It's my mistake, sorry for that, fixed

Can we modify SharedMarkdown.State to include quotationsCount explicitly? And what do you think about renaming it to quotationsDepth to indicate that we're not simply counting the occurrences?

Agree, good call. All inline comments are addressed now.

Before this change, all of the types underneath "simple-markdown types" were copy-pasted from simple-markdown.

After this change, you have broken this property.

When you introduce incompatibilities in types like this it can cause issues down the line. I'm not sure what issues might occur here, but I think @tomek's suggestion to update this type and make it inconsistent with what is defined in simple-markdown was a bad idea.

This code differs from the old version in two ways:

1. It doesn't handle state.inline in the same way. Do you know what that is about? It seems to me like you might be causing a regression here, where previously blockQuote would not be matched in the inline case, but now it will. It's disappointing that this was not discussed in the diff review... ideally @kamil would explain why he made this change in the diff description, but at a baseline the diff should not have been accepted without that explanation. It seems possilbe that @kamil has no idea what state.inline is about... we should never land a diff in this state.
2. The function we're using here doesn't have a .regex property. What is the significance of setting matchFunction.regex? This change should be explained, but it doesn't appear to be covered in the diff at all... same issue as above, @kamil should've identified and explained this change, or at the very least the reviewers should've caught it.

It's because the differences between parsing this on native/web.
blockQuoteStripFollowingNewlineRegex allow as to remove additional new line, more context here: https://phab.comm.dev/D8?id=27#inline-89

For test, result after using only blockQuoteRegex on both platforms:
web: {F162651}
mobile: {F162652}