Parallel by data: nemo.data merges data from top-level keys about nemo

[sainianubhav]

@grawk Could you let me know where to start on this issue?

[grawk]

Hey @sainianubhav sorry I just saw your message here.

I'm thinking the issue must be introduced by, and may be resolved by changing, this function:
https://github.com/krakenjs/nemo/blob/master/lib/flow.js#L128-L166

[grawk]

Hey @dwbruhn I tried to reproduce this. Please see the following commit:
5d7ab87

Here is the output I got:

dataset: 1
dataset: 2
{ dataset: 2, baseUrl: 'https://www.google.com' }
{ dataset: 1,
  B: { 'This should be the only property A can access under B': 1 },
  baseUrl: 'https://www.google.com' }

It looks like there isn't pollution from one dataset to the other in my example.

[grawk]

Can't reproduce. Closing.

[dwbruhn]

Hey @grawk, sorry for the delay (I was out on PTO).

The contamination only happens when the data is in the base profile. Keep your same test, but move the data section to the base profile, then you'll see it.

Here's the minimal config.json needed to reproduce:

{
    "output": {
        "reports": "path:report"
    },
    "profiles": {
        "base": {
            "maxConcurrent": 1,
            "tests": "path:./*.js",
            "driver": {
                "browser": "chrome"
            },
            "mocha": {
                "timeout": 180000,
                "reporter": "mochawesome"
            },
            "parallel": "data",
            "data": {
                "A": {
                    "dataset": 1,
                    "B": {
                        "This should be the only property A can access under B": 1
                    }
                },
                "B": {
                    "dataset": 2
                }
            }
        }
    }
}

[grawk]

Thanks for the additional info. I'm able to reproduce and have a fix in progress.