Text-based snapshot-testing

[akaDuality]

We should create an instrument that allows to compare the designer's layout with the result of developer's work. Any kind of snapshot-testing can do it.

We have great AccessibilitySnapshot, but the library generates images. Screenshot-testing has several downsides for our purpose:

• The image from designer will never match to the result of app's rendering.
• The test if fragile to changes of the layout: test would fail even if VoiceOver works correctly.

I would like to create some kind of text description of a view that can be generated directly from the macOS app. Example:

Real image example

Text description:

Close. Button.

| --- Description ---
| The roasted Veggie. Header, Image.
| Medium, signature dough, 572 grams.
| Nutrition facts. Button.
| Tomato sauce, Mozzarella Cheese, Roasted Red Onions...
| Remove ingredients. Button.
| Size: Medium, 2 of 2. Adjustable.
| ---

| --- Toppings ---
| Add to pizza. Header.
| Swap to BBQ base. Button
| Swap to truffle base. Button.
| ...
| ---

| --- Summary --- 
| Selected, Add to basket for £15. Button
| ---

Core part of view parsing for accessibility should be the same with AccessibilitySnapshot, because it's great! In addition we can suggest to it the text-strategy of snapshot generation.

As a result, text-based snapshot can be generated from the app and a develop can use AccessibilitySnapshot to match it with his result.

[akaDuality]

Issue