At some point we might want to break validateContent() into smaller functions (perhaps validateReportHeader, validateSystemInfo, etc.) -- Although this will make totaling the expected number of tests trickier.

LGTM, and I ran a CI on it:
https://ci.nodejs.org/view/post-mortem/job/nodereport-continuous-integration/100/
needs a rebase though

Rebased. Started new CI run: https://ci.nodejs.org/view/post-mortem/job/nodereport-continuous-integration/109/

CI failed on Windows to download a Node runtime. Raised #66.
Also failed on AIX: https://ci.nodejs.org/view/post-mortem/job/nodereport-continuous-integration/109/MACHINE=aix61-ppc64/console

# Subtest: test/test-fatal-error.js
    1..0
not ok 9 - test/test-fatal-error.js # time=30022.646ms
  ---
  timeout: 30000
  file: test/test-fatal-error.js
  failure: timeout
  results:
    ok: false
    count: 0
    pass: 0
    plan:
      start: 1
      end: 0
      skipAll: true
  signal: SIGTERM
  command: >-
    /home/iojs/build/workspace/nodereport-continuous-integration/MACHINE/aix61-ppc64/node-v8.0.0-nightly2017030187a039d721-aix-ppc64/bin/node
  arguments:
    - test/test-fatal-error.js
  ...

I think I've seen the AIX failure before. I'd have thought that 30 secs would be plenty of time to fill the 20Mb heap, but maybe the machine is very slow (it does seem to be by far the slowest platform to run in the CI).

@rnchamberlain if you think there is a flaky test can you open an issue for it so that it gets investigated ?

I opened an issue for an AIX failure as I ran on master and it there was a failure so it may be persistent as opposed to flaky: #69

LGTM

Landed as 918ef91