Fixed bug #42488 (SoapServer reports an encoding error and the error itself breaks).

1 files changed, 41 insertions, 4 deletions

diff --git a/ext/soap/php_encoding.c b/ext/soap/php_encoding.c
index 65e703d5d2..8354f87ace 100644
--- a/ext/soap/php_encoding.c
+++ b/ext/soap/php_encoding.c
@@ -864,13 +864,50 @@ static xmlNodePtr to_xml_string(encodeTypePtr type, zval *data, int style, xmlNo
 			efree(str);
 			str = estrdup((char*)xmlBufferContent(out));
 			new_len = n;
-		} else if (!php_libxml_xmlCheckUTF8(BAD_CAST(str))) {
-			soap_error1(E_ERROR,  "Encoding: string '%s' is not a valid utf-8 string", str);
 		}
 		xmlBufferFree(out);
 		xmlBufferFree(in);
-	} else if (!php_libxml_xmlCheckUTF8(BAD_CAST(str))) {
-		soap_error1(E_ERROR,  "Encoding: string '%s' is not a valid utf-8 string", str);
+	}
+
+	if (!php_libxml_xmlCheckUTF8(BAD_CAST(str))) {
+		char *err = emalloc(new_len + 8);
+		char c;
+		int i;
+		
+		memcpy(err, str, new_len+1);
+		for (i = 0; (c = err[i++]);) {
+			if ((c & 0x80) == 0) {
+			} else if ((c & 0xe0) == 0xc0) {
+				if ((err[i] & 0xc0) != 0x80) {
+					break;
+				}
+				i++;
+			} else if ((c & 0xf0) == 0xe0) {
+				if ((err[i] & 0xc0) != 0x80 || (err[i+1] & 0xc0) != 0x80) {
+					break;
+				}
+				i += 2;
+			} else if ((c & 0xf8) == 0xf0) {
+				if ((err[i] & 0xc0) != 0x80 || (err[i+1] & 0xc0) != 0x80 || (err[i+2] & 0xc0) != 0x80) {
+					break;
+				}
+				i += 3;
+			} else {
+				break;
+			}
+		}
+		if (c) {
+			err[i-1] = '\\';
+			err[i++] = 'x';
+			err[i++] = ((unsigned char)c >> 4) + ((((unsigned char)c >> 4) > 9) ? ('a' - 10) : '0');
+			err[i++] = (c & 15) + (((c & 15) > 9) ? ('a' - 10) : '0');
+			err[i++] = '.';
+			err[i++] = '.';
+			err[i++] = '.';
+			err[i++] = 0;
+		}
+
+		soap_error1(E_ERROR,  "Encoding: string '%s' is not a valid utf-8 string", err);
 	}
 
 	text = xmlNewTextLen(BAD_CAST(str), new_len);