mzaks / flexbuffersswift Goto Github PK

Is a diagnostic export. Will represent the buffer as a graph with infos of type linking and color coding of broken references

JSON format deines that \u is followed by 4 characters. This is not sufficient for UTF32 characters like emojis.
An example how it suppose to work can be found here:
https://github.com/DanToml/Jay/blob/master/Tests/JayTests/ParsingTests.swift#L512

FlexBuffer has a type for fixed typed tuples triples and quadruples. This can be used to encode CGPoint, CGRect and other iOS types.

Design a mechanism where user can reserve a vector of particular type an particular size, to add values at later point.

The naming and inner functions are a total mess.
Also consider just storing the offset and length of current token and property name.

Hello, do you plan on adding support for SwiftPM + Linux?

It should be possible to at least get to the performance of JSON encoding.
Also as the data is completely known it could be possible to optimise for binary size.

Shouldn't it read
// {vec:[-100,"Fred",4.0], bar:[1,2,3], bar3:[1,2,3], foo:100, mymap:{foo:"Fred"}}
instead of
// {vec:[-100,"Fred",4.0],bar:[1,2,3],bar3:[1,2,3]foo:100,mymap{foo:"Fred"}}