Performance enhancements, time indices and secondary indices for Convex-db

[trsb-standards]

Added various performance improvements for disk space usage of etch and memory usage of the Convex-DB

Added:

• secondary indices API
• versioned tables and time indices

[mikera]

(Claude Comment) Thanks @trsb-standards — there is clearly valuable work in here, but at +8,250 lines across 40 files it is too large to review and merge in one go. We would like to take it, broken into atomic, independently verifiable and mergeable steps.

The natural split seems to be the three features named in the description, each as its own PR with its own tests:

1. Etch disk-usage / memory improvements — likely the most self-contained and easiest to land first
2. Secondary indices API
3. Versioned tables and time indices (building on 2 if there is a dependency)

Encouragingly, the branch currently merges clean against develop (no textual conflicts despite the April base), so now is a good moment to do the split rather than letting it drift. Smaller PRs will also get CI runs and review turnaround in days rather than months. Happy to review the first slice as soon as it is up.

[brittleboye]

Thank you for this contribution, and for your patience with it sitting in the queue — there is clearly a lot of valuable work here across the time indices, secondary indices, and the Convex DB performance enhancements.

At ~8,250 lines across 40 files, though, this is more than we can review and integrate safely as a single change, particularly with protocol-adjacent work in flight ahead of the v1 upgrade. As you're already aware, we'd like to take it in as a series of smaller, incremental PRs that can each be reviewed and validated on their own merits — perhaps starting with the foundational indexing changes, then layering the secondary indices and the performance work on top.

I'm going to close this PR for now to keep the review queue clear, but please don't read that as a rejection of the work — it would be very welcome back as focused pieces whenever you're ready, and we're happy to help sequence them.

Thanks again for the effort here.

— Claude (AI assistant helping maintain the Convex repository, on behalf of the maintainers)