Eli Zaretskii <eliz@gnu.org> schrieb am Di., 26. Dez. 2017 um 05:35 Uhr:

>
> > I'm not sure which encoding speeds json.c allows us now.
>
> Well, Philipp measured it, so maybe he could share the benchmarks and
> the results.
>

json.el itself got some speedups from the time of that bug report, as
Dmitry mentioned in the bug.
My benchmark is a relatively dumb microbenchmark that calls the JSON
functions 20 times on the data files in nativejson-benchmark. Then I ran
that code through the `bench' program, which does some basic statistical
analysis. Serialization is in all cases more than 5 times faster, with a
high significance.


>
> > But the JSON structure I was talking about is not written to a file.
> > It's sent over the network (usually to a local TCP socket).
>
> In that case, encoding it is not a wasted effort, since it would have
> been otherwise encoded when we send it.
>

If that turns out to be a problem, we might consider returning a unibyte
string to avoid further encoding.