Applying Linter on Tests

[hiranya911]

So far we haven't been linting our test sources. Now that we have upgraded to the latest version of tslint (#194), we should start linting our unit and integration tests. This PR adds the following npm tasks:

npm run lint:src # Lints the source files (everything under src/)
npm run lint:unit # Lints unit tests
npm run lint:integration # Lints integration tests

The aggregate task 'npm run lintruns all of the above, which will run as part ofnpm test` as well (this causes all lint jobs to run with CI).

I also fixed all the existing lint errors in test files. I relaxed the linter rules a little bit for tests by introducing a separate tslint-test.json file.

[merlinnot]

Did you consider adding lint:*:fix too? 🙃

[merlinnot]

npm-run-all supports glob-like pattern matching, so you can do run-s lint:*.

Why did you decide to run these in sequence instead of parallel?

[hiranya911]

@merlinnot What should lint:*:fix do? Run tslint with the --fix flag?

Good idea to run the linter tasks in parallel. I'll make the change.

[merlinnot]

Since tslint is a dev dependency (which is a good thing!) one should use node_modules/tslint/bin/... —config ... —fix to fix all auto-fixable issues, which is a bit annoying. It’s much easier to yarn run lint:unit:fix or yarn run lint:fix. You can even leverage existing scripts like so:

{
  "lint:unit:fix": "run-s \"lint:unit —fix\""
}

One day we can even add precommit hooks on top of this, so it fixes everything before committing using husky and lint-staged, but that shouldn’t be a part of this PR.

[avishalom]

LGTM