This is a pretty meta…ish problem and really not important nor urgent. However, I wanted to talk about it.
AFAIK, CM unloads invisible parts of longer texts inside the editor to save resources. This may — as you’re surely aware of — lead to difficulties with third party features which rely on the entire content being there inside the DOM.
The most simple example is the built-in Cmd+F search of every browser. It doesn’t work for longer texts in CM. A workaround for this is using CM’s own search module.
Another example: I use LT (https://languagetool.org) as a browser extension for spell checking. This also breaks in longer (partly unloaded) texts. Whenever LT tries to replace a typo, my document gets mixed up. I think this has something to do with LT not setting the correct indexes of characters within the text, because parts of it have been unloaded in the meantime.
I could probably work around these issues by increasing the buffer size of CM. Is this possible? But I’m wondering if there are better solutions.