Hi, I spent a few days going through the behavior of the WebDAV Servlet, expanding on a test scenario, and looking at the most recent RFC. I found many items to fix.
Three items remain: - Propatch support: it is useless without support for dead properties (dead properties are custom properties, which are basically XML fragments). Since storing them is a major endeavor, I would recommend not implementing this. Also processing of propatch is supposed to be atomic, that would be another implementation difficulty. - If header support. The if header is rather convoluted, but it would still be useful to have. - Lock null has been removed from the specification and this email is about that. Lock null is one of the most convoluted features of WebDAV (along with the if header and the XML fragment properties, of course), and it has been removed in RFC 4918. http://www.webdav.org/specs/rfc4918.html#lock-null instead servers are required to create an empty resource (not a collection) when locking a non existing path, and then create a lock on it. This interoperates nicely with the existing behavior except for some edge cases (the user could have wanted to create a collection, and so on). I verified that my test scenario only needs one trivial change to pass (a PUT will now return NO_CONTENT instead of CREATED since it will now overwrite the locked empty resource; most likely a real client will be happy with both status codes) and I believe this gives a good idea on the impact of the change. Since this is removed in the specification, my plan is to drop the lock null feature. However, since this is definitely a behavior change, maybe it could be done only in main and 11 ? Or only in main ? Comments ? Rémy --------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscr...@tomcat.apache.org For additional commands, e-mail: dev-h...@tomcat.apache.org
