From f14177381433618a5b4f5bcff83e4e1a19487f02 Mon Sep 17 00:00:00 2001
From: Adam Tauber <asciimoo@gmail.com>
Date: Fri, 19 Dec 2014 22:40:37 +0100
Subject: [PATCH] [mod][fix] https rewrite refactor ++ fixes

---
 searx/https_rewrite.py           | 68 ++++++++++++++++++++++++++++++--
 searx/https_rules/Soundcloud.xml |  2 +-
 searx/webapp.py                  | 59 +--------------------------
 3 files changed, 68 insertions(+), 61 deletions(-)

diff --git a/searx/https_rewrite.py b/searx/https_rewrite.py
index 9faf3599d..408474a44 100644
--- a/searx/https_rewrite.py
+++ b/searx/https_rewrite.py
@@ -16,6 +16,7 @@ along with searx. If not, see < http://www.gnu.org/licenses/ >.
 '''
 
 import re
+from urlparse import urlparse
 from lxml import etree
 from os import listdir
 from os.path import isfile, isdir, join
@@ -86,15 +87,23 @@ def load_single_https_ruleset(filepath):
 
             # TODO hack, which convert a javascript regex group
             # into a valid python regex group
-            rule_from = ruleset.attrib.get('from').replace('$', '\\')
-            rule_to = ruleset.attrib.get('to').replace('$', '\\')
+            rule_from = ruleset.attrib['from'].replace('$', '\\')
+            if rule_from.endswith('\\'):
+                rule_from = rule_from[:-1]+'$'
+            rule_to = ruleset.attrib['to'].replace('$', '\\')
+            if rule_to.endswith('\\'):
+                rule_to = rule_to[:-1]+'$'
 
             # TODO, not working yet because of the hack above,
             # currently doing that in webapp.py
             # rule_from_rgx = re.compile(rule_from, re.I)
 
             # append rule
-            rules.append((rule_from, rule_to))
+            try:
+                rules.append((re.compile(rule_from, re.I | re.U), rule_to))
+            except:
+                # TODO log regex error
+                continue
 
         # this child define an exclusion
         elif ruleset.tag == 'exclusion':
@@ -143,3 +152,56 @@ def load_https_rules(rules_path):
         https_rules.append(ruleset)
 
     print(' * {n} https-rules loaded'.format(n=len(https_rules)))
+
+
+
+def https_url_rewrite(result):
+    skip_https_rewrite = False
+    # check if HTTPS rewrite is possible
+    for target, rules, exclusions in https_rules:
+
+        # check if target regex match with url
+        if target.match(result['parsed_url'].netloc):
+            # process exclusions
+            for exclusion in exclusions:
+                # check if exclusion match with url
+                if exclusion.match(result['url']):
+                    skip_https_rewrite = True
+                    break
+
+            # skip https rewrite if required
+            if skip_https_rewrite:
+                break
+
+            # process rules
+            for rule in rules:
+                try:
+                    new_result_url = rule[0].sub(rule[1], result['url'])
+                except:
+                    break
+
+                # parse new url
+                new_parsed_url = urlparse(new_result_url)
+
+                # continiue if nothing was rewritten
+                if result['url'] == new_result_url:
+                    continue
+
+                # get domainname from result
+                # TODO, does only work correct with TLD's like
+                #  asdf.com, not for asdf.com.de
+                # TODO, using publicsuffix instead of this rewrite rule
+                old_result_domainname = '.'.join(
+                    result['parsed_url'].hostname.split('.')[-2:])
+                new_result_domainname = '.'.join(
+                    new_parsed_url.hostname.split('.')[-2:])
+
+                # check if rewritten hostname is the same,
+                # to protect against wrong or malicious rewrite rules
+                if old_result_domainname == new_result_domainname:
+                    # set new url
+                    result['url'] = new_result_url
+
+            # target has matched, do not search over the other rules
+            break
+    return result
diff --git a/searx/https_rules/Soundcloud.xml b/searx/https_rules/Soundcloud.xml
index 0baa5832b..6958e8cbc 100644
--- a/searx/https_rules/Soundcloud.xml
+++ b/searx/https_rules/Soundcloud.xml
@@ -89,7 +89,7 @@
 	<rule from="^http://([aiw]\d|api|wis)\.sndcdn\.com/"
 		to="https://$1.sndcdn.com/" />
 
-	<rule from="^http://((?:api|backstage|blog|connect|developers|ec-media|eventlogger|help-assets|media|visuals|w|www)\.)?soundcloud\.com/"
+	<rule from="^http://((?:api|backstage|blog|connect|developers|ec-media|eventlogger|help-assets|media|visuals|w|www)\.|)soundcloud\.com/"
 		to="https://$1soundcloud.com/" />
 
 	<rule from="^https?://scbackstage\.wpengine\.netdna-cdn\.com/"
diff --git a/searx/webapp.py b/searx/webapp.py
index a2a135e9a..915fb3564 100644
--- a/searx/webapp.py
+++ b/searx/webapp.py
@@ -41,15 +41,12 @@ from searx.utils import (
     UnicodeWriter, highlight_content, html_to_text, get_themes
 )
 from searx.version import VERSION_STRING
-from searx.https_rewrite import https_rules
 from searx.languages import language_codes
+from searx.https_rewrite import https_url_rewrite
 from searx.search import Search
 from searx.query import Query
 from searx.autocomplete import backends as autocomplete_backends
 
-from urlparse import urlparse
-import re
-
 
 static_path, templates_path, themes =\
     get_themes(settings['themes_path']
@@ -215,59 +212,7 @@ def index():
         if settings['server']['https_rewrite']\
            and result['parsed_url'].scheme == 'http':
 
-            skip_https_rewrite = False
-
-            # check if HTTPS rewrite is possible
-            for target, rules, exclusions in https_rules:
-
-                # check if target regex match with url
-                if target.match(result['url']):
-                    # process exclusions
-                    for exclusion in exclusions:
-                        # check if exclusion match with url
-                        if exclusion.match(result['url']):
-                            skip_https_rewrite = True
-                            break
-
-                    # skip https rewrite if required
-                    if skip_https_rewrite:
-                        break
-
-                    # process rules
-                    for rule in rules:
-                        try:
-                            # TODO, precompile rule
-                            p = re.compile(rule[0])
-
-                            # rewrite url if possible
-                            new_result_url = p.sub(rule[1], result['url'])
-                        except:
-                            break
-
-                        # parse new url
-                        new_parsed_url = urlparse(new_result_url)
-
-                        # continiue if nothing was rewritten
-                        if result['url'] == new_result_url:
-                            continue
-
-                        # get domainname from result
-                        # TODO, does only work correct with TLD's like
-                        #  asdf.com, not for asdf.com.de
-                        # TODO, using publicsuffix instead of this rewrite rule
-                        old_result_domainname = '.'.join(
-                            result['parsed_url'].hostname.split('.')[-2:])
-                        new_result_domainname = '.'.join(
-                            new_parsed_url.hostname.split('.')[-2:])
-
-                        # check if rewritten hostname is the same,
-                        # to protect against wrong or malicious rewrite rules
-                        if old_result_domainname == new_result_domainname:
-                            # set new url
-                            result['url'] = new_result_url
-
-                    # target has matched, do not search over the other rules
-                    break
+            result = https_url_rewrite(result)
 
         if search.request_data.get('format', 'html') == 'html':
             if 'content' in result: