fix(imessage): drop dangling surrogates in debounced merge preview and probe snippet

wangmiao0668000666 · wangmiao0668000666 · commit 6fb65781883e · 2026-06-30T14:45:16.000+08:00
diff --git a/extensions/imessage/src/monitor/monitor-provider.truncation.test.ts b/extensions/imessage/src/monitor/monitor-provider.truncation.test.ts
@@ -0,0 +1,37 @@
+// Imessage tests cover the UTF-16-safe debounced-merge preview in
+// monitor-provider.ts:731. Verifies that `sliceUtf16Safe` drops a surrogate
+// pair that straddles the 50-char truncation boundary instead of leaving a
+// lone high-surrogate half in the preview, and that the existing conditional
+// `"..."` ellipsis at L732 still fires correctly when the original input
+// exceeded the cap.
+import { describe, expect, it } from "vitest";
+import { sliceUtf16Safe } from "openclaw/plugin-sdk/text-utility-runtime";
+
+describe("monitor-provider debounced-merge preview truncation", () => {
+  const emoji = "🎉";
+
+  it("sliceUtf16Safe drops a trailing surrogate straddling the 50-char boundary", () => {
+    const input = "a".repeat(49) + emoji;
+    const out = sliceUtf16Safe(input, 0, 50);
+    expect(out.length).toBe(49);
+    expect(out.charCodeAt(out.length - 1)).toBeLessThan(0xd800);
+  });
+
+  it("ellipsis marker is appended when the original text exceeded 50 chars (no surrogate regression)", () => {
+    const input = "a".repeat(60);
+    const out = sliceUtf16Safe(input, 0, 50);
+    const ellipsis = input.length > 50 ? "..." : "";
+    expect(out + ellipsis).toBe("a".repeat(50) + "...");
+  });
+
+  it("empty input stays empty", () => {
+    expect(sliceUtf16Safe("", 0, 50)).toBe("");
+  });
+
+  it("emoji fully inside the window is preserved (no false-positive drops)", () => {
+    const input = emoji + "a".repeat(50);
+    const out = sliceUtf16Safe(input, 0, 50);
+    expect(out.startsWith(emoji)).toBe(true);
+    expect(out.length).toBe(50);
+  });
+});
diff --git a/extensions/imessage/src/monitor/monitor-provider.ts b/extensions/imessage/src/monitor/monitor-provider.ts
@@ -45,7 +45,10 @@ import {
   resolveSendPolicy,
   resolveStorePath,
 } from "openclaw/plugin-sdk/session-store-runtime";
-import { truncateUtf16Safe } from "openclaw/plugin-sdk/text-utility-runtime";
+import {
+  sliceUtf16Safe,
+  truncateUtf16Safe,
+} from "openclaw/plugin-sdk/text-utility-runtime";
 import { waitForTransportReady } from "openclaw/plugin-sdk/transport-ready-runtime";
 import { resolveIMessageAccount } from "../accounts.js";
 import { pollPendingIMessageApprovalReactions } from "../approval-reaction-poller.js";
@@ -728,7 +731,7 @@ export async function monitorIMessageProvider(opts: MonitorIMessageOpts = {}): P
       const combined = combineIMessagePayloads(messages);
       if (shouldLogVerbose()) {
         const text = combined.text ?? "";
-        const preview = text.slice(0, 50);
+        const preview = sliceUtf16Safe(text, 0, 50);
         const ellipsis = text.length > 50 ? "..." : "";
         logVerbose(
           `[imessage] merged ${entries.length} debounced messages: "${preview}${ellipsis}"`,
diff --git a/extensions/imessage/src/probe.test.ts b/extensions/imessage/src/probe.test.ts
@@ -1,5 +1,6 @@
 // Imessage tests cover probe plugin behavior.
 import { afterEach, describe, expect, it, vi } from "vitest";
+import { truncateUtf16Safe } from "openclaw/plugin-sdk/text-utility-runtime";
 import {
   clearCachedIMessagePrivateApiStatus,
   getCachedIMessagePrivateApiStatus,
@@ -135,3 +136,35 @@ describe("iMessage private API status cache", () => {
     expect(getCachedIMessagePrivateApiStatus("imsg-overflow-private-clock")).toBeUndefined();
   });
 });
+
+// Verifies the `lines[0]?.slice(0, 120)` snippet site in probe.ts:161 drops a
+// surrogate pair that straddles the truncation boundary instead of leaving a
+// lone high-surrogate half in the diagnostic preview.
+describe("probe first-line snippet truncation", () => {
+  const emoji = "🎉";
+
+  it("truncateUtf16Safe drops a trailing surrogate straddling the 120-char boundary", () => {
+    const input = "a".repeat(119) + emoji;
+    const out = truncateUtf16Safe(input, 120);
+    expect(out.length).toBe(119);
+    expect(out.charCodeAt(out.length - 1)).toBeLessThan(0xd800);
+  });
+
+  it("plain ASCII under the cap passes through unchanged", () => {
+    const input = "x".repeat(50);
+    expect(truncateUtf16Safe(input, 120)).toBe(input);
+  });
+
+  it("undefined input is preserved (no truncation attempted)", () => {
+    const lines: string[] = [];
+    const snippet = lines[0] ? truncateUtf16Safe(lines[0], 120) : undefined;
+    expect(snippet).toBeUndefined();
+  });
+
+  it("emoji fully inside the 120-char window is preserved", () => {
+    const input = emoji + "a".repeat(120);
+    const out = truncateUtf16Safe(input, 120);
+    expect(out.startsWith(emoji)).toBe(true);
+    expect(out.length).toBe(120);
+  });
+});
diff --git a/extensions/imessage/src/probe.ts b/extensions/imessage/src/probe.ts
@@ -13,6 +13,7 @@ import {
   normalizeLowercaseStringOrEmpty,
   normalizeStringEntries,
 } from "openclaw/plugin-sdk/string-coerce-runtime";
+import { truncateUtf16Safe } from "openclaw/plugin-sdk/text-utility-runtime";
 import { createIMessageRpcClient } from "./client.js";
 import { DEFAULT_IMESSAGE_PROBE_TIMEOUT_MS } from "./constants.js";
 import {
@@ -158,7 +159,7 @@ function parseStatusPayload(stdout: string): {
   // No JSONL line parsed. Surface a small snippet of the first non-empty
   // line so the operator can grep imsg release notes if the status output
   // schema has shifted.
-  const snippet = lines[0]?.slice(0, 120);
+  const snippet = lines[0] ? truncateUtf16Safe(lines[0], 120) : undefined;
   return { payload: null, firstLineSnippet: snippet };
 }
 
