Open flippmoke opened 5 years ago
I think we don't need a separate bytes_values
, just reuse the string_values
. All we need is the bit that says "interpret as bytes" vs. "interpret as string", the internal encoding is the same.
Also we might want to renumber the IDs while we can and make it id 0?
@joto the weird part is that some protobuf encoder/decoders will enforce this to a string encoding, this isn't an issue in C++ but in other cases it is an issue. I suppose we could work around this by writing custom protobuf writers in each language but I could see this being a gotcha.
@flippmoke Makes sense. Doesn't cost us anything really, so that's fine. Still think we should fix order though.
This is a possible solution to #142 /cc @joto @ericfischer