@mem-weave/server 0.2.0 → 0.2.2

package/dist/cli-entry.js

@@ -25,7 +25,7 @@ async function main() {
         env: process.env,
         configPath: process.env.MEMWEAVE_CONFIG
     });
-    // Print message + data, and exit non-zero on failure
+    // Print message + data
     if (result.message) {
         if (result.ok) {
             // eslint-disable-next-line no-console
@@ -40,6 +40,10 @@ async function main() {
         // eslint-disable-next-line no-console
         console.log(JSON.stringify(result.data, null, 2));
     }
+    // For the `start` command, the HTTP server keeps the event loop alive —
+    // do NOT call process.exit() or it would kill the server immediately.
+    if (parsed.command === 'start')
+        return;
     process.exit(result.ok ? 0 : 1);
 }
 main().catch((err) => {

package/dist/db/database.js

@@ -1,8 +1,10 @@
 import { mkdirSync } from 'node:fs';
 import { dirname } from 'node:path';
+import { createRequire } from 'node:module';
 import Database from 'better-sqlite3';
 import { SCHEMA_SQL } from './schema.js';
 import { logger } from '../server/logger.js';
+const _require = createRequire(import.meta.url);
 /**
  * Default dimensions for the memory_vectors vec0 table. The actual configured
  * dimensions come from the EmbeddingConfig; the table is created lazily by
@@ -36,7 +38,7 @@ export function openDatabase(path, options = {}) {
             // Use a dynamic import to avoid hard-failing test environments that
             // exercise only non-vector code paths.
             // eslint-disable-next-line @typescript-eslint/no-var-requires
-            const sqliteVec = require('sqlite-vec');
+            const sqliteVec = _require('sqlite-vec');
             sqliteVec.load(db);
             const dims = options.vectorDimensions ?? VECTOR_DEFAULT_DIMENSIONS;
             ensureVecTable(db, dims);

package/package.json

@@ -1,12 +1,12 @@
 {
   "name": "@mem-weave/server",
-  "version": "0.2.0",
+  "version": "0.2.2",
   "description": "MemWeave local-first memory infrastructure for AI agents: structured memory, 4-layer retrieval (BM25 + vector + graph + causal), token-budgeted injection, server-side write deduplication, and background consolidation. Server process: Fastify REST API + CLI.",
   "type": "module",
   "main": "./dist/cli-entry.js",
   "types": "./dist/cli-entry.d.ts",
   "bin": {
-    "memweave": "dist/cli-entry.js"
+    "memweave": "./dist/cli-entry.js"
   },
   "files": [
     "dist",