build: use c8 for coverage

[JustinBeckwith]

No description provided.

[codecov]

Codecov Report

Merging #236 into master will not change coverage.
The diff coverage is n/a.

@@          Coverage Diff           @@
##           master    #236   +/-   ##
======================================
  Coverage    34.6%   34.6%           
======================================
  Files          39      39           
  Lines        6768    6768           
======================================
  Hits         2342    2342           
  Misses       4426    4426

---

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 2071b7d...b43849d. Read the comment docs.

[bcoe]

@JustinBeckwith the root cause of this significant drop in coverage is that --all now works for c8, and there's a pretty large swath of this codebase we're not exercising.

I was hoping that adding coverage for samples might help boost our numbers, but it only brought things up 0.10%...

I think this problem will be addressed when we convert this codebase to TypeScript, so we might just want to land the lower coverage.

---

One other thing, I'm tempted to start running samples like this:

    "samples-test": "c8 npm run samples-test-suite",
    "samples-test-suite": "cd samples/ && npm link ../ && npm test && cd ../",

Which means we then do get merged coverage reports in our repos.

I'd also like to remove the **/samples ignore rule from our .nycrc, replacing it with:

**/samples/test
**/samples/system-test.

☝️ My feeling being that, knowing if we've written appropriate tests for our samples is perhaps even more valuable than knowing whether we've written good tests for our auto-generated libraries.