Comments on: RESTEasy MOM: An exercise in JAX-RS RESTful WS design

By: javamonkey79

javamonkey79 — Mon, 23 Jan 2012 15:53:14 +0000

So if I read you right, it is possible to have only one MBR and MBW instead of one for each class in our DOM? If this is the case, then I wonder why this functionality isn’t provided out of the box?

Restlet does have a JAX-RS implementation for the server side, but not the client side. That being said, I don’t see the point. You don’t get the dynamic path mapping that JAX-RS buys you.

By: billburke

billburke — Mon, 23 Jan 2012 15:41:44 +0000

I don’t know what your custom XML marshalling looks like or how it works, but MessageBodyReader/Writer can be tied to a specific mediatype via @Produces/Consumes annotations, also, look at the interface for MBR and MBW, you’ll see isReadable, isWritable methods. If you can easily match your custom classes using these methods, then there’s no reason you can’t have one MBR/MBW for your business logic.

I hate to correct you on the competition, but, I thought Restlet had a JAX-RS implementation?

By: javamonkey79

javamonkey79 — Mon, 23 Jan 2012 15:35:17 +0000

We have a large in house library for our XML representation of our business logic. Something like:

…

This is a trivial representation. In reality there are hundreds of classes.

So in order to make this work I would need to write a MessageBodyReader/Writer for every class? eg: foo, bar, baz, etc.? Or would it be possible to use reflection and only do this once for all of them? If reflection is possible, do you have any thoughts on why the library itself doesn’t provide it?

I’ve evaluated Restlet as well, and one of the things I really like about Restlet is that Object marshaling\unmarshaling is handled with no developer interaction – you just add XStream to you POM and it does the work for you. However, Restlet is not JAX-RS compliant.

By: billburke

billburke — Mon, 23 Jan 2012 14:01:21 +0000

Not sure what your constraints are, but you could write your own MessageBodyReader/Writer for the types you need it done for.

By: javamonkey79Shaun

javamonkey79Shaun — Sat, 21 Jan 2012 01:59:55 +0000

How would use use JAX-RS if you can not apply JAXB to the objects you are passing over the wire? Is this possible?

By: billburke

billburke — Thu, 24 Feb 2011 03:01:59 +0000

Check out http://jboss.org/hornetq/rest and let me know if it fits your needs, or is overkill.

By: nav

nav — Thu, 24 Feb 2011 02:53:13 +0000

Hi bill,

We are using RESTeasy. We wish to forward the request coming to RESTful service to a queue. The Listener of the queue will invoke a JBPM process.
Can you put your ideas on this flow?

Thanks
Naveen Gayar

By: John Harby

John Harby — Mon, 12 Oct 2009 20:55:13 +0000

Hey Bill,

I wondered if anyone has built an async interface to REST without JMS as in JAX-WS.

Thanks,
John Harby

By: Ross Duncan

Ross Duncan — Sun, 31 May 2009 07:56:56 +0000

Hi Bill,

I know this is a little old, but Im really intrerested in this approach to ubiquitous access to a messaging system.

I believe the need is there for a greater decoupling of messaging clients from the servers implementation. Projects I have seen recently have run into severe difficulties talking to messaging systems from disparate JMS providers, due to the clients needing to very carefully manage the messaging client libraries for each provider concurrently. Without an OSGi system, this can get really tricky indeed.

REST as an architectural style seems to bring a lot to the table when trying to solve this problem, particularly providing a standard coms protocol, if we assume HTTP.

My main concern or challenge I see with providing a RESTful interface to a messaging system (JMS for example) would be allowing for XA interactions with the messaging provider. Of course this more of a WS problem generally. Is this something you have considered?

Cheers,
Ross

By: Sergei

Sergei — Mon, 03 Nov 2008 15:05:48 +0000

the standard HTTP 1.1 definition of PUT tells the user of our messaging system that sending a message is idempotent and they don’t have to worry about performing resends

Strictly speaking this is untrue. RFC-2616 (HTTP/1.1) explicitly states that "HTTP/1.1 does not define how a PUT method affects the state of an origin server."