perf: reduce concurrent thread overhead

[ronak-guliani]

What Changed

• Cap SQLite WAL files at 32 MiB and verify the actual file shrinks after exceeding the limit.
• Suppress idle port-poll and Git Trace2 setup spans that add tracing work without useful timing data.
• Add migration 111 to remove two projection indexes that are strict prefixes of retained covering indexes.
• Fold message shell-summary fields directly and use an indexed O(1) client activity append path, with sorted fallback handling for snapshots, out-of-order events, and replays.

Why

• These changes reduce disk growth, tracing overhead, index write amplification, and repeated full-history work when multiple threads stream concurrently.
• The implementations follow pingdotgg/t3code#13684, pingdotgg/t3code#13756, pingdotgg/t3code#12847, and pingdotgg/t3code#9152. The index migration intentionally drops only the two strict prefixes proven redundant in this fork's divergent schema.
• Validation passed: pnpm fmt:check, pnpm lint, pnpm typecheck, pnpm test, pnpm --filter t3 build:bundle, and git diff --check. Two earlier full-suite runs each hit a different existing timing-sensitive test timeout; both tests passed immediately in isolation, and the final full pnpm test run passed.

Checklist

• This PR is small and focused
• I explained what changed and why
• I included before/after screenshots for any UI changes
• I included a video for animation/interaction changes