Integer Overflow leads to Buffer Underflow vulnerability in GLib’s Base64 encoding implementation. The flaw is caused by storing the output length returned by g_base64_encode_step() in a signed integer (gint). When encoding very large attacker-controlled input, this value may overflow and become negative. The resulting pointer arithmetic in g_base64_encode_close() leads to a buffer underflow and out-of-bounds heap write. Exploitation requires that an application passes untrusted or excessively large input to the Base64 encoding API and may result in application crashes or potential memory corruption.