fix: keep surrogate numeric character references from failing the mail list

packages/integrations/src/google.ts

@@ -188,7 +188,10 @@ function decodeSnippet(value: string): string {
         entity[1].toLowerCase() === "x"
           ? Number.parseInt(entity.slice(2), 16)
           : Number.parseInt(entity.slice(1), 10);
-      return code >= 0 && code <= 0x10ffff ? String.fromCodePoint(code) : original;
+      // Surrogates are not scalar values: fromCodePoint throws on them, and one
+      // such reference in a snippet must not reject the whole message list.
+      const scalar = code >= 0 && code <= 0x10ffff && !(code >= 0xd800 && code <= 0xdfff);
+      return scalar ? String.fromCodePoint(code) : original;
     },
   );
 }

tests/google.test.ts

@@ -139,6 +139,19 @@ test("header decoding preserves spaces encoded inside words and next to plain te
   assert.equal((await client.getThread("thread1"))[0].subject, "Re: Visit résumé notes");
 });
 
+test("a numeric character reference to a surrogate stays literal instead of failing the list", async () => {
+  // String.fromCodePoint throws on surrogate code points; a sender can put the
+  // literal text � in a snippet, and one such message must not reject
+  // the whole list (Promise.all over per-message parses).
+  const client = clientWith((request) =>
+    new URL(request.url).pathname.endsWith("/messages")
+      ? json({ messages: [{ id: "msg1" }] })
+      : json({ id: "msg1", threadId: "thread1", snippet: "Entity notes: � and �" }),
+  );
+  const [mail] = await client.listMail();
+  assert.equal(mail.body, "Entity notes: � and �");
+});
+
 test("a body charset TextDecoder does not know falls back to UTF-8 instead of failing the list", async () => {
   const charsets: Record<string, string> = { known: "utf-8", unknown: "unknown-8bit" };
   const bodies: Record<string, string> = { known: "Opening hours", unknown: "Café tickets" };