geminispace.info

commit 9de31b2aa50f5c35b13f33eb5506d560f180844b
parent e24313a14c4ad496f418f47c1cd6162373234ef3
Author: René Wagner <rwa@clttr.info>
Date:   Tue, 16 Aug 2022 19:38:41 +0200

fix normalizing of URIs with default port

It seems that urlsplit.port can not be string-compared for some reason.
So we simply always replace the port portion - i assume this is fine
for almost everything out there.

fixes #55

Diffstat:
M
gus/lib/gemini.py
 | 
11
+++++------
M
tests/gus/lib/test_gemini.py
 | 
5
++---

2 files changed, 7 insertions(+), 9 deletions(-)
diff --git a/gus/lib/gemini.py b/gus/lib/gemini.py
@@ -392,12 +392,11 @@ class GeminiResource:
 
     def _get_normalized_url_and_host(self):
         url_normalized = unquote(self.fetchable_url.rstrip("/"))
-        if self.urlsplit.port == "1965":
-            url_normalized = url_normalized.replace(
-                self.urlsplit.hostname.lower() + ":1965",
-                self.urlsplit.hostname.lower(),
-                1,
-            )
+        url_normalized = url_normalized.replace(
+            self.urlsplit.hostname.lower() + ":1965",
+            self.urlsplit.hostname.lower(),
+            1,
+        )
         host_normalized = self.urlsplit.hostname.lower()
         return url_normalized, host_normalized
 
diff --git a/tests/gus/lib/test_gemini.py b/tests/gus/lib/test_gemini.py
@@ -51,8 +51,9 @@ text
         (["/bar", "gemini://gus.guru/foo/", "gus.guru"], [True, "gemini://gus.guru/bar", "gus.guru"]),
         (["bar", "gemini://gus.guru/foo", "gus.guru"], [True, "gemini://gus.guru/bar", "gus.guru"]),
         (["bar", "gemini://gus.guru/foo/", "gus.guru"], [True, "gemini://gus.guru/foo/bar", "gus.guru"]),
-        # (["bar", "gemini://gus.guru/foo", "gus.guru"], [True, "gemini://gus.guru/foo/bar", "gus.guru"]),
         (["//foo.com", None, None], [True, "gemini://foo.com", "foo.com"]),
+        (["gemini://gem.splatt9990.com/index.gmi", None, None], [True, "gemini://gem.splatt9990.com/index.gmi", "gem.splatt9990.com"]),
+        (["gemini://gem.splatt9990.com:1965/index.gmi", None, None], [True, "gemini://gem.splatt9990.com/index.gmi", "gem.splatt9990.com"]),
 
     ])
     def test_url_parsing(self, test_input, expected_result):
@@ -134,12 +135,10 @@ Allow: /""")
 Disallow: /cgi-bin/wp.cgi/view
 Disallow: /cgi-bin/wp.cgi/media
 Disallow: /cgi-bin/wp.cgi/search
-
 Disallow: /cgi-bin/waffle.cgi/article
 Disallow: /cgi-bin/waffle.cgi/feed
 Disallow: /cgi-bin/waffle.cgi/links
 Disallow: /cgi-bin/waffle.cgi/view
-
 Disallow: /cgi-bin/witw.cgi/play
 """)
         self._assert_fetchable(rp, "/cgi-bin/waffle.cgi/feed/link", False)