dynamicity engine update

lib/controller/checks.py

@@ -15,7 +15,6 @@ from difflib import SequenceMatcher
 
 from lib.core.agent import agent
 from lib.core.common import beep
-from lib.core.common import getFilteredPageContent
 from lib.core.common import getUnicode
 from lib.core.common import randomInt
 from lib.core.common import randomStr
@@ -28,6 +27,7 @@ from lib.core.data import kb
 from lib.core.data import logger
 from lib.core.data import paths
 from lib.core.exception import sqlmapConnectionException
+from lib.core.exception import sqlmapGenericException
 from lib.core.exception import sqlmapNoneDataException
 from lib.core.exception import sqlmapUserQuitException
 from lib.core.exception import sqlmapSilentQuitException
@@ -219,6 +219,8 @@ def checkStability():
     time.sleep(1)
     secondPage, _ = Request.queryPage(content=True)
 
+    conf.seqMatcher.set_seq1(firstPage)
+
     kb.pageStable = (firstPage == secondPage)
 
     if kb.pageStable:
@@ -283,6 +285,11 @@ def checkStability():
         else:
             checkDynamicContent(firstPage, secondPage)
 
+            if not Request.queryPage():
+                errMsg  = "target url is too dynamic. unable to continue. consider using other methods"
+                logger.error(errMsg)
+                raise sqlmapSilentQuitException
+
     return kb.pageStable
 
 def checkString():
@@ -386,8 +393,7 @@ def checkConnection():
     logger.info(infoMsg)
 
     try:
-        page, _ = Request.getPage()
+        Request.getPage()
-        conf.seqMatcher.set_seq1(page if not conf.textOnly else getFilteredPageContent(page))
 
     except sqlmapConnectionException, errMsg:
         errMsg = getUnicode(errMsg)

lib/core/common.py

@@ -41,6 +41,7 @@ from lib.core.data import kb
 from lib.core.data import logger
 from lib.core.data import paths
 from lib.core.data import queries
+from lib.core.convert import htmlunescape
 from lib.core.convert import urlencode
 from lib.core.exception import sqlmapFilePathException
 from lib.core.exception import sqlmapGenericException
@@ -1102,6 +1103,8 @@ def getFilteredPageContent(page):
         while retVal.find("  ") != -1:
             retVal = retVal.replace("  ", " ")
 
+        retVal = htmlunescape(retVal)
+
     return retVal
 
 def getPageTextWordsSet(page):

lib/core/convert.py

@@ -93,7 +93,7 @@ def utf8decode(string):
     return string.decode("utf-8")
 
 def htmlescape(string):
-    return string.replace('&', '&amp;').replace('<', '&lt;').replace('>', '&gt;').replace('"', '&quot;').replace("'", '&#39;')
+    return string.replace('&', '&amp;').replace('<', '&lt;').replace('>', '&gt;').replace('"', '&quot;').replace("'", '&#39;').replace(' ', '&nbsp;')
 
 def htmlunescape(string):
-    return string.replace('&amp;', '&').replace('&lt;', '<').replace('&gt;', '>').replace('&quot;', '"').replace('&#39;', "'")
+    return string.replace('&amp;', '&').replace('&lt;', '<').replace('&gt;', '>').replace('&quot;', '"').replace('&#39;', "'").replace('&nbsp;', ' ')

lib/request/comparison.py

@@ -9,7 +9,6 @@ See the file 'doc/COPYING' for copying permission
 
 import re
 
-from lib.core.common import getFilteredPageContent
 from lib.core.common import wasLastRequestError
 from lib.core.data import conf
 from lib.core.data import kb
@@ -50,7 +49,7 @@ def comparison(page, headers=None, getSeqMatcher=False, pageLength=None):
         if conf.regexp:
             return re.search(conf.regexp, page, re.I | re.M) is not None
 
-        # Dynamic content lines to be excluded before calculating page hash
+        # Dynamic content lines to be excluded before comparison
         if not kb.nullConnection:
             for item in kb.dynamicMarkings:
                 prefix, postfix = item
@@ -72,7 +71,7 @@ def comparison(page, headers=None, getSeqMatcher=False, pageLength=None):
         if ratio > 1.:
             ratio = 1. / ratio
     else:
-        conf.seqMatcher.set_seq2(page if not conf.textOnly else getFilteredPageContent(page))
+        conf.seqMatcher.set_seq2(page)
         ratio = round(conf.seqMatcher.ratio(), 3)
 
     if kb.locks.seqLock:

lib/request/connect.py

@@ -18,6 +18,7 @@ import traceback
 from lib.contrib import multipartpost
 from lib.core.agent import agent
 from lib.core.common import readInput
+from lib.core.common import getFilteredPageContent
 from lib.core.common import getUnicode
 from lib.core.convert import urlencode
 from lib.core.common import urlEncodeCookieValues
@@ -366,6 +367,9 @@ class Connect:
 
         if not pageLength:
             page, headers = Connect.getPage(url=uri, get=get, post=post, cookie=cookie, ua=ua, silent=silent, method=method, auxHeaders=auxHeaders, response=response, raise404=raise404)
+            
+        if conf.textOnly:
+            page = getFilteredPageContent(page)
 
         if content or response:
             return page, headers