Prosemirror-markdown - Block level marks

Violeta-Amza · April 27, 2021, 11:51am

Question1: Let’s say that some marks are not allowed for paragraph nodes. For example, paragraph nodes should accept only inline marks as their content.

const mySchema = new Schema({
  nodes: {
    text: {
      group: "inline"
    },
    paragraph: {
      group: "block",
      content: "inline*",
      marks: 'inline',
      toDOM() { return ["p", 0] },
      parseDOM: [{tag: "p"}]
    },
    doc: {content: "block+"}
  },
  marks: {
    strong: {
      group: 'inline',
      toDOM() { return ["strong", 0] }
    },
  }
})

But when a text node is added, there is no check to test if registered marks are allowed, see

github.com

ProseMirror/prosemirror-markdown/blob/master/src/from_markdown.js#L33


push(elt) {
  if (this.stack.length) this.top().content.push(elt)
}


// : (string)
// Adds the given text to the current position in the document,
// using the current marks as styling.
addText(text) {
  if (!text) return
  let nodes = this.top().content, last = nodes[nodes.length - 1]
  let node = this.schema.text(text, this.marks), merged
  if (last && (merged = maybeMerge(last, node))) nodes[nodes.length - 1] = merged
  else nodes.push(node)
}


// : (Mark)
// Adds the given mark to the set of active marks.
openMark(mark) {
  this.marks = mark.addToSet(this.marks)
}

I think node creation process should respect the node/marks specs. Is there a chance this will be changed? Thanks!

Question2: Currently, block level marks are allowed: Discussion: What are marks - #15 by marijn. But when a block node is added the registered marks are not saved for the current parsed node, see: prosemirror-markdown/from_markdown.js at master · ProseMirror/prosemirror-markdown · GitHub
Why registered marks are reset before adding the current block node? By doing so, the block level marks will be ignored. Is there a chance this will be changed? Thanks!

marijn · April 27, 2021, 3:00pm

It assumes that the markdown parser only creates marks at places where they are valid. Are you customizing the parser, or did you run into a situation where the regular markdown parser creates nodes for marks in the wrong place?

Block-level marks are allowed on the raw node API level, but most of the convenience code for marks (commands, the markdown parser) assumes marks are inline-only.