Should we parse the output of the API?

[dave-irvine]

At the moment we complete callbacks with the raw JSON output of the API... what about parsing the output?

This lets us call functions on the created objects:

This:

gitlab.projects.all (projects) ->
   for project in projects
      project.members.all (members) ->
         console.log members

rather than:

gitlab.projects.all (projects) ->
   for project in projects
      gitlab.projects.members.list project.id, (members) ->
         console.log members

Not the best example, but it seems simpler?

[moul]

👍 +1 for me

[mdsb100]

+1.
In cli-gitlab

stringifyFormat = (data) ->
  console.log(stringify(data, {space: 2})) if data?

As u say, "data" will be a object, and "data" need a function "toJSON".

But sometimes "data" is "true".

So this:

str = ""
if data.toJSON?
  str = stringify(data.toJSON(), {space: 2})
else
  str = stringify(data, {space: 2})

[mdsb100]

CLI
this:

gitlab membersWithProject --filter "project.owner.id==9"

rather than:

gitlab projects --filter "item.owner.id==9"

[
{
    .......
    "id": 29,
    ......
    "owner": {
      "avatar_url": "http://gitlab.baidao.com//uploads/user/avatar/9/764b89699bcc946e4239b04f28002ce9.jpeg",
      "id": 9,
      "name": "Cao Jun",
      "state": "active",
      "username": "mdsb100"
    },
    .......
  }
]

get project id 29

gitlab members 29