@Dadoo - I’m super excited about this release, too! A couple general notes on your stuff above (and it looks like our release notes were a little ambiguous, so we’ll get those updated to reflect this too!)
- The search changes and the system which had the memory and processing leak were actually independent issues. We swapped out the search because Tantivy offered much more consistent results and a more effective approach to memory as compared to our proprietary index. That said, we still use SymSpell, and we still use cloud search. This change swaps out our old index for Tantivy and paves the way for a number of other future changes and optimizations to the suggestions we pass into the cloud.
- While we aimed for parity between the indexes for this at first, this change creates an opportunity for us to improve search results in the future (while this change just focused on improving its speed and memory).
- The memory and processing leak was in the heart of the system that managed all your views of your data. We had a leak and other inefficiencies in the area that kept it sorted which prompted us to change out the structural approach and the sorting approach without changing the architectural approach (for now!)
We will edit the release notes to make that less ambiguous! For what its worth, the only reason I wanted to clarify is because the search and the memory issues are NOT related to one another, so if you experience issues with one its unlikely to affect the other!
Cheers,
Steve