improved link parser (works even without "http://")

reject submissions with suspicious usernames or content

wortschlucker/src/wortschlucker.py

@@ -296,8 +296,17 @@ def get_markup_with_links(text):
         # only take the TLD part of the url
         short_name = url.split("/")[2]
         return """%s<a href="%s" rel="nofollow">%s</a>%s""" % (prefix, url, short_name, suffix)
+    def expand_protocol(match):
+        prefix, url, suffix = match.groups()
+        if url.lower().startswith("www") or suffix.startswith("/"):
+            # just prepend "http://"
+            return prefix + "http://" + url + suffix
+        else:
+            return prefix + url + suffix
     # surround all urls with html markup
     text = genshi.escape(text)
+    text = re.sub(r"(\A|\s|\()([a-zA-Z_\-\.]+\.[a-zA-Z]{2,4})(/|\)|\s|\Z)",
+            expand_protocol, text)
     text = re.sub(r"(\A|\s|\()(https?://[\w/\?\.\#=;,_\-\~&]*)(\)|\s|\Z)",
             get_link_markup, text)
     return get_markup_with_formatted_linebreaks(text, "<br />")
@@ -527,6 +536,16 @@ def publish_twitter_alert(text, key, secret,access_key,access_secret):
     except urllib2.URLError, e:
         print e.reason
 
+def check_spam_submitter_name(name):
+    lower_text = re.sub("[^a-z]", "", name)
+    upper_text = re.sub("[^A-Z]", "", name)
+    return (len(lower_text) + len(upper_text) == len(name)) and \
+            (len(lower_text) > 3) and (len(upper_text) > 3) and \
+            (len(name) >= 8) and (not name.startswith(upper_text))
+
+def check_spam_content(text):
+    return bool(re.search(r"(<a\s|\shref=|</a>)", text.lower()))
+
 @bobo.query('/profile/logout')
 def user_logout(bobo_request):
     # default start page
@@ -705,18 +724,30 @@ def new_poll(bobo_request, submit=None, cancel=None, author=None, title=None,
 @bobo.query('/:hash_key/submit')
 def submit_content(bobo_request, hash_key=None, submitter=None, content=None):
     value_dict = get_default_values(bobo_request)
-    data = {"submitter": submitter, "content": content}
+    value_dict["errors"] = {}
+    data = {}
+    if content and check_spam_content(content):
+        value_dict["errors"]["content"] = \
+                "Spam-Verdacht: Inhalt darf keine HTML-Tags enthalten"
+    else:
+        data["content"] = content
+    if submitter and check_spam_submitter_name(submitter):
+        value_dict["errors"]["submitter"] = \
+                "Spam-Verdacht: bitte den Namen korrigieren"
+    else:
+        data["submitter"] = submitter
     poll_id = get_poll_id(hash_key)
     if not poll_id is None:
         poll = Poll.get(poll_id)
         value_dict["poll"] = poll
-        errors = {}
         try:
             data = forms.SubmitForm.to_python(data)
         except formencode.Invalid, errors_packed:
+            # merge errors with previous ones - but never overwrite existing ones
             errors = errors_packed.unpack_errors()
-        if errors or poll.is_closed():
+            errors.update(value_dict["errors"])
             value_dict["errors"] = errors
+        if value_dict["errors"] or poll.is_closed():
             return render("poll_details.html", input_data=data, **value_dict)
         else:
             # create the new submission content