On Tue, Sep 29, 2009 at 12:41, alopecoid <alopec...@gmail.com> wrote: > >> Given that the serialized bytes have to be able to *deserialize* back >> to the original messages, surely if those original messages aren't >> equal, the serialized forms would have to be different too - assuming >> we're talking about the same message type > > But, as in my example, that doesn't seem to be the case (necessarily). > Again, for example, let's say you have two messages, both of the same > type. The proto defines two optional fields, both of type variable > int64. > > Say message A poopulates both optional fields: > [1 byte tag] [3 byte value] [1 byte tag] [2 byte value] = 7 bytes > > And message B populates only one optional field: > [1 byte tag] [6 byte value] = 7 bytes
The varints are self synchronizing http://code.google.com/apis/protocolbuffers/docs/encoding.html#varints i.e. the first bit is always set in the bytes except for the last one. So the 3 byte value will have something like 1xxxxxxx 1xxxxxxx 0xxxxxxx while the 6 byte value will have have a msb set to 1 at the third byte. So they will always be different. So yes, they will be different. As Jon said: the protocol decoder needs to be able to decode it properly - a confusion between a (3byte + tag + 2 byte varint) vs. (6 byte varint) would not work. So two different messages of the same message type are always different (however two messages of different type could theoretically encode two the same). The thing you have to worry about more is the _sequence_ in which the tags are encoded. The decoder does not care in which sequence the fields are encoded, so it could be that messages with the same content can be encoded in different ways. However the encoding in the Google implementation guarantees that the fields are always in a consistent order (I guess too many people relied on the fact that messages can be used as a key/can be hashed). -h > > Couldn't these generate, by chance, the same 7 bytes? Yes, using > deserialize will correctly parse two unequal messages, but if you look > at the raw serialized byte sequences, they could actually be the same. > > > --~--~---------~--~----~------------~-------~--~----~ You received this message because you are subscribed to the Google Groups "Protocol Buffers" group. To post to this group, send email to protobuf@googlegroups.com To unsubscribe from this group, send email to protobuf+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/protobuf?hl=en -~----------~----~----~----~------~----~------~--~---
