Dmitry comments

Results 134 comments of


Dmitry

Loss of precision during serialization/parsing plain double ?

For your example I get this string: ``` {"timestamp":11000,"longitude":5E0,"latitude":2.155E2,"altitude":null,"skewAngle":5.5E0,"heading":3.027E2,"speed":8.52E1,"yaw":1.59E1,"pitch":1.23E1,"roll":1.234E2} ``` Are you _sure_ you are using Boost.JSON and not something else? Some people confuse Boost.JSON with Boost.PropertyTree.

Loss of precision during serialization/parsing plain double ?

I don't understand your issue, then.

Loss of precision during serialization/parsing plain double ?

So, #890 technically fixed this, but I'll keep the issue open until we are able to use a faster implementation than strtod.

Loss of precision during serialization/parsing plain double ?

Now that #905 was merged, this was properly fixed.

json stream: Fix out-of-bound memory access by using start+size instead of end

I fixed all asan failures I've managed to reproduce locally with the options you mentioned. Can you test current develop branch if your failures are reproducible still?

json stream: Fix out-of-bound memory access by using start+size instead of end

> hello, will I get your changes if I try Boost "Version 1.83.0 beta 1"? Yes.

json stream: Fix out-of-bound memory access by using start+size instead of end

I'm closing this for now. If the problem still occurs, please reopen.

json::string convert to std::string

Still no issues about this. I'm inclined to close this.

NaN and Infinity support

Why specifically `Infinity` and `NaN`? My cursory research shows that the only implementation that outputs these is Python.

NaN and Infinity support

While JS has them, it doesn't parse them as numbers. But I guess, this is a common enough approach to pick it. One question. I'm not sure we actually need...