[XML-SIG] Problem using xmlrpclib.

[Brendan Rankin]

Just forwarding the fact that I solved my problem.  I'm now using the API
requests that require the login information for both cookies and XMLRPC
session_id/userID fields.

Thanks to both John Lee and M.A. Lemburg for the confirmation I was looking
for.  Thanks especially to John Lee for a good suggestion of methodology.

Regards,

- Brendan

-----Original Message-----
From: Brendan Rankin [mailto:brendan at magma-da.com]
Sent: Thursday, August 07, 2003 3:43 PM
To: John J Lee
Subject: RE: [XML-SIG] Problem using xmlrpclib.


John,

It *is* (or was) the problem.  I'm now doing *exactly* what you said and
everything appears to be functioning.  I can access any further API
Requests.

Actually, it was contained in their documentation, though I still don't
understand the need for it.  To me, it seems quite redundant.  Even stranger
is the fact that the SOAP server does not have this requirement.  One would
think that they would, at the very least, try to be consistent....

Yes, I read the doc headers for the Cookie.py file included with the distro.
and it is for servers....  Thanks for the cool ClientCookie class.  Saves me
a little bit of regexing the headers on my own!

Cheers,

- Brendan

-----Original Message-----
From: John J Lee [mailto:jjl at pobox.com]
Sent: Thursday, August 07, 2003 3:07 PM
To: Brendan Rankin
Subject: RE: [XML-SIG] Problem using xmlrpclib.


On Thu, 7 Aug 2003, Brendan Rankin wrote:
[...]
> Also, their error message lead me to believe that it was an XML or XMLRPC
> problem (malformed request, etc.), since, I felt than any HTML issues
should
> have triggered a HTML Protocol Error....  Something like "Missing Cookies,
> can't authenticate" would have prevented me from debugging most of the
> xmlrpclib. :-)
[...]

Well, I don't guarantee the missing Cookie header in your requests *is*
the problem, but it seems like a reasonable guess.

In general, HTTP servers for web sites are usually misconfigured badly
enough that you can't rely on an error response to even have a non-200
HTTP response code, so from that PoV, nothing surprises me.  One might
expect better of an XML-RPC server, it's true!  But (to somebody who's
never used XML-RPC, anyway), it seems extremely odd in the first place
that those Set-Cookie headers are there, and even stranger that they
appear to be a necessary part of the XML-RPC interaction... how does the
XML-RPC server even get to find out about these cookies??  <shrug>


> P.S.:  Are you aware of the Cookie.py file included in the Python distro.?

That's for servers, not clients.


John