handler: replaces Handler.Schema with Handler.Params (graphql.Params)

[chris-ramon]

Overview

• handler: replaces *Handler.Schema with Handler.Params (graphql.Params).
  • So we can actually send to graphql.Do(...) the params we exactly need, like Context, RootObject, etc.
• handler_test:
  • TestHandler_Params_NilParams: removes it since we want to graphql.Do handle validation errors such us schema is nil.
  • TestHandler_Params_RootObject: adds test unit to check we are passing correctly graphql.Params.RootObject — thanks a lot @nossila 👍
• graphiql_test: updates it to use handler.Config.Params.
• Closes: Feature request! #1

Action required

• Add the following within Breaking Changes section when shipping a new release including this PR:

---

• Handler.Schema was replaced by Handler.Params graphql.Params:
• Before:

h := handler.New(&handler.Config{
  Schema: testutil.StarWarsSchema,
  GraphiQL: true,
})

• Now:

h := handler.New(&handler.Config{
  Params: graphql.Params{
    Schema: testutil.StarWarsSchema,
  },
  GraphiQL: true,
})

---

Test plan

• Unit test:
• Manual test, via GraphiQL:
  

package main

import (
        "net/http"
        "github.com/graphql-go/handler"
        "github.com/graphql-go/graphql"
        "github.com/graphql-go/graphql/testutil"
)

func main() {
        h := handler.New(&handler.Config{
                // Before:
                // Schema: &testutil.StarWarsSchema,

                // Now:
                Params: graphql.Params{
                        Schema: testutil.StarWarsSchema,
                },

                Pretty: true,
                GraphiQL: true,
        })
        http.Handle("/graphql", h)
        http.ListenAndServe(":8080", nil)
}

[coveralls]

Coverage decreased (-2.0%) to 84.375% when pulling feba169 on nossila/master into b7e0284 on master.

[lwc]

Being able to set RootObject is great, but this PR currently breaks request scoped context rendering this library useless for use with typical middleware based authentication & tracing techniques.

[chris-ramon]

Totally agree with you @lwc, thanks a lot for taking a look to this PR — closing this one in favor of: #42