From dadcfb574f5415e22c927a0608404c6b2750831c Mon Sep 17 00:00:00 2001
From: Peter Steinberger <steipete@gmail.com>
Date: Fri, 17 Apr 2026 19:38:53 +0100
Subject: [PATCH] test: trim surrogate chunk fixtures

---
 packages/memory-host-sdk/src/host/internal.test.ts | 7 +++----
 src/memory-host-sdk/host/internal.test.ts          | 7 +++----
 2 files changed, 6 insertions(+), 8 deletions(-)

diff --git a/packages/memory-host-sdk/src/host/internal.test.ts b/packages/memory-host-sdk/src/host/internal.test.ts
index 45adf405867..6aa17953741 100644
--- a/packages/memory-host-sdk/src/host/internal.test.ts
+++ b/packages/memory-host-sdk/src/host/internal.test.ts
@@ -368,11 +368,10 @@ describe("chunkMarkdown", () => {
   });
   it("does not break surrogate pairs when splitting long CJK lines", () => {
     // "𠀀" (U+20000) is a surrogate pair: 2 UTF-16 code units per character.
-    // A line of 500 such characters = 1000 UTF-16 code units.
-    // With tokens=99 (odd), the fine-split must not cut inside a pair.
+    // With an odd token budget, the fine-split must not cut inside a pair.
     const surrogateChar = "\u{20000}"; // 𠀀
-    const longLine = surrogateChar.repeat(500);
-    const chunks = chunkMarkdown(longLine, { tokens: 99, overlap: 0 });
+    const longLine = surrogateChar.repeat(120);
+    const chunks = chunkMarkdown(longLine, { tokens: 31, overlap: 0 });
     for (const chunk of chunks) {
       // No chunk should contain the Unicode replacement character U+FFFD,
       // which would indicate a broken surrogate pair.
diff --git a/src/memory-host-sdk/host/internal.test.ts b/src/memory-host-sdk/host/internal.test.ts
index a68d1a98137..298a3e78dd4 100644
--- a/src/memory-host-sdk/host/internal.test.ts
+++ b/src/memory-host-sdk/host/internal.test.ts
@@ -360,11 +360,10 @@ describe("chunkMarkdown", () => {
   });
   it("does not break surrogate pairs when splitting long CJK lines", () => {
     // "𠀀" (U+20000) is a surrogate pair: 2 UTF-16 code units per character.
-    // A line of 500 such characters = 1000 UTF-16 code units.
-    // With tokens=99 (odd), the fine-split must not cut inside a pair.
+    // With an odd token budget, the fine-split must not cut inside a pair.
     const surrogateChar = "\u{20000}"; // 𠀀
-    const longLine = surrogateChar.repeat(500);
-    const chunks = chunkMarkdown(longLine, { tokens: 99, overlap: 0 });
+    const longLine = surrogateChar.repeat(120);
+    const chunks = chunkMarkdown(longLine, { tokens: 31, overlap: 0 });
     for (const chunk of chunks) {
       // No chunk should contain the Unicode replacement character U+FFFD,
       // which would indicate a broken surrogate pair.