test(privacy): verify full wipe across core and Pro

Author: alichherawalla

docs/FUNCTIONAL_TEST_STRATEGY.md

@@ -28,10 +28,10 @@ really a test of a third-party framework or engine we merely depend on:
   75% branches / 77% functions / 79% lines on the TESTABLE surface (ratchet floor 77/74/76/78,
   climbing toward the 85% goal). ~1200 unit/integration tests.
 - **Swift, `npm run test:swift`:** 37 XCTest cases over the text-extractor pure classifiers.
-- **DB integration, `npm run test:db`:** 61 cases over `database.ts` + `rag/store.ts` against a
-  real temp SQLite. Kept OUT of the default gate because it needs `better-sqlite3` rebuilt for
-  the node ABI (the app builds it for Electron's ABI); the script rebuilds + restores. KNOWN GAP:
-  wire `test:db` into CI as an isolated step so database.ts coverage is enforced there.
+- **DB integration, `npm run test:db`:** 104 cases over core and Pro persistence against real
+  temp SQLite. Kept OUT of the default gate because it needs `better-sqlite3` rebuilt for the
+  node ABI (the app builds it for Electron's ABI); the script rebuilds + restores. KNOWN GAP:
+  wire `test:db` into CI as an isolated step so database coverage is enforced there.
 - The default coverage gate EXCLUDES (with a real alternative suite each): vendored/built
   (`packages/**`, `**/dist/**`); native/DB/spawn shells (database.ts, rag/store.ts, imagegen.ts,
   mflux.ts, sd-server.ts, model-server.ts, media-server.ts, whisper/parakeet/whisper-server) -

docs/P0_P2_INTEGRATION_COVERAGE.md

@@ -8,13 +8,13 @@ manual claim does not count as complete integration coverage.
 ## Current status - 2026-07-17
 
 - Scope: 155 journeys - 74 P0, 71 P1, and 10 P2.
-- Covered by direct integration or E2E evidence: 20 - 13 P0, 6 P1, and 1 P2.
-- Left: 135 - including partially covered journeys whose exact release behavior is not yet proven.
+- Covered by direct integration or E2E evidence: 21 - 14 P0, 6 P1, and 1 P2.
+- Left: 134 - including partially covered journeys whose exact release behavior is not yet proven.
 - Green gates today:
   - `npm test`: 203 files passed, 1 skipped; 2,190 tests passed, 1 skipped.
   - `npm run test:coverage`: 96.77% statements, 91.54% branches, 96.25% functions,
     97.55% lines.
-  - `npm run test:db`: 12 files and 103 real SQLite integration tests passed; Electron ABI
+  - `npm run test:db`: 13 files and 104 real SQLite integration tests passed; Electron ABI
     restored afterward.
   - `npm run test:e2e`: 22 Playwright Electron tests passed against fresh synthetic temp profiles.
   - Both TypeScript projects pass.
@@ -51,6 +51,10 @@ manual claim does not count as complete integration coverage.
   attachment persistence across lock and unlock.
 - #128 - Vault recovery and backup. `vault-service.test.ts` and `vault-recovery.test.ts` exercise
   real KDBX export bytes, recovery setup, wrong phrases, and recovery to a new password.
+- #136 - Delete all is complete. Core and Pro DB integration tests seed projects, chats, memory,
+  knowledge, connectors, encrypted tokens, profile data, every registered Pro table, and every
+  personal-data directory. They run the real delete-all registry, close and reopen the encrypted
+  database, and verify personal data stays gone while models and ordinary preferences survive.
 
 ## Covered P1 journeys
 
@@ -154,10 +158,10 @@ manual claim does not count as complete integration coverage.
 
 - #129 - Runtime residency toggles persist; #130 - Chat residency stays required; #131 - Resource
   mode applies; #132 - Settings survive relaunch; #133 - Storage usage is truthful; #134 - Clear
-  cache preserves user data; #135 - Delete category is scoped; #136 - Delete all is complete; #137
-  - Core locked Pro tabs; #138 - Pro license activates; #139 - Invalid or exhausted license fails
-    clearly; #140 - Offline entitlement behavior; #141 - Core and Pro override behavior; #142 - Manual
-    update check; #143 - Update channel persists.
+  cache preserves user data; #135 - Delete category is scoped; #137 - Core locked Pro tabs; #138 -
+  Pro license activates; #139 - Invalid or exhausted license fails
+  clearly; #140 - Offline entitlement behavior; #141 - Core and Pro override behavior; #142 - Manual
+  update check; #143 - Update channel persists.
 
 ## Left - resilience and desktop polish
 

scripts/test-db.sh

@@ -1,5 +1,5 @@
 #!/usr/bin/env bash
-# DB integration tests (src/main/__tests__/*.dbtest.ts) exercise the real data layer against
+# DB integration tests (core + pro *.dbtest.ts) exercise the real data layer against
 # a temp SQLite DB. They need better-sqlite3-multiple-ciphers built for the TEST RUNNER's node
 # ABI - but the app builds it for ELECTRON's ABI (via electron-builder install-app-deps). So:
 # rebuild for node, run the tests, then ALWAYS restore Electron's build - even on failure or

vitest.db.config.ts

@@ -1,11 +1,18 @@
 import { defineConfig } from 'vitest/config'
+import { resolve } from 'path'
 
-// Dedicated config for the DB integration tests (*.dbtest.ts). Kept separate from the
+// Dedicated config for core + Pro DB integration tests (*.dbtest.ts). Kept separate from the
 // default vitest run because they load the better-sqlite3 native module (see
 // scripts/test-db.sh). Run via `npm run test:db`.
 export default defineConfig({
+  resolve: {
+    alias: {
+      '@offgrid/core': resolve(__dirname, 'src'),
+      '@offgrid/pro': resolve(__dirname, 'pro')
+    }
+  },
   test: {
-    include: ['src/main/__tests__/*.dbtest.ts'],
+    include: ['src/main/__tests__/*.dbtest.ts', 'pro/main/__tests__/*.dbtest.ts'],
     exclude: ['node_modules/**', 'out/**', 'e2e/**']
   }
 })