Error in 'wu list'

[wukefe]

Before I check the command 'wu list' in the directory of 'Ostrich2', I have followed the steps below.

1. Sync wu-wei-benchmarking-toolkit, and update the executable file 'wu'
2. Sync Ostrich2

The error message is pasted here.

(Canopy 64bit) HF:Ostrich2 wu list
fs.js:844
  return binding.stat(pathModule._makeLong(path));
                 ^

Error: ENOENT: no such file or directory, stat '/Users/wukefe/Documents/GitHub/wu/Ostrich2/.wu/setup.json'
    at Error (native)
    at Object.fs.statSync (fs.js:844:18)
    at Object.<anonymous> (/Users/wukefe/Documents/GitHub/wu-wei-benchmark-toolkit/bin/list:39:8)
    at Module._compile (module.js:410:26)
    at Object.Module._extensions..js (module.js:417:10)
    at Module.load (module.js:344:32)
    at Function.Module._load (module.js:301:12)
    at Function.Module.runMain (module.js:442:10)
    at startup (node.js:136:18)
    at node.js:966:3

[elavoie]

Do you have a 'setup.json' file in your .wu directory at the root of your Ostrich2 repository?

It should look like:

{
    "platform": "mba-2011"
}

[wukefe]

You are right. Because I deleted the old repository and forked the latest one, the setup.json is missing.

After I add it into the .wu directory, the message returned from 'wu list' seems incorrect. What does 'no (or more than one) schemas match' mean?

(Canopy 64bit) HF:.wu wu list
[
  {
    "field": "data",
    "message": "no (or more than one) schemas match",
    "value": {
      "type": "platform",
      "short-name": "Hanfeng’s MacBook Pro",
      "name": "Darwin-15.3.0-x86_64-i386-64bit",
      "cpu": "Intel Core i5",
      "gpu": "Intel Iris Graphics 6100",
      "memory": "8 GB",
      "os": "Mac OS X 10.11.3",
      "supported-languages": [
        "x86_64"
      ],
      "supported-formats": []
    }
  }
]

[elavoie]

The message means that the JSON description listed in the value field does not conform to an expected schema. In your case, the short name you are using uses non-supported characters such as the apostrophe (') and spaces (' '). Look at the regexp for 'short-name' in the schema file to see what is valid:
https://github.com/Sable/wu-wei-benchmarking-toolkit/blob/master/lib/config-schema.json

Also, while the toolkit will not reject it for that, try to use really short names for the 'short-name' field so that it does not require a lot of typing when using it. I would suggest 'hf-mbp' rather than 'Hanfeng's MacBook Pro'. The commandline interace requires the user to type the short-name in an exact way to refer to the whole description. Ex: If we were to filter all runs according to the platform their where obtain from, it is a lot more convenient to type the second version.

It would be good if your script reflected those constraints in the choice of a short-name. Otherwise we will have to ask people to change them manually.

[elavoie]

I added this issue Sable/wu-wei-benchmarking-toolkit#15 to eventually fix it and make it more convenient.

[wukefe]

Thanks a lot.

The illegal characters in the field of short-name lead to the error. I also find a new field environments is added into the platform.json. I will modify platform script to generate the new field.