const.SRL_TYPE_SHORT_BINARY_0 + 32):
return self._decode_short_binary(tag)
else:
raise exception.SrlError('bad tag: unsupported tag {}'.format(tag))
def _get_copy(self):
if self.copy_depth > 0:
raise exception.SrlError('bad nested copy tag: recursive copy tag found')
copy_pos = self.reader.read_varint()
copy_pos += self.body_offset-1
curr_pos = self.reader.tell()
self.reader.seek(copy_pos, os.SEEK_SET)
self.copy_depth += 1
copy = self._decode_bytes()
self.copy_depth -= 1
self.reader.seek(curr_pos, os.SEEK_SET)
return copy
def _decode_tag(self, tag):
if tag >= const.SRL_TYPE_POS_0 and tag < const.SRL_TYPE_POS_0 + 16:
return int(tag)
elif tag >= const.SRL_NEG_16 and tag < const.SRL_NEG_16 + 16:
return int(tag) - 32
elif tag == const.SRL_TYPE_FLOAT:
return self.reader.read_float()
elif tag == const.SRL_TYPE_DOUBLE:
return self.reader.read_double()
elif tag == const.SRL_TYPE_VARINT:
return self.reader.read_varint()
elif tag == const.SRL_TYPE_ZIGZAG:
return self._decode_zigzag()
elif tag == const.SRL_TYPE_UNDEF:
return None
elif tag == const.SRL_TYPE_BINARY:
return self._decode_binary(tag)
elif tag == const.SRL_TYPE_STR_UTF8:
return self._decode_str_utf8(tag)
elif tag == const.SRL_TYPE_REFN:
return self._decode_refn(tag)
elif tag == const.SRL_TYPE_REFP:
return self._decode_refp()
elif tag == const.SRL_TYPE_HASH:
ln = self.reader.read_varint()
return self._decode_hash(ln)
elif tag == const.SRL_TYPE_ARRAY:
ln = self.reader.read_varint()
return self._decode_array(ln)
elif tag == const.SRL_TYPE_OBJECT:
return self._decode_object()
elif tag == const.SRL_TYPE_OBJECTV:
return self._decode_objectv()
elif tag == const.SRL_TYPE_REGEXP:
pattern = self._decode_bytes()
flags = self._decode_bytes()
python_flags = reduce(
lambda x, y: x | re.RegexFlag[str(y).upper()], flags, 0)
return re.compile(pattern, python_flags)
elif tag == const.SRL_TYPE_CANONICAL_UNDEF:
return None
elif tag == const.SRL_TYPE_FALSE:
return False
elif tag == const.SRL_TYPE_TRUE:
return True
elif tag == const.SRL_TYPE_COPY:
return self._get_copy()
elif tag >= const.SRL_TYPE_ARRAYREF_0 and tag < const.SRL_TYPE_ARRAYREF_0 + 16:
ln = tag & 15
return self._decode_array(ln)
elif tag >= const.SRL_TYPE_HASHREF_0 and tag < const.SRL_TYPE_HASHREF_0 + 16:
ln = tag & 15
return self._decode_hash(ln)
elif tag >= const.SRL_TYPE_SHORT_BINARY_0 and tag < const.SRL_TYPE_SHORT_BINARY_0 + 32:
# Note: self.reader.read_str() automatically decodes as utf-8.
# Thus, this returns a string, instead of returning bytes.
return self._decode_short_binary(tag)
else:
raise exception.SrlError('bad tag: unsupported tag {}'.format(tag))
