Interoperability testing with Xythos DAV server

I report on preliminary interoperation testing between the Python DAV
client and the Xythos server
(at 216.102.212.251, port 80, using base collection /~webdav/)

So far as I know, Xythos is the only other WebDAV server that supports
locking on collections (MS supports locking, but not on collections.  I am
unsure about the IBM implementation, someone should set me straight.)

I tested OPTIONS, PUT, GET, HEAD, PROPFIND, PROPPATCH, MKCOL, DELETE, LOCK,
and UNLOCK, including
some combinations.

I find the following problems:

1. The reply from server  (to a PROPFIND) has neither a content-length nor
does it use chunked encoding.  I believe this is a violation of HTTP/1.1 rules.

PROPFIND /~webdav HTTP/1.1
Host: 216.102.212.251
 
HTTP/1.1 207 Multi-Status
Server: JavaWebServer/2.0
Content-Type: text/xml; charset=UTF-8
Vary: Accept-Encoding
Connection: close
Date: Mon, 12 Jul 1999 16:49:20 GMT

(seems like the server just closes the connection each time.  Is this
allowed?  It's certainly not efficient.  But anyway on the assumption that
it is legal, I modified my test code to work with this, so I could test
more features)

2) the server does not seem to be preserving the xml:lang attribute

3) dead properties are preserved across deletion of a resource.  That is, I
PROPPATCH foo, delete it, then PUT foo again, and the property I set the
first time is still there

4) PROPPATCH changes a property even if the resource is locked.  I could
set and delete properties

on the other hand, I could not *delete* a property when I did pass a token

5) PUT could not create resource in a locked collection even with token

6) attempt to create collection /~webdav/newcoll/ when parent was locked
got code 405 (NotAllowed).  shouldn't this be 409 (conflict) instead?

7) Deleted resource /~webdav/foo.html when parent was locked (without
passing token)

8) attempt to lock a non-existant resource gets NotFound

I was not able to test operation of locked resources within locked
collections because of these bugs.

When they are corrected, I'll try again.

I congratulate the Xythos team on a good (though not yet perfect) server.

Jim

Received on Monday, 12 July 1999 14:04:20 UTC