Turned out it is. The counting was off in too many cases and the way we limited the content led to strange behaviour in a lot of cases. Thanks to all the bug reports we’ve got a pretty picture of what’s wrong, what people want and how complex it is to get even that seemingly small task right.
Philipp took all your input and did a complete rewrite of the extension. Here are my favorite cases we handle now (but didn’t before):
- Dispatching a transaction that adds content before and after the document? That’ll get blocked.
- Paste a really long text and it’s just trimmed to the configured length.
- Paste content with a table where the limit would chop off half the table: Blocked.
With the new extension you can also switch between two different counting modes, count characters or words, and can even count characters just for a specific node.
Read more about the rewrite in the Pull Request on GitHub and more about the CharacterCount extension in the docs.
