Maxim Zaks

Lets have a look at JWT (internals of JWT explained in this video if you are not familiar with it https://www.youtube.com/watch?v=_XbXkVdoG_0). As mentioned in the video at the end, a...

@krojew > You can't compare FB to JWT, since JWT will not crash your process when it contains invalid data. It all depends on how your code access the data...

@krojew I guess we just need to agree to disagree 😁. We both stated our arguments. The decision is anyways not in our hand. An I am for one quite...

Sorry, one thing I would like to add though. Regarding this: > Implementing it in a generic way is quite complicated. I don't see it as complicated at all. Let...

Ok now I think I understand your main concern. I totally agree that FlatBuffers should be completely separated from the hashing function / algorithm. The only thing that FlatBuffers could...

> You don't need to add empty space as a part of a buffer. Just add it in your custom protocol, like we can today. It's not a burden and...

> Just a bunch of logic which **should** be in the protocol. Logic which you can rely on as it is battle tested and used by many users. Custom protocol...

On more thing regarding the offsets. Currently the offsets are relative. Having a relative offset has it's benefits, but I think those benefits are not "used". The offsets are always...

@mikkelfj was your message in regards to my message or something else :)

> And absolute offsets are worse in many cases, because you need to know the start of the buffer to be able to convert them to a pointer, whereas relative...