Another patch related to #1539 (simplifying unicode bad chars and preventing double encoding of safe chars)

stamparm · stamparm · commit 5593bf2fee7c · 2015-11-16T15:02:30.000+01:00
diff --git a/extra/safe2bin/safe2bin.py b/extra/safe2bin/safe2bin.py
@@ -19,15 +19,18 @@
 # Regex used for recognition of hex encoded characters
 HEX_ENCODED_CHAR_REGEX = r"(?P<result>\\x[0-9A-Fa-f]{2})"
 
-# Regex used for recognition of representation for hex encoded invalid unicode characters
-INVALID_UNICODE_CHAR_REGEX = r"(?P<result>\\\?[0-9A-Fa-f]{2})"
-
 # Raw chars that will be safe encoded to their slash (\) representations (e.g. newline to \n)
 SAFE_ENCODE_SLASH_REPLACEMENTS = "\t\n\r\x0b\x0c"
 
 # Characters that don't need to be safe encoded
 SAFE_CHARS = "".join(filter(lambda x: x not in SAFE_ENCODE_SLASH_REPLACEMENTS, string.printable.replace('\\', '')))
 
+# Prefix used for hex encoded values
+HEX_ENCODED_PREFIX = r"\x"
+
+# Strings used for temporary marking of hex encoded prefixes (to prevent double encoding)
+HEX_ENCODED_PREFIX_MARKER = "__HEX_ENCODED_PREFIX__"
+
 # String used for temporary marking of slash characters
 SLASH_MARKER = "__SLASH__"
 
@@ -45,6 +48,7 @@ def safecharencode(value):
 
     if isinstance(value, basestring):
         if any(_ not in SAFE_CHARS for _ in value):
+            retVal = retVal.replace(HEX_ENCODED_PREFIX, HEX_ENCODED_PREFIX_MARKER)
             retVal = retVal.replace('\\', SLASH_MARKER)
 
             for char in SAFE_ENCODE_SLASH_REPLACEMENTS:
@@ -53,6 +57,7 @@ def safecharencode(value):
             retVal = reduce(lambda x, y: x + (y if (y in string.printable or isinstance(value, unicode) and ord(y) >= 160) else '\\x%02x' % ord(y)), retVal, (unicode if isinstance(value, unicode) else str)())
 
             retVal = retVal.replace(SLASH_MARKER, "\\\\")
+            retVal = retVal.replace(HEX_ENCODED_PREFIX_MARKER, HEX_ENCODED_PREFIX)
     elif isinstance(value, list):
         for i in xrange(len(value)):
             retVal[i] = safecharencode(value[i])
@@ -83,12 +88,6 @@ def safechardecode(value, binary=False):
         if binary:
             if isinstance(retVal, unicode):
                 retVal = retVal.encode("utf8")
-            while True:
-                match = re.search(INVALID_UNICODE_CHAR_REGEX, retVal)
-                if match:
-                    retVal = retVal.replace(match.group("result"), chr(ord(binascii.unhexlify(match.group("result").lstrip("\\?")))))
-                else:
-                    break
 
     elif isinstance(value, (list, tuple)):
         for i in xrange(len(value)):
diff --git a/lib/core/common.py b/lib/core/common.py
@@ -2954,6 +2954,10 @@ def decodeIntToUnicode(value):
                     retVal = unichr(value)
                 else:
                     retVal = getUnicode(raw, conf.charset)
+
+                if Backend.isDbms(DBMS.MYSQL):
+                    import pdb
+                    pdb.set_trace()
             else:
                 retVal = getUnicode(chr(value))
         except:
diff --git a/lib/core/settings.py b/lib/core/settings.py
@@ -587,7 +587,7 @@
 LIMITED_ROWS_TEST_NUMBER = 15
 
 # Format used for representing invalid unicode characters
-INVALID_UNICODE_CHAR_FORMAT = r"\?%02x"
+INVALID_UNICODE_CHAR_FORMAT = r"\x%02x"
 
 # Regular expression for XML POST data
 XML_RECOGNITION_REGEX = r"(?s)\A\s*<[^>]+>(.+>)?\s*\Z"
diff --git a/xml/queries.xml b/xml/queries.xml
@@ -3,7 +3,7 @@
 <root>
     <!-- MySQL -->
     <dbms value="MySQL">
-        <cast query="CAST(%s AS CHAR)"/>
+        <cast query="CAST(%s AS CHAR CHARACTER SET latin1)"/>
         <length query="CHAR_LENGTH(%s)"/>
         <isnull query="IFNULL(%s,' ')"/>
         <delimiter query=","/>
