feat(event cache): connect the room event cache and persistent storage

[bnjbvr]

This implements both directions for interacting between the event cache and the storage backend. I think this is fine design… as long as we don't have any other process writing into the cache, in which case we'll need to reset at different places.

writing

The main idea in this PR is to use RoomEventCacheState::with_events_mut(func: impl FnMut(&mut RoomEvents)) to perform any write to the underlying RoomEvent data structure. The function takes care of maintaining the storage, after updates have been performed.

Additionally: this behavior isn't enabled by default; callers have to call EventCache::enable_storage() when they start using the event cache.

I've put the RoomEventCacheState data structure in a private module, so that even the same module can't access its private fields, and we don't make mistakes by accidentally misusing events.

reading

Reading happens by reloading the chunks initially, and prefilling the deduplicator with the reloaded events information. It's mostly boring changes at this point.

---

Part of #3280. On top of #4362. Split and reworked from #4308.

[codecov]

Codecov Report

Attention: Patch coverage is 88.17204% with 11 lines in your changes missing coverage. Please review.

Project coverage is 85.15%. Comparing base (5da36d1) to head (fca4131).
Report is 5 commits behind head on main.

Files with missing lines	Patch %	Lines
crates/matrix-sdk/src/event_cache/deduplicator.rs	63.63%	4 Missing ⚠️
crates/matrix-sdk/src/event_cache/pagination.rs	82.60%	4 Missing ⚠️
crates/matrix-sdk/src/event_cache/mod.rs	78.57%	3 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #4347      +/-   ##
==========================================
+ Coverage   85.13%   85.15%   +0.02%     
==========================================
  Files         280      280              
  Lines       30666    30716      +50     
==========================================
+ Hits        26107    26157      +50     
  Misses       4559     4559              

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[bnjbvr]

Getting back to a draft for this; I suspect a bit more work is required because the back-pagination tokens ought to belong in this state too…

[bnjbvr]

Calling this ready for review, after a rebase from main. I think the backpagination issue I've thought about, should be resolved independently.

@Hywan Whenever the time is right for you, PTAL ☺️

[Hywan]

It looks good. I like the new API to mutate the state. Thanks for putting everything together.

[Hywan]

It's part of the PR I'm preparing I suppose, with VectorDiff being emitted by EventCache?

[bnjbvr]

Exactly. With your PR, we'd get a remove(0) and a push_back (or insert(1) ^^). Right now, these are not sent via the updates, and that's why we can't use this by default yet 👀