fix: globby dep

[kabirbaidhya]

Resolves #101

• Move globby to package dependencies.
• This package globby is being used by the package but is only installed in devDependencies. As it eventually cascades up to the user's CLI project, it would fail if the user do not have globby installed in their package.json. It should be a core dependency as per the usage here:
  

  config/src/plugin.ts

  Line 2 in 6c800f2

  import * as Globby from 'globby'

See issue #101 for details.

[salesforce-cla]

Thanks for the contribution! Before we can merge this, we need @kabirbaidhya to sign the Salesforce.com Contributor License Agreement.

[codecov]

Codecov Report

Merging #102 into master will decrease coverage by 66.81%.
The diff coverage is n/a.

@@            Coverage Diff            @@
##           master   #102       +/-   ##
=========================================
- Coverage   66.81%     0%   -66.82%     
=========================================
  Files           7      2        -5     
  Lines         452     12      -440     
  Branches      127      3      -124     
=========================================
- Hits          302      0      -302     
+ Misses        101     12       -89     
+ Partials       49      0       -49

Impacted Files	Coverage Δ
src/index.ts	0% <0%> (-100%)	⬇️
src/debug.ts	0% <0%> (-45.46%)	⬇️
src/ts-node.ts
src/util.ts
src/plugin.ts
src/config.ts
src/command.ts

---

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 b0c2c9e...fad8f0d. Read the comment docs.

[kabirbaidhya]

Not sure how there's this coverage decrease with just a line change in package.json. Is this legit?

[G-Rath]

@kabirbaidhya the weird coverage change is because of how the change is actually calculated, which is by comparing to the coverage reported by the last collection run against master.

This is instead of it running coverage on the target branch before & after this PR and comparing the results, which is what the report appears to be representing.

This is fine in most cases, except if something changed in the last run that causes coverage to change wildly, such as a node version being dropped, or CI failing drastically that means the majority of tests don't run and so half the application isn't touched.

In general, the oclif repos don't see a lot of attention these days, so this happens for most of their PRs.

[RasPhilCo]

🏄