Getter and setter for path component on Requests.

2 files changed, 40 insertions, 2 deletions

diff --git a/libmproxy/flow.py b/libmproxy/flow.py
index 6076d202..868419e5 100644
--- a/libmproxy/flow.py
+++ b/libmproxy/flow.py
@@ -18,7 +18,7 @@
     with their responses, and provide filtering and interception facilities.
 """
 import hashlib, Cookie, cookielib, copy, re, urlparse, os
-import time
+import time, urllib
 import tnetstring, filt, script, utils, encoding, proxy
 from email.utils import parsedate_tz, formatdate, mktime_tz
 from netlib import odict, http, certutils
@@ -397,6 +397,26 @@ class Request(HTTPMsg):
         self.headers["Content-Type"] = [HDR_FORM_URLENCODED]
         self.content = utils.urlencode(odict.lst)
 
+    def get_path_components(self):
+        """
+            Returns the path components of the URL as a list of strings.
+
+            Components are unquoted.
+        """
+        _, _, path, _, _, _ = urlparse.urlparse(self.get_url())
+        return [urllib.unquote(i) for i in path.split("/") if i]
+
+    def set_path_components(self, lst):
+        """
+            Takes a list of strings, and sets the path component of the URL.
+
+            Components are quoted.
+        """
+        lst = [urllib.quote(i, safe="") for i in lst]
+        path = "/" + "/".join(lst)
+        scheme, netloc, _, params, query, fragment = urlparse.urlparse(self.get_url())
+        self.set_url(urlparse.urlunparse([scheme, netloc, path, params, query, fragment]))
+
     def get_query(self):
         """
             Gets the request query string. Returns an ODict object.
diff --git a/test/test_flow.py b/test/test_flow.py
index 47a09360..53d92f25 100644
--- a/test/test_flow.py
+++ b/test/test_flow.py
@@ -741,7 +741,25 @@ class TestRequest:
         r.content = flow.CONTENT_MISSING
         assert not r._assemble()
 
-
+    def test_path_components(self):
+        h = flow.ODictCaseless()
+        c = flow.ClientConnect(("addr", 2222))
+        r = flow.Request(c, (1, 1), "host", 22, "https", "GET", "/", h, "content")
+        assert r.get_path_components() == []
+        r = flow.Request(c, (1, 1), "host", 22, "https", "GET", "/foo/bar", h, "content")
+        assert r.get_path_components() == ["foo", "bar"]
+        q = flow.ODict()
+        q["test"] = ["123"]
+        r.set_query(q)
+        assert r.get_path_components() == ["foo", "bar"]
+
+        r.set_path_components([])
+        assert r.get_path_components() == []
+        r.set_path_components(["foo"])
+        assert r.get_path_components() == ["foo"]
+        r.set_path_components(["/oo"])
+        assert r.get_path_components() == ["/oo"]
+        assert "%2F" in r.path
 
     def test_getset_form_urlencoded(self):
         h = flow.ODictCaseless()