conform now guarantees order of configuration options

[0xAX]

No description provided.

[0xAX]

@thz this patch provides functional that we talked about. Waiting for review from somebody.

It also provides test in the test/integration_test.exs ():

 test "for the complex data types" do
....

and config (test/confs/complex_example.conf) which represents current behaviour.

[grundrauschen]

I tried this test and get, for a more complex configuration, a response with wrong order.
For a configuration with test.*.foo and test.*.bar I tried the following in the app.conf

test.egg3.foo = "string1"
test.egg3.bar = "string2"
test.egg0.foo = "string3"
test.egg0.bar = "string4"
test.egg1.foo = "string5"
test.egg1.bar = "string6"

and the result in mix conform.effective was

...
{:test, 
  [egg3: [foo: "string1", bar: "string2"],
  egg1: [foo: "string5", bar: "string6"],
  egg0: [foo: "string3", bar: "string4]]},
...

So the order is not preserved in the resulting list.

[0xAX]

Hello

I've update the PR.

cc: @thz @grundrauschen

[grundrauschen]

@0xAX: as far as I can tell from a few tests, the branch works now as intended.

[surik]

The diff has a lot of unrelated changes with code style. Please clean it up.

[surik]

Does somebody know usecase of this PR?
Is it really important to have this

...
{:test, 
  [egg3: [foo: "string1", bar: "string2"],
   egg0: [foo: "string3", bar: "string4"],
   egg1: [foo: "string5", bar: "string6]]},
...

instead of

{:test, 
  [egg3: [foo: "string1", bar: "string2"],
   egg1: [foo: "string5", bar: "string6"],
   egg0: [foo: "string3", bar: "string4]]},
...

in sys.config for

test.egg3.foo = "string1"
test.egg3.bar = "string2"
test.egg0.foo = "string3"
test.egg0.bar = "string4"
test.egg1.foo = "string5"
test.egg1.bar = "string6"

?

[surik]

OK. Looks it is useful for some routing-rules with matching