Add ygnmi validator

[wenovus]

• Only compile ygot-generated code since ygnmi-generated code significantly increases compilation time and it is simpler so most issues will be caught at code-generation time.
• Make processing parallel, which reduced stage time from 7m2.326s to 1m26.953s

Tested @ openconfig/public#996 (see latest commits)

[github-actions]

Pull Request Test Coverage Report for Build 6853610196

• 0 of 5 (0.0%) changed or added relevant lines in 2 files are covered.
• 1 unchanged line in 1 file lost coverage.
• Overall coverage decreased (-0.2%) to 54.403%

---

Changes Missing Coverage	Covered Lines	Changed/Added Lines	%
cmd_gen/main.go	0	1	0.0%
commonci/github.go	0	4	0.0%

Files with Coverage Reduction	New Missed Lines	%
cmd_gen/main.go	1	30.42%

Totals
Change from base Build 6819486632:	-0.2%
Covered Lines:	939
Relevant Lines:	1726

---

💛 - Coveralls

[robshakir]

Should we keep both ygot and ygnmi? What's the logic for replacing ygot? (It seems like they do different jobs, and may have different flags/ways of generating structs -- e.g., uncompressed vs. compressed).

[wenovus]

Should we keep both ygot and ygnmi? What's the logic for replacing ygot? (It seems like they do different jobs, and may have different flags/ways of generating structs -- e.g., uncompressed vs. compressed).

ygnmi generates a full set of ygot GoStructs with the recommended set of flags for ygot. This is because users of ygnmi in general need the full ygot functionality.

However, there are small benefits to running ygot alongside ygnmi:

• We know at a glance that ygot is not broken, only ygnmi.
• We also test ygot's generator.go, but there is not much logic here.
• If ygot advances its version, ygnmi doesn't automatically advance it without a PR.

I didn't consider the last point here, and that seems to me like a good-enough reason to warrant the extra compute time, so let me add it as a second check.

[wenovus]

@robshakir @dplore PTAL.