perf(state): config-gated SQLite PRAGMA tuning for large DBs (salvage #71755)

[kshitijk4poor]

Context

Users with large state.dbs (hundreds of MB — heavy browser/tool sessions) can only tune SQLite via wal_autocheckpoint/journal_size_limit today. This adds three config-gated, non-durability pragmas — database.cache_size, database.mmap_size, database.temp_store — and applies the pragma pass to the read-only connection and the per-thread WAL readers (cache_size/mmap_size are per-connection, so readers never benefited before). WHO benefits: opt-in only — users who set these keys in config.yaml, i.e. large-DB power users; a default install is byte-for-byte unchanged.

Measured impact

211 MB synthetic DB (360k rows), heavy read scan, default page cache vs cache_size=-262144 + 256 MB mmap, median of 5:

scenario	before	after	delta
cold-ish full scan	0.484 s	0.468 s	+3.3%
warm re-scan	0.211 s	0.163 s	+22.9%

Honest caveat: config-gated and default-off; the win only exists for users who opt in on a large DB, and is workload-dependent (warm repeated reads benefit most; cold I/O-bound scans barely move).

Safety audit (the #64091 lesson)

No durability pragmas: the allowlisted tuple gains only cache_size, mmap_size, temp_store — no synchronous, journal_mode, locking_mode, or fullfsync. Values apply ONLY when present in config.yaml; integer coercion warns-and-skips garbage; apply_database_pragmas remains best-effort try/except so a failing pragma cannot break open. The docstring's journal-mode-ownership warning stays accurate. Per-reader cost is bounded: the reader connection is thread-local, so the pragma pass runs once per thread lifetime, and the config read is the mtime-cached fast path.

Provenance

Salvage of #71755 by @crayfish-ai (both commits authorship-preserved). Follow-up commit adds the missing guard test: config-set pragmas must reach the writer, read-only, and per-thread reader connections (with a discriminating -16000 vs SQLite's -2000 default so a revert can't accidentally pass), and a no-config run must leave defaults untouched. Mutation-checked: reverting the pragma wiring turns the test RED.

Verification

• 165 passed (tests/test_hermes_state.py, -p no:randomly) + the new end-to-end pragma test.
• Simplify pass: no material findings (coercion loop reused, not duplicated; per-thread cost traced to once-per-thread-lifetime).

Closes #71755.

[Ruanjq98]

Code Review: #77630

Verdict: Approve

DB PRAGMAs: well-documented addition of cache_size, mmap_size, temp_store. Safe best-effort.

LGTM - Reviewed diff. Changes are sound.