Take config file instead of piped output

[tgetgood]

The config file format of name-your-contributors solves our cli duplication conundrum.

Add a --config option which takes an nyc config file, sends it through nyc, and postprocesses the results.

[tgetgood]

Question: do we want to process each config file entry individually, or just have one grand total?

[RichardLitt]

Can you specify multiple configuration files? I think having one grand total makes more sense, but I may be missing something.

[tgetgood]

Yes you can:

$ whodidwhat --config configA.json -u tgetgood > out.json

Each config file can have multiple repos and orgs. See the README of nyc for the details.

By grand total I mean: say you have a config like

{
  "token": "abc...",
  "repos": [{
    "login": "mntnr",
    "repo": "whodidwhat"
  }, {
    "login": "mntnr",
    "repo": "name-your-contributors"
  }]
}

Do you want one object from names to stuff done, like:

{
  "tgetgood": {
    "commits": 7,
    "issuesCreated": 9
  }
}

Or broken down by repo like:

{
  "repos": [{
    "login": "mntnr",
    "repo": "whodidwhat",
    "contributions": {
      "tgetgood": {
        "commits": 6
      }
    }
  }, {
    "login": "mntnr",
    "repo": "name-your-contributors",
    "contributions": {
      "tgetgood": {
        "commits": 1,
        "issuesCreated": 5
      }
    }
  }]
}

If they're both useful we could make one the default and put a flag in to do the other.

From my understanding of the use case though, the grand total is what you're really after.

[RichardLitt]

If they're both useful we could make one the default and put a flag in to do the other.

They may both be useful. The default should be global. :)