@unito/integration-sdk/src/richTextFixtures/README.md

Integration SDK

# richTextFixtures

A corpus of canonical RichText ASTs known to be problematic for codec
implementations — nested lists, headers-in-list, soft breaks, mentions,
empty formatting, link-with-quote, malformed input, etc.

## Why

Every rich-text codec hits the same edge cases. This module collects
those shapes once, in canonical AST form, so a new integration inherits
the catalog instead of rediscovering each shape via production incidents.

## Usage

```ts
import { canonicalFixtures } from '@unito/integration-sdk/richTextFixtures';

for (const [name, ast] of Object.entries(canonicalFixtures)) {
  // Run whatever assertion your integration's codec needs.
  // E.g. snapshot the serializer output of one of your fields:
  assert.equal(yourField.serialize(ast), readFileSync(`./snapshots/${name}.txt`, 'utf-8'));
}
```

How you organise your tests — snapshot layout, which fields you cover,
how you generate expected outputs — is up to the integration. The SDK
only provides the corpus.

## Catching silent gaps

If you want every fixture to have a declared expected output (so a codec
can't accidentally ship doing nothing on, say, nested lists), let
`readFileSync` throw on the missing snapshot path. The ENOENT path
points at exactly the file the engineer needs to populate.