From 5cc40ee3ba51b6b9592642b2e1dd66a0505ecc46 Mon Sep 17 00:00:00 2001
From: Oliver Urs Lenz <oliver.urs.lenz@gmail.com>
Date: Wed, 8 Aug 2018 16:37:39 +0200
Subject: [PATCH] correct suffix order in ComplexHTTPRequestHandler

---
 pelican/server.py            | 49 +++++++++++++-----------------
 pelican/tests/test_server.py | 58 ++++++++++++++++++++++++++++++++++++
 2 files changed, 79 insertions(+), 28 deletions(-)
 create mode 100644 pelican/tests/test_server.py

diff --git a/pelican/server.py b/pelican/server.py
index 09eeb493..0ba7b996 100644
--- a/pelican/server.py
+++ b/pelican/server.py
@@ -42,8 +42,7 @@ def parse_arguments():
 
 
 class ComplexHTTPRequestHandler(srvmod.SimpleHTTPRequestHandler):
-    SUFFIXES = ['', '.html', '/index.html']
-    RSTRIP_PATTERNS = ['', '/']
+    SUFFIXES = ['.html', '/index.html', '/', '']
 
     def translate_path(self, path):
         # abandon query parameters
@@ -67,34 +66,28 @@ class ComplexHTTPRequestHandler(srvmod.SimpleHTTPRequestHandler):
 
     def do_GET(self):
         # cut off a query string
-        if '?' in self.path:
-            self.path, _ = self.path.split('?', 1)
+        original_path = self.path.split('?', 1)[0]
+        # try to find file
+        self.path = self.get_path_that_exists(original_path)
 
-        found = False
-        # Try to detect file by applying various suffixes and stripping
-        # patterns.
-        for rstrip_pattern in self.RSTRIP_PATTERNS:
-            if found:
-                break
-            for suffix in self.SUFFIXES:
-                if not hasattr(self, 'original_path'):
-                    self.original_path = self.path
-
-                self.path = self.original_path.rstrip(rstrip_pattern) + suffix
-                path = self.translate_path(self.path)
-
-                if os.path.exists(path):
-                    srvmod.SimpleHTTPRequestHandler.do_GET(self)
-                    logging.info("Found `%s`.", self.path)
-                    found = True
-                    break
-
-                logging.info("Tried to find `%s`, but it doesn't exist.", path)
-
-        if not found:
-            # Fallback if there were no matches
+        if not self.path:
             logging.warning("Unable to find `%s` or variations.",
-                            self.original_path)
+                            original_path)
+            return
+
+        logging.info("Found `%s`.", self.path)
+        srvmod.SimpleHTTPRequestHandler.do_GET(self)
+
+    def get_path_that_exists(self, original_path):
+        # Try to strip trailing slash
+        original_path = original_path.rstrip('/')
+        # Try to detect file by applying various suffixes
+        for suffix in self.SUFFIXES:
+            path = original_path + suffix
+            if os.path.exists(self.translate_path(path)):
+                return path
+            logging.info("Tried to find `%s`, but it doesn't exist.", path)
+        return None
 
     def guess_type(self, path):
         """Guess at the mime type for the specified file.
diff --git a/pelican/tests/test_server.py b/pelican/tests/test_server.py
new file mode 100644
index 00000000..508b75ff
--- /dev/null
+++ b/pelican/tests/test_server.py
@@ -0,0 +1,58 @@
+import os
+from shutil import rmtree
+from tempfile import mkdtemp
+
+from six import BytesIO
+
+from pelican.server import ComplexHTTPRequestHandler
+from pelican.tests.support import unittest
+
+
+class MockRequest(object):
+    def makefile(self, *args, **kwargs):
+        return BytesIO(b"")
+
+
+class MockServer(object):
+    pass
+
+
+class TestServer(unittest.TestCase):
+
+    def setUp(self):
+        self.server = MockServer()
+        self.temp_output = mkdtemp(prefix='pelicantests.')
+        self.old_cwd = os.getcwd()
+        os.chdir(self.temp_output)
+
+    def tearDown(self):
+        rmtree(self.temp_output)
+        os.chdir(self.old_cwd)
+
+    def test_get_path_that_exists(self):
+
+        handler = ComplexHTTPRequestHandler(MockRequest(), ('0.0.0.0', 8888),
+                                            self.server)
+        handler.base_path = self.temp_output
+
+        os.mknod(os.path.join(self.temp_output, 'foo.html'))
+        os.mkdir(os.path.join(self.temp_output, 'foo'))
+        os.mknod(os.path.join(self.temp_output, 'foo', 'index.html'))
+
+        os.mkdir(os.path.join(self.temp_output, 'bar'))
+        os.mknod(os.path.join(self.temp_output, 'bar', 'index.html'))
+
+        os.mkdir(os.path.join(self.temp_output, 'baz'))
+
+        for suffix in ['', '/']:
+            path = handler.get_path_that_exists('foo' + suffix)
+            self.assertEqual(path, 'foo.html')
+
+            path = handler.get_path_that_exists('bar' + suffix)
+            self.assertEqual(path, 'bar/index.html')
+
+            path = handler.get_path_that_exists('baz' + suffix)
+            self.assertEqual(path, 'baz/')
+
+            path = handler.get_path_that_exists('quux' + suffix)
+            self.assertIsNone(path)