parser: Update from go tool yacc into goyacc

[corona10]

Since Go 1.8 removed the go tool yacc command,
we should update it to use goyacc

reference: https://beta.golang.org/doc/go1.8#tool_yacc

[codecov-io]

Codecov Report

Merging #22 into master will decrease coverage by 0.07%.
The diff coverage is n/a.

@@            Coverage Diff             @@
##           master      #22      +/-   ##
==========================================
- Coverage    64.5%   64.43%   -0.08%     
==========================================
  Files          55       55              
  Lines        9740     9793      +53     
==========================================
+ Hits         6283     6310      +27     
- Misses       3010     3036      +26     
  Partials      447      447

Impacted Files	Coverage Δ
parser/y.go	67.64% <ø> (-7.57%)	⬇️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update ed3c651...cf2d78d. Read the comment docs.

[corona10]

@ncw @sbinet

PTAL

[corona10]

@ncw @sbinet
For reviewers, The code coverage results are not related with this CL.
We can ignore the result of this.

[corona10]

@ncw @sbinet
I notice that the code coverage metric is not exact.
I checked that the block which said "no-cover" is already covered on the unit test.

[sbinet]

I am on a train + mobile, so I can't easily check... But is the piece of code you say should be tested, is tested within the same package? (There's probably a way to treat a whole repo like a single code-coverage unit so that parts of pkgA that are tested from pkgB still are counted as tested. But I haven't found it.)

[corona10]

@sbinet
Yes, they are on the same package "parser".

[ncw]

I think that looks fine to me. I had a quick look through the auto generated code and it looks like it has just been tidied up. That may explain the code coverage oddities also since the line count of the generated code has increased.

[ncw]

LGTM