i found Python 3.1 xmlrpc lib use "<param>" not properly. and i have fixed it.
please have a look at the attachment

wrong.txt is the output of client_old.py, it use <param> not properly.
right.txt is the output of client.py, it use <param> properly.

Please could you generate and upload a patch against the Python source?  (For Windows, you can do this using the WinMerge tool, amongst others.)  I'm unable to open the file you attached on my machine:

"""No application knows how to open ... xmlrpc.client.modify.rar."""

If you could explain how "<param>" is being used improperly, that might also help.

I can't modify the issue. so i replay this email.
The old xmlrpc lib don't use <param> surround a param. please take a look at
wrong.xml and right.xml.

2010/2/22 Mark Dickinson <report@bugs.python.org>

>
> Mark Dickinson <dickinsm@gmail.com> added the comment:
>
> Please could you generate and upload a patch against the Python source?
>  (For Windows, you can do this using the WinMerge tool, amongst others.)
>  I'm unable to open the file you attached on my machine:
>
> """No application knows how to open ... xmlrpc.client.modify.rar."""
>
> If you could explain how "<param>" is being used improperly, that might
> also help.
>
> ----------
> nosy: +mark.dickinson
>
> _______________________________________
> Python tracker <report@bugs.python.org>
> <http://bugs.python.org/issue7977>
> _______________________________________
>

The patch causes failures in test_xmlrpc, using Py3k branch.

Can i have a look at the failure tests？
I just check out code at http://svn.python.org/projects/python/branches/py3k
It's very slow. and i'm afraid that i don't know how to run the tests.

The <param> tag use improperly is very clear.

the wrong case: very ugly, there is no difference between <param> and
<params>.
<params>
<param>
<value><data>
<value><int>1</int></value>
<value><string>sinojelly</string></value>
<value><string>123456</string></value>
<value><struct>
<member>
<name>description</name>
<value><string>Test description4</string></value>
</member>
<member>
<name>title</name>
<value><string>Test article3</string></value>
</member>
</struct></value>
<value><boolean>1</boolean></value>
</data></value>
</param>
</params>

the right one: there is a <param> around every parameter.
 <params>
  <param>
   <value>
    <string>479153</string>
   </value>
  </param>
  <param>
   <value>
    <string>sinojelly</string>
   </value>
  </param>
  <param>
   <value>
    <string>123456</string>
   </value>
  </param>
  <param>
   <value>
    <struct>
     <member>
      <name>title</name>
      <value>
       <string>Just Test.</string>
      </value>
     </member>
     <member>
      <name>description</name>
      <value>
       <string>&lt;p&gt;Just Test wlw post.&lt;/p&gt;</string>
      </value>
     </member>
     <member>
      <name>categories</name>
      <value>
       <array>
        <data />
       </array>
      </value>
     </member>
    </struct>
   </value>
  </param>
  <param>
   <value>
    <boolean>1</boolean>
   </value>
  </param>
 </params>

2010/2/22 Florent Xicluna <report@bugs.python.org>

>
> Florent Xicluna <laxyf@yahoo.fr> added the comment:
>
> The patch causes failures in test_xmlrpc, using Py3k branch.
>
> ----------
> nosy: +flox
> priority:  -> normal
> stage:  -> test needed
> versions: +Python 3.2
>
> _______________________________________
> Python tracker <report@bugs.python.org>
> <http://bugs.python.org/issue7977>
> _______________________________________
>

I didn't look at this issue in details.
I'm not a specialist of xmlrpc. But it is unlikely that the patch will be accepted, if it fails the regression tests.

Could you investigate this failure and provide a patch which fixes the tests? You could add a specific test related to this issue, too.

Command to run the tests (on linux):
        python -m test.regrtest test_xmlrpc
 
I've tested with 3.1 branch, and test_xmlrpc fails too.

Attached, the output of the test with "-v test_xmlrpc".

I know why those two tests run failed, and I found another way to solve the
problem, do not make the existing tests broken, and at the same time I have
added another test case.
I encountered this problem due to a argument is already being a tuple but
re-packaged into another tuple, It appears a high probability. Maybe modify
the client.py according to this patch is a solution.

2010/2/22 Florent Xicluna <report@bugs.python.org>

>
> Florent Xicluna <laxyf@yahoo.fr> added the comment:
>
> Attached, the output of the test with "-v test_xmlrpc".
>
> ----------
> Added file: http://bugs.python.org/file16308/issue7977_test_xmlrpc.log
>
> _______________________________________
> Python tracker <report@bugs.python.org>
> <http://bugs.python.org/issue7977>
> _______________________________________
>

But what if I want to pass a tuple through xmlrpc?
The function in the test case:
+        def execute(server, methodname, *args):
+            r = getattr(server, methodname)(args) # params become a tuple
+            return r
looks wrong to me. It has the effect to modify the nature of the arguments. 
If I write instead:
+        def execute(server, methodname, *args):
+            r = getattr(server, methodname)(*args)
+            return r
(note the additional '*') the function may receive several arguments, and will call the server method with several arguments as well.
Tuple unpacking occurs where tuple packing was made. The library does not need to be changed.