Issue 26609: Wrong request target in test_httpservers.py
Issue26609
Created on 2016-03-22 08:56 by xiang.zhang, last changed 2022-04-11 14:58 by admin. This issue is now closed.
| Files | ||||
|---|---|---|---|---|
| File name | Uploaded | Description | Edit | |
| request_target_in_test_httpservers.patch | xiang.zhang, 2016-03-22 08:56 | review | ||
| request_target_in_test_httpservers_v2.patch | xiang.zhang, 2016-04-05 02:28 | review | ||
| Messages (6) | |||
|---|---|---|---|
| msg262171 - (view) | Author: Xiang Zhang (xiang.zhang) *  |
Date: 2016-03-22 08:56 | |
When requesting a resource from an origin server, the request-target in request line should always starts with a back slash. But in SimpleHTTPServerTestCase in test_httpservers.py, almost all the requests are sent without the back slash though the handler handles it well. The request lines are like 'GET tmpXXXXX HTTP/1.1'. I add the back slashes. Maybe in SimpleHTTPRequestHandler, we should reject such invalid request-targets and then return BAD_REQUEST. And then issue2254 won't happen. |
|||
| msg262803 - (view) | Author: Martin Panter (martin.panter) * |
Date: 2016-04-02 23:03 | |
Perhaps tempdir_name could get the slash prefixed in setUp() rather than in every test. There is one case that creates index.html which could probably be changed to use self.tempdir, and then the rest could be renamed to say self.base_url. Rejecting requests that do not start with a slash could be a compatibility problem. Do you think potential security benefits (i.e. having unexpected variations on the format of the URL) outweigh that? Either way, it would be good to retain a test without a slash. |
|||
| msg262886 - (view) | Author: Xiang Zhang (xiang.zhang) * |
Date: 2016-04-05 02:28 | |
Get the slash prefixed path in Setup() is a good idea. I change the patch. I retain self.tempdir_name so we can use it in a test for no leading slash. The case creating index.html is OK with self.tempdir_name since we have changed our working directory to basetempdir. I didn't think about compatibility but I know it's important. So rejecting the invalid request-targets is not a good idea to me now. |
|||
| msg263096 - (view) | Author: Roundup Robot (python-dev)  |
Date: 2016-04-09 13:44 | |
New changeset 0e19f421dc9e by Martin Panter in branch '3.5': Issue #26609: Fix HTTP server tests to request an absolute URL path https://hg.python.org/cpython/rev/0e19f421dc9e New changeset 34ebf79acd78 by Martin Panter in branch 'default': Issue #26609: Merge HTTP tests from 3.5 https://hg.python.org/cpython/rev/34ebf79acd78 New changeset 2691f81a89a7 by Martin Panter in branch '2.7': Issue #26609: Fix HTTP server tests to request an absolute URL path https://hg.python.org/cpython/rev/2691f81a89a7 |
|||
| msg263097 - (view) | Author: Roundup Robot (python-dev) |
Date: 2016-04-09 13:46 | |
New changeset 4f64b1c87a56 by Martin Panter in branch '2.7': Issue #26609: Fix up Python 2 port https://hg.python.org/cpython/rev/4f64b1c87a56 |
|||
| msg263115 - (view) | Author: Martin Panter (martin.panter) * |
Date: 2016-04-09 22:31 | |
Thanks for the patch |
|||
| History | |||
|---|---|---|---|
| Date | User | Action | Args |
| 2022-04-11 14:58:28 | admin | set | github: 70796 |
| 2016-04-09 22:31:02 | martin.panter | set | status: open -> closed versions: + Python 2.7, Python 3.5 messages: + msg263115 resolution: fixed |
| 2016-04-09 13:46:08 | python-dev | set | messages: + msg263097 |
| 2016-04-09 13:44:39 | python-dev | set | nosy:
+ python-dev messages: + msg263096 |
| 2016-04-05 02:28:37 | xiang.zhang | set | files:
+ request_target_in_test_httpservers_v2.patch messages: + msg262886 |
| 2016-04-03 04:40:49 | martin.panter | link | issue26585 dependencies |
| 2016-04-02 23:03:16 | martin.panter | set | messages:
+ msg262803 stage: patch review |
| 2016-03-22 08:56:20 | xiang.zhang | create | |