codeclimate: less memory allocation

[lasiar]

Use json encoder to write codeclimate report.

[boring-cyborg]

Hey, thank you for opening your first Pull Request !

[CLAassistant]

All committers have signed the CLA.

[ldez]

Hello,

do you have a problem with the code climate reporter?
Can you explain your problem?

[lasiar]

@ldez

Hello,

do you have a problem with the code climate reporter? Can you explain your problem?

Hello, I don't have any problem, this PR just reduce allocation and memory usage.

Source code of benchmark.

result on my machine:

go test -run='^$' -bench=. -count=10

name	old time/op	new time/op	delta
Climate-10	2.79µs ± 1%	2.56µs ± 1%	-8.00%	(p=0.000 n=9+8)

name	old alloc/op	new alloc/op	delta
Climate-10	3.36kB ± 1%	2.23kB ± 3%	-33.66%	(p=0.000 n=10+9)

name	old alloc/op	new alloc/op	delta
Climate-10	34.0 ± 0%	31.0 ± 0%	-8.82%	(p=0.000 n=10+10)

[ldez]

Hello,

The problem with this kind of benchmark is in the real interest of this optimization.

The basic rule of optimization is to do it only when you have a performance problem.
There is a name for that: premature optimization.

In a general way, json.Marshal or json.Unmarshal produce more useful errors than json.NewDecoder and json.NewEncoder.

In the current situation, there's little chance of an error, and the change doesn't impact the maintainability, so we can accept this PR.
But I want to be clear that performance optimization should be led by performance problems and not micro-benchmarks.