Hi! Would you consider an option in @codemirror/view to render the entire document without JavaScript-based virtualization?
My main motivation is compatibility with browser-native features and tools that inspect the DOM. Browser extensions, including AI agents, can miss content that isn’t mounted, and browser-native “Find in page” cannot search the complete document.
I’d even be okay to sacrifice (JavaScript/CSS) virtualization completely for documents that I already know aren’t going to exceed certain number of lines or size purely to support these use cases.
In my own experiments with content-visibility: auto, grouping multiple lines into blocks rather than applying it to individual lines, I found that it performed reasonably well in Chrome and Firefox. Safari struggled more, though that may be specific to my implementation.