Changeset 1484

Timestamp:

12/07/07 17:22:03 (9 months ago)

Author:

jose

Message:

[phoneyc]
- handle the <meta> redirect stuff as an href
- fix up URLs that lack a needed trailing '/'

Files:

• phoneyc/trunk/honeyclient.py (modified) (4 diffs)

phoneyc/trunk/honeyclient.py

@@ -379 +379 @@
-    # build, store, return js and vbs content
-    def __init__(self, contents, url):
+        url = self.fixup(url)
-        self.url = url
-        self.baseurl = '/'.join(url.split('/')[:-1])  # for relative links
@@ -442 +443 @@
-        self.links = self.hrefs + self.iframes + self.objects + self.imgs + self.frames
-
+    def fixup(self, url):
+        """
+        fixes a URL that may lack a required trailing slash. without it
+        we don't get proper relative URLs.
+        """
+        c = 0
+        for i in url:
+            if i == '/': c += 1
+        if c < 3: return '%s/' % url
+        else: return url
-
-class PageParser(SGMLParser):
@@ -520 +531 @@
-                                
-    def end_a(self):
+        pass
+
+    def start_meta(self, attrs):
+        for k, v in attrs:
+            if k.lower() == 'content' and 'url=' in v.lower():
+                url = v.split(';')
+                for u in url:
+                    if u.lower().startswith('url='):
+                        url = re.sub('URL=', '', u, re.IGNORECASE)
+                        url = url.replace('"', '')
+                        if url not in self.hrefs: self.hrefs.append(url)
+
+    def end_meta(self):
-        pass
-
@@ -592 +616 @@
-        if self.js_inScript: self.js_body.append(text)
-        if self.vbs_inScript: self.vbs_body.append(text)
-
-
-LINKS = ('http://www.google.com/', )