signed integer encoding

[Bernd Späth]

Hi together,

to me the BSON-spec seems to be missing encoding information regarding both signed integer types: int32 as well as int64.

As it even dictates byte-order in little-endian format it seemed to me you are trying to accomplish binary-interchangablity with this spec.

If this is the case the information 32-bit signed integer is quite worthless if everyone uses her own encoding format.

Just in case my assumtion about the goals of your spec are right, I would like to suggest you add appropriate information regarding the encoding-format of referenced integer types.

I assume the intended encoding is a two's complement as this seems the one the one used in the bson code coming along with the mongo-c-driver.

best regards,

  Bernd.

[Dwight Merriman]

that is the intent.

and yes that makes sense.