Merge pull request #457 from gratipay/base_url

add base URL canonicalization to algorithm

aspen/algorithms/website.py

@@ -61,6 +61,10 @@ def raise_200_for_OPTIONS(request):
         raise Response(200)
 
 
+def redirect_to_base_url(website, request):
+    website.canonicalize_base_url(request)
+
+
 def dispatch_request_to_filesystem(website, request):
 
     if website.list_directories:

aspen/configuration/__init__.py

@@ -20,7 +20,7 @@
 import aspen.logging
 from aspen.configuration import parse
 from aspen.exceptions import ConfigurationError
-from aspen.utils import ascii_dammit
+from aspen.utils import ascii_dammit, BaseURLCanonicalizer
 from aspen.typecasting import defaults as default_typecasters
 import aspen.body_parsers
 
@@ -30,7 +30,8 @@
 
     # 'name':               (default,               from_unicode)
 KNOBS = \
-    { 'changes_reload':     (False,                 parse.yes_no)
+    { 'base_url':           (None,                  parse.identity)
+    , 'changes_reload':     (False,                 parse.yes_no)
     , 'charset_dynamic':    ('UTF-8',               parse.charset)
     , 'charset_static':     (None,                  parse.charset)
     , 'indices':            (default_indices,       parse.list_)
@@ -220,6 +221,9 @@ def safe_getcwd(errorstr):
             self.media_type_json: aspen.body_parsers.jsondata
             }
 
+        # install a base URL canonicalizer
+        self.canonicalize_base_url = BaseURLCanonicalizer(self.base_url)
+
         # load renderers
         self.renderer_factories = {}
         for name in aspen.BUILTIN_RENDERERS:

aspen/utils.py

@@ -386,7 +386,7 @@ def nice(t):
             raise TypeError(msg)
 
 
-# Hostname canonicalization
+# Base URL Canonicalization
 # =========================
 
 def _extract_scheme(request):
@@ -395,16 +395,16 @@ def _extract_scheme(request):
 def _extract_host(request):
     return request.headers['Host']  # will 400 if missing
 
-def Canonizer(expected, permanent_redirect=False, extract_scheme=_extract_scheme,
+def BaseURLCanonicalizer(expected, permanent_redirect=False, extract_scheme=_extract_scheme,
         extract_host=_extract_host):
-    """Takes a netloc such as http://localhost:8080 (no path part).
+    """Takes a base_url such as http://localhost:8080 (no path part).
     """
 
     def noop(request):
         pass
 
-    def canonize(request):
-        """Enforce a certain network location.
+    def func(request):
+        """Enforce a certain base URL.
         """
 
         scheme = extract_scheme(request)
@@ -424,7 +424,7 @@ def canonize(request):
                 uri += '/'
             request.redirect(uri, permanent=permanent_redirect)
 
-    return expected and canonize or noop
+    return expected and func or noop
 
 
 if __name__ == '__main__':

doc/api/website/index.html.spt

@@ -13,6 +13,7 @@ arguments to Website:</p>
 
 <table>
     <tr><td><b><u>attribute</u></b></td><td><b><u>default</u></b></td> </tr>
+    <tr><td>base_url</td><td>''</td> </tr>
     <tr><td>changes_reload</td><td>False</td> </tr>
     <tr><td>charset_dynamic</td><td>UTF-8</td> </tr>
     <tr><td>charset_static</td><td>None</td> </tr>

doc/configuration/index.html.spt

@@ -10,7 +10,8 @@ like <code>ASPEN_NAME</code> and a keyword argument to `aspen.website.Website`.
 Here is an example configuration:</p>
 
 <table>
-    <tr><td><b><u>name</u></b></td>         <td><b><u>default</u></b></td> </tr>
+    <tr><td><b><u>name</u></b></td>         <td><b><u>example value</u></b></td> </tr>
+    <tr><td>base_url</td>                   <td>http://www.example.com</td> </tr>
     <tr><td>changes_reload</td>             <td>yes</td> </tr>
     <tr><td>charset_dynamic</td>            <td>ISO-8859-1</td> </tr>
     <tr><td>charset_static</td>             <td>windows-1252</td> </tr>

tests/test_utils.py

@@ -56,3 +56,48 @@ def test_to_rfc822():
     expected = 'Thu, 01 Jan 1970 00:00:00 GMT'
     actual = to_rfc822(datetime(1970, 1, 1))
     assert actual == expected
+
+
+def test_base_url_canonicalizer_canonicalizes_base_url(harness):
+    harness.fs.www.mk(('index.html', 'Greetings, program!'))
+    harness.client.hydrate_website(base_url='http://example.com')
+    response = harness.client.GxT()
+    assert response.code == 302
+    assert response.headers['Location'] == 'http://example.com/'
+
+def test_base_url_canonicalizer_includes_path_and_qs_for_GET(harness):
+    harness.fs.www.mk(('index.html', 'Greetings, program!'))
+    harness.client.hydrate_website(base_url='http://example.com')
+    response = harness.client.GxT('/foo/bar?baz=buz')
+    assert response.code == 302
+    assert response.headers['Location'] == 'http://example.com/foo/bar?baz=buz'
+
+def test_base_url_canonicalizer_redirects_to_homepage_for_POST(harness):
+    harness.fs.www.mk(('index.html', 'Greetings, program!'))
+    harness.client.hydrate_website(base_url='http://example.com')
+    response = harness.client.PxST('/foo/bar?baz=buz')
+    assert response.code == 302
+    assert response.headers['Location'] == 'http://example.com/'
+
+def test_base_url_canonicalizer_allows_good_base_url(harness):
+    harness.fs.www.mk(('index.html', 'Greetings, program!'))
+    harness.client.hydrate_website(base_url='http://localhost')
+    response = harness.client.GET()
+    assert response.code == 200
+    assert response.body == 'Greetings, program!'
+
+def test_base_url_canonicalizer_is_noop_without_base_url(harness):
+    harness.fs.www.mk(('index.html', 'Greetings, program!'))
+    harness.client.hydrate_website()
+    response = harness.client.GET()
+    assert response.code == 200
+    assert response.body == 'Greetings, program!'
+
+def test_base_url_canonicalizer_is_really_noop_without_base_url(harness):
+    harness.client.hydrate_website()
+    harness.client.website.algorithm['redirect_to_base_url'](harness.client.website, None)
+
+def test_base_url_canonicalizer_is_not_noop_with_base_url(harness):
+    harness.client.hydrate_website(base_url='foo')
+    with raises(AttributeError):
+        harness.client.website.algorithm['redirect_to_base_url'](harness.client.website, None)