Another update related to the sqlmapproject#1539

stamparm · stamparm · commit 94639d11a340 · 2015-11-16T15:33:05.000+01:00
diff --git a/lib/core/agent.py b/lib/core/agent.py
@@ -37,6 +37,7 @@
 from lib.core.settings import CUSTOM_INJECTION_MARK_CHAR
 from lib.core.settings import DEFAULT_COOKIE_DELIMITER
 from lib.core.settings import DEFAULT_GET_POST_DELIMITER
+from lib.core.settings import DEFAULT_MYSQL_CHARACTER_SET
 from lib.core.settings import GENERIC_SQL_COMMENT
 from lib.core.settings import PAYLOAD_DELIMITER
 from lib.core.settings import REPLACEMENT_MARKER
@@ -400,7 +401,10 @@ def nullAndCastField(self, field):
                 nulledCastedField = field
             else:
                 if not (Backend.isDbms(DBMS.SQLITE) and not isDBMSVersionAtLeast('3')):
-                    nulledCastedField = rootQuery.cast.query % field
+                    if Backend.isDbms(DBMS.MYSQL):
+                        nulledCastedField = rootQuery.cast.query.replace(")", " CHARACTER SET %s)") % (field, DEFAULT_MYSQL_CHARACTER_SET)
+                    else:
+                        nulledCastedField = rootQuery.cast.query % field
                 if Backend.getIdentifiedDbms() in (DBMS.ACCESS,):
                     nulledCastedField = rootQuery.isnull.query % (nulledCastedField, nulledCastedField)
                 else:
diff --git a/lib/core/settings.py b/lib/core/settings.py
@@ -224,6 +224,10 @@
 
 HSQLDB_DEFAULT_SCHEMA = "PUBLIC"
 
+# Default character set used in MySQL
+# Reference: http://pieroxy.net/blog/2013/05/28/mysql_charset_hell.html
+DEFAULT_MYSQL_CHARACTER_SET = "latin1"
+
 # Names that can't be used to name files on Windows OS
 WINDOWS_RESERVED_NAMES = ("CON", "PRN", "AUX", "NUL", "COM1", "COM2", "COM3", "COM4", "COM5", "COM6", "COM7", "COM8", "COM9", "LPT1", "LPT2", "LPT3", "LPT4", "LPT5", "LPT6", "LPT7", "LPT8", "LPT9")
 
diff --git a/lib/techniques/blind/inference.py b/lib/techniques/blind/inference.py
@@ -8,6 +8,7 @@
 import threading
 import time
 
+from extra.safe2bin.safe2bin import safechardecode
 from extra.safe2bin.safe2bin import safecharencode
 from lib.core.agent import agent
 from lib.core.common import Backend
@@ -18,6 +19,7 @@
 from lib.core.common import filterControlChars
 from lib.core.common import getCharset
 from lib.core.common import getCounter
+from lib.core.common import getUnicode
 from lib.core.common import goGoodSamaritan
 from lib.core.common import getPartRun
 from lib.core.common import hashDBRetrieve
@@ -35,6 +37,7 @@
 from lib.core.enums import PAYLOAD
 from lib.core.exception import SqlmapThreadException
 from lib.core.settings import CHAR_INFERENCE_MARK
+from lib.core.settings import DEFAULT_MYSQL_CHARACTER_SET
 from lib.core.settings import INFERENCE_BLANK_BREAK
 from lib.core.settings import INFERENCE_UNKNOWN_CHAR
 from lib.core.settings import INFERENCE_GREATER_CHAR
@@ -589,6 +592,10 @@ def blindThread():
         raise KeyboardInterrupt
 
     _ = finalValue or partialValue
+
+    if Backend.isDbms(DBMS.MYSQL) and safechardecode(_) != _:
+        _ = getUnicode(safechardecode(_).encode(DEFAULT_MYSQL_CHARACTER_SET))
+
     return getCounter(kb.technique), safecharencode(_) if kb.safeCharEncode else _
 
 def queryOutputLength(expression, payload):
diff --git a/plugins/dbms/mysql/syntax.py b/plugins/dbms/mysql/syntax.py
@@ -8,6 +8,7 @@
 import binascii
 
 from lib.core.convert import utf8encode
+from lib.core.settings import DEFAULT_MYSQL_CHARACTER_SET
 from plugins.generic.syntax import Syntax as GenericSyntax
 
 class Syntax(GenericSyntax):
@@ -26,7 +27,7 @@ def escaper(value):
             try:
                 retVal = "0x%s" % binascii.hexlify(value)
             except UnicodeEncodeError:
-                retVal = "CONVERT(0x%s USING utf8)" % "".join("%.2x" % ord(_) for _ in utf8encode(value))
+                retVal = "CONVERT(0x%s USING %s)" % ("".join("%.2x" % ord(_) for _ in utf8encode(value)), DEFAULT_MYSQL_CHARACTER_SET)
             return retVal
 
         return Syntax._escape(expression, quote, escaper)
diff --git a/plugins/generic/enumeration.py b/plugins/generic/enumeration.py
@@ -31,6 +31,7 @@ def __init__(self):
         kb.data.banner = None
         kb.data.hostname = ""
         kb.data.processChar = None
+        kb.data.characterSet = None
 
         Custom.__init__(self)
         Databases.__init__(self)
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 CHARACTER SET latin1)"/>
+        <cast query="CAST(%s AS CHAR)"/>
         <length query="CHAR_LENGTH(%s)"/>
         <isnull query="IFNULL(%s,' ')"/>
         <delimiter query=","/>
