Details
-
Bug
-
Status: Open
-
Major
-
Resolution: Unresolved
-
None
-
None
Description
This issue was reported by 'durin42' on #svn-dev yesterday.
What steps will reproduce the problem?
Using svn from trunk:
$ svn co http://code.djangoproject.com/svn/django/trunk /tmp/wc
Output: svn will return a 400 Bad Request error
What is the expected output? What do you see instead?
The 400 Bad Request error is returned on the first PROPFIND request to the
remote server.
When looking at the exact response with wireshark, we get this error message:
"perlbal: Can't pipeline to HTTP/1.0"
The code fragment in perlbal that reports this error refers to the missing
Content-Length field in the request header:
http://code.sixapart.com/trac/perlbal/changeset/577
If my reading of the comment in that fragment is correct, perlbal requires
the Content-Length field when the client tries to pipeline multiple
requests in order to avoid flooding attacks.