test(matrix): prove UTF-16 truncation boundaries

steipete · steipete · commit 7d7aba160485 · 2026-07-09T06:07:50.000+01:00
diff --git a/extensions/matrix/src/matrix/monitor/handler.test.ts b/extensions/matrix/src/matrix/monitor/handler.test.ts
@@ -3184,6 +3184,31 @@ describe("matrix monitor handler draft streaming", () => {
     await finish();
   });
 
+  it("keeps truncated Matrix tool progress UTF-16 safe", async () => {
+    const { dispatch } = createStreamingHarness({
+      streaming: "progress",
+      previewToolProgressEnabled: true,
+      accountConfig: {
+        streaming: {
+          mode: "progress",
+          progress: { label: false, maxLineChars: 500 },
+        },
+      } as never,
+    });
+    const { opts, finish } = await dispatch();
+    const progressPrefix = "x".repeat(298);
+
+    const progressText = `${progressPrefix}🎉tail`;
+    await opts.onItemEvent?.({ progressText });
+    await opts.onItemEvent?.({ progressText });
+
+    await vi.waitFor(() => {
+      expect(sendSingleTextMessageMatrixMock).toHaveBeenCalledTimes(1);
+    });
+    expect(singleTextMessageBody()).toBe(`- \`${progressPrefix}...\``);
+    await finish();
+  });
+
   it("replaces recovered Matrix command progress instead of leaving stale failed text", async () => {
     const { dispatch } = createStreamingHarness({
       streaming: "progress",
diff --git a/extensions/matrix/src/matrix/sdk/event-helpers.test.ts b/extensions/matrix/src/matrix/sdk/event-helpers.test.ts
@@ -41,35 +41,12 @@ describe("event-helpers", () => {
     expect(fromText.statusCode).toBe(500);
   });
 
-  it("buildHttpError keeps the 500-char plain-text body truncation UTF-16 safe at emoji boundaries", () => {
-    // 498 ASCII chars + emoji (😀 = U+1F600, UTF-16 surrogate pair) = 500 code units.
-    // A naive .slice(0, 500) would land on the high surrogate and return a dangling
-    // high surrogate (charCode 0xD83D), which downstream string consumers render
-    // as U+FFFD / mojibake. truncateUtf16Safe trims the orphan high surrogate
-    // and returns 499 code units ending cleanly before the emoji.
-    const ascii = "a".repeat(498);
-    const emoji = "😀"; // 2 UTF-16 code units
-    const bodyText = ascii + emoji + "tail";
-    const error = buildHttpError(500, bodyText);
-    expect(error.message.length).toBe(499);
-    expect(error.message).toBe(ascii);
-    expect(error.message.charCodeAt(error.message.length - 1)).not.toBe(0xd83d);
-  });
+  it("keeps truncated HTTP error bodies UTF-16 safe", () => {
+    const parsedPrefix = `{"detail":"${"a".repeat(488)}`;
+    const invalidPrefix = `not-json ${"b".repeat(490)}`;
 
-  it("buildHttpError keeps a non-parseable JSON body truncation UTF-16 safe at emoji boundaries", () => {
-    // 498 ASCII chars + emoji = 500 code units. The try/catch branch falls back
-    // to truncateUtf16Safe(bodyText, 500); before this fix the same dangling
-    // high surrogate would have leaked into the error message.
-    const ascii = "b".repeat(498);
-    const emoji = "🎉"; // 2 UTF-16 code units
-    const bodyText = "not-json " + ascii + emoji + "more";
-    const error = buildHttpError(502, bodyText);
-    // The slice lands on the high surrogate of 🎉, which truncateUtf16Safe
-    // trims — final length is 499 code units (one less than the cap).
-    expect(error.message.length).toBe(499);
-    // Final code unit is the 498th ASCII 'b' (0x62), not a dangling high surrogate.
-    expect(error.message.charCodeAt(error.message.length - 1)).toBe(0x62);
-    expect(error.message).not.toMatch(/�/);
+    expect(buildHttpError(500, `${parsedPrefix}😀"}`).message).toBe(parsedPrefix);
+    expect(buildHttpError(502, `${invalidPrefix}🎉tail`).message).toBe(invalidPrefix);
   });
 
   it("serializes Matrix events and resolves state key from available sources", () => {
diff --git a/extensions/matrix/src/matrix/sdk/event-helpers.ts b/extensions/matrix/src/matrix/sdk/event-helpers.ts
@@ -1,5 +1,5 @@
-import type { MatrixEvent } from "matrix-js-sdk/lib/matrix.js";
 // Matrix helper module supports event helpers behavior.
+import type { MatrixEvent } from "matrix-js-sdk/lib/matrix.js";
 import { truncateUtf16Safe } from "openclaw/plugin-sdk/text-utility-runtime";
 import type { MatrixRawEvent } from "./types.js";
 
