fix(core): harden fusion.db pragmas against mid-write crashes

Switch synchronous from NORMAL to FULL and restore the default
wal_autocheckpoint of 1000 (was 100). The aggressive checkpoint
cadence + NORMAL fsync left a wide window for torn pages whenever
a writer crashed mid-checkpoint, which is exactly what happened
when node:sqlite SIGSEGV'd inside pager_write and corrupted the
production db.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

.changeset/fusion-db-synchronous-full.md

@@ -0,0 +1,5 @@
+---
+"@runfusion/fusion": patch
+---
+
+Harden `fusion.db` against process crashes: switch `PRAGMA synchronous` from `NORMAL` to `FULL` and restore the default `wal_autocheckpoint = 1000` (was 100). Repeated node:sqlite SIGSEGVs inside `pager_write` had been corrupting the db; the previous settings left a wide window for torn pages whenever a writer crashed mid-checkpoint. The small fsync cost is worth the durability win.

packages/core/src/__tests__/db.test.ts

@@ -194,8 +194,8 @@ describe("Database", () => {
       const autoCheckpoint = db.prepare("PRAGMA wal_autocheckpoint").get() as { wal_autocheckpoint: number };
       const journalSizeLimit = db.prepare("PRAGMA journal_size_limit").get() as { journal_size_limit: number };
 
-      expect(synchronous.synchronous).toBe(1); // NORMAL
-      expect(autoCheckpoint.wal_autocheckpoint).toBe(100);
+      expect(synchronous.synchronous).toBe(2); // FULL
+      expect(autoCheckpoint.wal_autocheckpoint).toBe(1000);
       expect(journalSizeLimit.journal_size_limit).toBe(4_194_304);
     });
 

packages/core/src/db.ts

@@ -1319,12 +1319,15 @@ export class Database {
       this.db.exec(`PRAGMA busy_timeout = ${this.busyTimeoutMs}`);
       // Enable WAL mode for concurrent reader/writer access
       this.db.exec("PRAGMA journal_mode = WAL");
-      // In WAL mode NORMAL is nearly as durable as FULL with much lower fsync cost.
-      this.db.exec("PRAGMA synchronous = NORMAL");
-      // Checkpoint every 100 pages (~400 KB) to keep WAL small and reduce
-      // corruption risk. More aggressive than the default 1000, but paired
-      // with journal_size_limit to prevent WAL bloat.
-      this.db.exec("PRAGMA wal_autocheckpoint = 100");
+      // FULL fsyncs on every commit. Slightly slower than NORMAL, but the only
+      // setting that survives a process crash mid-checkpoint without torn pages
+      // — repeated node:sqlite SIGSEGVs inside pager_write have corrupted this
+      // db before.
+      this.db.exec("PRAGMA synchronous = FULL");
+      // Default (1000) checkpoint cadence. The previous value of 100 made the
+      // db spend most of its life mid-checkpoint, multiplying corruption risk
+      // when a writer crashed. journal_size_limit below still caps WAL growth.
+      this.db.exec("PRAGMA wal_autocheckpoint = 1000");
       // Bound WAL growth between checkpoints/maintenance cycles.
       this.db.exec("PRAGMA journal_size_limit = 4194304");
     } else {