A complicated aspect of Loops is captions.
The challenging thing with toots or captions is that they can contain hashtags, mentions and urls.
That's fine, until you're building a mobile app and need to handle these links differently.
Instead of complicated regexes, I'm going to copy Bluesky and use token indices.
This will allow us to normalize these entities while allowing spaces in hashtags and mentions at the backend level, because RN doesn't have full regex support.
@dansup I'd like to see a FEP expanding Tag objects to include indices within the Content for those tags, if they are in the content. Parsing HTML here sucks