Output in text and markdown not equivalent (text output incomplete)

[misko321]

The text output which is printed by default by the CLI doesn't show all the information that the markdown output shows:

spec1.yaml

components:
  schemas:
    SomeDTO:
      someProperty:
        type: string
        enum:
          - ENUM_VALUE_1
          - ENUM_VALUE_2

spec2.yaml

components:
  schemas:
    SomeDTO:
      someProperty:
        type: string
        enum:
          - ENUM_VALUE_1
          - ENUM_VALUE_2
          - ENUM_VALUE_3

---

markdown output

#### What's Changed
---

##### `GET` /resources1/{id}/resources2


###### Return Type:

Changed response : **200 OK**
> OK


* Changed content type : `application/json`

    Changed items (object):

    * Changed property `someProperty` (string)

        Added enum value:

        * `ENUM_VALUE_3`

text output

==========================================================================
==                            API CHANGE LOG                            ==
==========================================================================
             API for Umetrics Core Project Management Service             
--------------------------------------------------------------------------
--                            What's Changed                            --
--------------------------------------------------------------------------
  Return Type:
    - Changed 200 OK
      Media types:
        - Changed application/json
          Schema: Broken compatibility
- GET    /resources1/{id}/resources2
                 API changes broke backward compatibility                 
--------------------------------------------------------------------------

text output tells nothing about what actually broke backwards compatibility, but since it's the stdout of the process it should show everything that's needed to determine the reason of the problem.

Alternative solution: give an option to decide which kind of output: text/markdown/json should be printed to stdout.

[rennyeb]

I've had a similar experience when the response schema has a difference in its "oneOf".

I see that MarkdownRender invokes schema.getOneOfSchema() to get these details, whereas ConsoleRender and HtmlRender don't.

The net result is that the output from MarkdownRender contains important details about how the schema has broken compatibility, but the other renders omit these details.

[DrSatyr]

All kinds of renders and console output are not coupled to each other, like separate features created for specific needs. Agree - it introduces some mess. Good point for improvement.