Merge branch 'fix/env_based_dereferer_config' into 'develop'

allow env-based configuration of dereferer

See merge request hub/hub!578

src/hub/settings/base.py

@@ -14,6 +14,7 @@ from email.utils import getaddresses
 import json
 import os
 from pathlib import Path
+import re
 
 from django.utils.translation import gettext_lazy as _
 import environ
@@ -85,6 +86,9 @@ env = environ.FileAwareEnv(
 
     PLAINUI_BASE_URL=(str, ''),
     PLAINUI_CONFERENCE=(str, None),
+    PLAINUI_DEREFERER_URL=(str, '//dereferer/{quoted_target}'),
+    PLAINUI_DEREFERER_ALLOWLIST=(list, []),
+    PLAINUI_DEREFERER_COUNTLIST=(list, []),
 
     SHIBBOLEET_WA_ROOM_ID=('str', None),
 )
@@ -522,17 +526,35 @@ PLAINUI_CONFERENCE = env('PLAINUI_CONFERENCE')
 
 # Absolute url to use when dereferring links. Replaces `{quoted_target}` with the urlquoted target url
 # If you modify this on a non-empty database you should run `manage.py rerender_markdown` to update cashed markdown that contains old urls.
-PLAINUI_DEREFERER_URL = 'https://rc3.world/2021/dereferrer/{quoted_target}'
+PLAINUI_DEREFERER_URL = env('PLAINUI_DEREFERER_URL')
+
+# helper function for parsing DEREFERER_*LIST (allow hostname only, regex and https:// prefixed strings)
+re_host = re.compile('([\w\d-]+\.)+\w+(:\d+)?$')
+def _handle_hostpattern_list(target: list, items: list):
+    for item in items:
+        if item.startswith('/') and item.endswith('/'):
+            try:
+                target.append(re.compile('https://' + item[1:-1]))
+            except re.error as err:
+                print('WARNING: failed to parse regex pattern "', item, "', ignoring it: ", err, sep='')
+        elif item.startswith('https://') or item.startswith('http://'):
+            target.append(item)
+        elif re_host.match(item):
+            target.append('https://' + item)
+        else:
+            print('WARNING: failed to parse pattern "', item, "' (expected hostname, url or slash-enclosed regex), skipping it", sep='')
 
 # URI patterns that are auto-allowed (never get dereferrer'd)
 # Only looks at schema and host/port, so for example for the url https://rc3.world:4321/some/where/ the entry should be `'https://rc3.world'`.
 # Can also be a regex, for example re.compile(r'https://.*\.rc3\.world')
 # DON'T ADD TRAILING SLASHES
 DEREFERRER_GLOBAL_ALLOWLIST = []
+_handle_hostpattern_list(DEREFERRER_GLOBAL_ALLOWLIST, env('PLAINUI_DEREFERER_ALLOWLIST'))
 
 # URI patterns where following outgoing links is counted in the database
 # rules are the same as with DEREFERRER_GLOBAL_ALLOWLIST
 DEREFERRER_COUNT_ACCESS = []
+_handle_hostpattern_list(DEREFERRER_COUNT_ACCESS, env('PLAINUI_DEREFERER_COUNTLIST'))
 
 # Shibboleet Mode: an entrypoint /shibboleet which adds the registered user to the conference and redirects to a new WA session
 # ID of the WorkAdventure room which shall be used as the entrypoint in the new WA session