fix: Ensure that getLastCrashReport() is actually the last crash report

[felixrieseberg]

This PR fixes #11749 by ensuring that getLastCrashReport() actually returns the last crash report (and not just a certain position in the array).

Also: A test that actually calls the method under test 😅

[codebytere]

LGTM 💯 when we merge this we can close the other PR

[ckerr]

This PR is short and sweet. LGTM. One nit inline.

[ckerr]

It's not clear from this block that the new code is being tested.

Would it be possible to prepend an assert(reports.length > 1) to show that comparison happened?

[ckerr]

electron-linux-x64:

AssertionError [ERR_ASSERTION]: false == true
    at Context.it (/home/builduser/project/spec/api-crash-reporter-spec.js:266:7)

electron-linux-ia32:

not ok 987 crashReporter module getLastCrashReport correctly returns the most recent report
  AssertionError [ERR_ASSERTION]: false == true
      at Context.it (/var/lib/jenkins/workspace/ron_ARM64_electron_PR-12253-WP3JY47LEEJ2OR6WRBG4W43ZPAXD4YTS3ZGJCPY3GXIOPM3Q5EKQ@2/spec/api-crash-reporter-spec.js:266:7)
      at runCallback (timers.js:763:18)
      at tryOnImmediate (timers.js:734:5)
      at processImmediate (timers.js:716:5)

seems legit

[trop]

An error occurred while attempting to backport this PR to "1-8-x", you will need to perform this backport manually

[trop]

We have automatically backported this PR to "2-0-x", please check out #12255

[MarshallOfSound]

@felixrieseberg Could you backport this to 1-8-x 😄 The bot did 2-0-x for you 👍

[ckerr]

@MarshallOfSound IMO assert(reports.length > 1) was still warranted and we're literally in the same TZ so 👎 on the dismissal.

Not a blocker though so ¯\(ツ)/¯ on this