deps: non-breaking updates

[bnb]

Dependency updates. Stepped through each outdated dependency individually and ran tests to ensure that there's no breakages.

Module updates that worked without breakages:

deps

All of these are semver major changes

• chalk@3.0.0
• npm-package-arg@8.0.0
• semver@7.1.1
• strip-ansi@6.0.0
• tar@6.0.0
• yargs@15.1.0

devDeps

• prettier@1.19.1

Three modules broke the tests on update:

• mkdirp@1.0.3
  • metric boatload of usage that breaks. I'm not sure if we'd want to switch to the implementation in Node.js or update it to use the newer version of mkdirp.
• update-notifier@4.0.0.
  • the callback property was deprecated in 4.0.0. The .fetchInfo() method is the replacement (yeoman/update-notifier#158).
• execa@4.0.0
  • breaks Node.js 8.x builds. Once ci: remove Node.js 8 from CI suite #783 or equivalent lands, this can be updated.

If it's helpful, I can create issues for these.

Checklist

• npm test passes

[codecov-io]

Codecov Report

❗ No coverage uploaded for pull request base (master@6ae20c8). Click here to learn what that means.
The diff coverage is 100%.

@@            Coverage Diff            @@
##             master     #782   +/-   ##
=========================================
  Coverage          ?   96.27%           
=========================================
  Files             ?       27           
  Lines             ?      887           
  Branches          ?        0           
=========================================
  Hits              ?      854           
  Misses            ?       33           
  Partials          ?        0

Impacted Files	Coverage Δ
lib/unpack.js	100% <100%> (ø)
lib/temp-directory.js	100% <100%> (ø)
lib/git-clone.js	100% <100%> (ø)

---

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 6ae20c8...3595613. Read the comment docs.

[bnb]

As far as I can tell this is not an issue with the module itself, especially since the changelog does not list anything that would result in such a breakage.

Happy to revert if folks think it's needed.

From Travis:

test/npm/test-npm-author-name.js 2> /Users/travis/build/nodejs/citgm/node_modules/strip-ansi/index.js:1
test/npm/test-npm-author-name.js 2> 'use strict';
test/npm/test-npm-author-name.js 2> 
test/npm/test-npm-author-name.js 2> 
test/npm/test-npm-author-name.js 2> TypeError: undefined is not a function
test/npm/test-npm-author-name.js 2>     at /Users/travis/build/nodejs/citgm/node_modules/strip-ansi/index.js:1
test/npm/test-npm-author-name.js 2>     at Script.runInThisContext (vm.js:116:20)
test/npm/test-npm-author-name.js 2>     at Object.Module._compile (/Users/travis/build/nodejs/citgm/node_modules/tap/node_modules/source-map-support/source-map-support.js:525:25)
test/npm/test-npm-author-name.js 2>     at Module._extensions..js (internal/modules/cjs/loader.js:991:10)

[MylesBorins]

Now that we've dropped support for Node.js 8 I've gone ahead and rebased this and update a few other dependencies. Biggest changes include

• removing mkdirp (using fs.promises.mkdir /w recursive option)
• updating xml builder which changes the output of the junit reporter... only changes should be formatting though, other tests were unaffected. Will run CITGM with this branch to confirm

[MylesBorins]

Run of CITGM (no build) to test the output of the XML reporter

https://ci.nodejs.org/view/Node.js-citgm/job/citgm-smoker-nobuild/732/

[MylesBorins]

The strip ansi failure on travis is super odd... fails only only on windows and only on travis. (various windows runs on github all passed)

[BridgeAR]

LGTM with green CI

[BridgeAR]

While being on this, we might also switch to using assert.rejects(). Otherwise we do not verify that the catch clause is actually reached.

[MylesBorins]

Feel free to push to this PR, although if you do please use the built in tap asserts

https://node-tap.org/docs/api/asserts/#trejectspromise--fn-expectederror-message-extra

TBH the tests across the board could use a refactor so I'd prefer to save fixing this for a larger refactoring

[MylesBorins]

LGTM