lexical_mark 1.7.4

Mark nodes (annotations and comment ranges) for lexical_core, wire-compatible with Lexical's @lexical/mark.

lexical_mark

Mark nodes — annotations and comment ranges — for lexical_core. Pure Dart.

dependencies:
  lexical_mark: ^1.0.0

final editor = LexicalEditor(nodes: markNodes);

editor.update(() {
  $getRoot().append(
    $createParagraphNode()
      ..append($createMarkNode(['comment-1'])
        ..append($createTextNode('markiert'))),
  );
}, discrete: true);

A mark wraps inline content and carries a set of identifiers. Overlapping annotations are represented by nesting marks rather than by letting one node belong to two ranges — which is why ids is a list: the innermost mark of an overlap carries every identifier that covers it.

Wire-compatible with @lexical/mark 0.48.x.

Commenting a selection

registerMark(editor);                             // once
final id = 'comment-42';                          // yours
editor.dispatchCommand(addMarkCommand, id);       // mark the selection
editor.dispatchCommand(removeMarkCommand, id);    // resolve it

Only the id reaches the document. Author, text, replies and timestamps belong to the application, keyed by that id — which is what lets a comment be written, answered and resolved without touching the document, and keeps a document with fifty threads the same size as one with none.

$getMarkedText(id) reads back what the mark covers now, so a sidebar's quote follows the text as it is edited instead of quoting something that no longer exists. $getMarkIdsAtSelection() answers which threads the caret is inside.

Marking the same range twice adds the second id to the mark already there; marking an overlapping range nests, because a node has one parent. A mark left with no ids is unwrapped rather than kept — an annotation nobody refers to is invisible, and leaving it behind would grow the document by one element per resolved comment forever.

Licence

MIT. Derived from Lexical, © Meta Platforms, Inc., also MIT. See NOTICE.