Merge pull request #1046 from DataDog/marcosaia/fix/replace-buffer-with-func

[FIX] Replaced 'Buffer' with custom function to get UTF8 string byte length

Author: marco-saia-datadog

packages/core/src/rum/instrumentation/resourceTracking/requestProxy/XHRProxy/baggageHeaderUtils.ts

@@ -110,7 +110,7 @@ export function formatBaggageHeader(entries: Set<string>): string | null {
     }
 
     const headerValue = formattedParts.join(',');
-    const byteLength = Buffer.byteLength(headerValue, 'utf8');
+    const byteLength = utf8ByteLength(headerValue);
 
     if (byteLength > MAX_BYTES) {
         InternalLog.log(
@@ -122,6 +122,40 @@ export function formatBaggageHeader(entries: Set<string>): string | null {
     return headerValue;
 }
 
+/**
+ * Returns the number of bytes needed to encode a string in UTF-8.
+ *
+ * Useful as a lightweight alternative to Node.js `Buffer.byteLength()`
+ * for older environments that do not support it.
+ *
+ * @param text - The input string.
+ * @returns The UTF-8 byte length of the string.
+ */
+function utf8ByteLength(text: string): number {
+    let byteLength = text.length;
+    for (let i = text.length - 1; i >= 0; i--) {
+        const code = text.charCodeAt(i);
+
+        // 2-byte characters (U+0080 to U+07FF)
+        if (code > 0x7f && code <= 0x7ff) {
+            byteLength++;
+        }
+        // 3-byte characters (U+0800 to U+FFFF)
+        else if (code > 0x7ff && code <= 0xffff) {
+            byteLength += 2;
+        }
+
+        // Handle surrogate pairs (4-byte characters, e.g. emoji)
+        // These characters already count as 2 in the initial length
+        // Encountering the low surrogate already accounts for the full 4 bytes
+        // (2 from the initial length + 2 for the 3-byte characters logic above)
+        if (code >= 0xdc00 && code <= 0xdfff) {
+            i--; // prevents double counting the same character by skipping high surrogate
+        }
+    }
+    return byteLength;
+}
+
 /**
  * Returns a set of valid baggage header characters.
  */