Speed up history for large amounts of commits

[joepio]

If I open a large document like the test document and view its history, my M1 macbook takes 10+ seconds to load.

It creates resources for every single change.

• Fetching takes 1.6s for all collections
• CPU time takes 15s

Maybe rendering of the 50.000 DOM nodes is the slow thing? Maybe it's creating the resources?

Some ways to fix this

Find stupid stuff in the code

This would be my hunch

Fetch representations from the server

For every (say 100) verions, just use the /versions endpoint.
Will be faster I think, but also more complex as code

Only build representations client-side on click

Maybe again use the server to fetch representations?

[joepio]

Polle's solution reduces the issue by quite a large margin (x4?), although it will still be slow for even larger resources. The addition of the loading page makes this more bearable at least.