[feature request] Cookie support

Keith Beckman kbeckman at mcg.edu
Tue Sep 15 11:49:39 EDT 2009


I noticed that cadaver very purposely doesn't include cookie support
(http://www.webdav.org/cadaver/history.html, release 0.14.1). However,
several commercial servers, including Novell's oneNet NetStorage,
require session cookies for use (NetStorage maintains a persistent 401
if proper Basic credentials are offered without the session cookie).

While philosophical issues about the WebDAV RFC abound -- and this may
indeed be considered a server bug -- cookies are part of HTTP and WebDAV
is considered an extension to HTTP; so there's a philosophical point to
be made that cookies _are_ part of WebDAV simply by inclusion of all
elements of a subset by a superset. cadaver's current behaviour,
whatever the philosophical issues around cookie support, nonetheless
fails to live up to the Robustness Principle ("/Be conservative in what
you do, be liberal in what you accept from others/.").

cadaver is unusable with several common commercially-available servers,
and its utility is not simply limited by that, but is in environments
which use such servers (my university, for example) exactly nil. cadaver
does not support possibly-nonstandard but accepted server behaviour,
which, for a client, is pointless.

The fact that support was accidentally added and later removed shows
that adding this support should be trivial, as it's supported by the
underlying HTTP library (libneon). I hope you'll consider restoring
cookie support to cadaver so this utility can be useful not only in the
ideal, RFC-compliant-sunshine world, but also in the real world,
populated by real servers that may not always abide by the letter of the
RFCs.

Thanks! and Cheers!
     Keith Beckman



More information about the cadaver mailing list