Merge ca85d20204 into 10d3af84b8

The query string send to DDG must not be qouted.

The query string was URL-qouted in #4011, but the URL-qouted query string result
in unexpected *URL decoded* and other garbish results as reported in #4019
and #4020.  To test compare the results of a query like::

    !ddg Häuser und Straßen :de
    !ddg Häuser und Straßen :all
    !ddg 房屋和街道 :all
    !ddg 房屋和街道 :zh

Closed:

- [#4019] https://github.com/searxng/searxng/issues/4019
- [#4020] https://github.com/searxng/searxng/issues/4020

Related:

- [#4011] https://github.com/searxng/searxng/pull/4011

Signed-off-by: Markus Heiser <markus.heiser@darmarit.de>

docs/admin/settings/settings_server.rst

@@ -14,6 +14,7 @@
        limiter: false
        public_instance: false
        image_proxy: false
+       method: "GET"
        default_http_headers:
          X-Content-Type-Options : nosniff
          X-Download-Options : noopen
@@ -50,8 +51,14 @@
 ``image_proxy`` : ``$SEARXNG_IMAGE_PROXY``
   Allow your instance of SearXNG of being able to proxy images.  Uses memory space.
 
+``method`` : ``GET`` | ``POST``
+  HTTP method.  By defaults ``GET`` is used / The ``POST`` method has the
+  advantage with some WEB browsers that the history is not easy to read, but
+  there are also various disadvantages that sometimes severely restrict the ease
+  of use for the user (e.g. back button to jump back to the previous search
+  page and drag & drop of search term to new tabs do not work as expected).
+
 .. _HTTP headers: https://developer.mozilla.org/en-US/docs/Web/HTTP/Headers
 
 ``default_http_headers`` :
   Set additional HTTP headers, see `#755 <https://github.com/searx/searx/issues/715>`__
-

searx/engines/duckduckgo.py

@@ -6,7 +6,7 @@ DuckDuckGo Lite
 
 from typing import TYPE_CHECKING
 import re
-from urllib.parse import urlencode, quote_plus
+from urllib.parse import urlencode
 import json
 import babel
 import lxml.html
@@ -263,7 +263,7 @@ def request(query, params):
 
     params['url'] = url
     params['method'] = 'POST'
-    params['data']['q'] = quote_plus(query)
+    params['data']['q'] = query
 
     # The API is not documented, so we do some reverse engineering and emulate
     # what https://html.duckduckgo.com/html does when you press "next Page" link
@@ -381,7 +381,11 @@ def response(resp):
     zero_click_info_xpath = '//div[@id="zero_click_abstract"]'
     zero_click = extract_text(eval_xpath(doc, zero_click_info_xpath)).strip()
 
-    if zero_click and "Your IP address is" not in zero_click and "Your user agent:" not in zero_click:
+    if zero_click and (
+        "Your IP address is" not in zero_click
+        and "Your user agent:" not in zero_click
+        and "URL Decoded:" not in zero_click
+    ):
         current_query = resp.search_params["data"].get("q")
 
         results.append(

searx/settings.yml

@@ -98,7 +98,7 @@ server:
   http_protocol_version: "1.0"
   # POST queries are more secure as they don't show up in history but may cause
   # problems when using Firefox containers
-  method: "POST"
+  method: "GET"
   default_http_headers:
     X-Content-Type-Options: nosniff
     X-Download-Options: noopen
@@ -145,11 +145,11 @@ ui:
 # Lock arbitrary settings on the preferences page.  To find the ID of the user
 # setting you want to lock, check the ID of the form on the page "preferences".
 #
-# preferences:
-#   lock:
+preferences:
+  lock:
+    - method
 #     - language
 #     - autocomplete
-#     - method
 #     - query_in_title
 
 # searx supports result proxification using an external service: