#blocker

Since the last comments here, we've done 📌 Render the navigation toolbar in a placeholder Active which is not directly related to this issue but does help.

And 📌 Review cache bin and cache tags of access policy caching Active which added the 'not worth caching' interface discussed here, although not yet applying it to render arrays yet.

For any render array that appears on a lot of pages (e.g. the new navigation toolbar), if you both set cache keys and also #create_placeholder TRUE, then it gets its own render cache item, but also it's only included in the dynamic page cache as a placeholder. That was already true before that issue, but now it will also skip big pipe rendering replacement and be fetched direct from the render cache. It also means that the cache tags are isolated from dynamic page cache item (which might help for things like node:list cache tags and similar).

This means less 'double caching' of items between the dynamic page cache and the render cache, although the placeholder metadata itself still ends up in the dynamic page cache so for tiny items it probably wouldn't save anything.

I think if we combine more placeholdering for some things, with the not-worth-caching logic for others, we should be getting closer to what this issue was trying to do.

I think there's quite a bit of overlap with 📌 Automatically set revision user/log information/created time on entity revisions Needs work here. Given this issue hasn't had much action in a long time (existing patches are against code that no longer exists) I wonder if we should close this as outdated?

Some updates for MR
- Coding Standards changes - as reported in the MR threads
Btw - I have no idea why MR didn't say anything about those issues.

+ Update for the thread about ConfigTranslationEntityDisplayListBuilder inheritance, added extra class to share with existing ConfigTranslationFieldListBuilder ... instead of a "direct one"

still things to do ...