Hi Stephan
On 01/02/12 06:45, Klevenz, Stephan wrote:
Hi,
would it make sense to open a Jira issue?
I'm interested in getting this fixed
sure, this can be fixed very easily.
and also interested in getting a
better separation of JAX-RS runtime classes if an other implementation
like Jersey is reachable by classloader.
I'm not sure yet how realistic this can be, having Jersey loaded by the
same loader which loads CXF, and have it working.
We can indeed, as suggested by Dan, replace all calls like
MediaType.valueOf() (and similar) with direct calls to concrete
implementation classes. This may be worth it on its own, it will be
CXF's best effort.
But we will be 'powerless' once the application code or filter does
Response.ok() which will create a custom ResponseBuilder impl, or use
some CacheControl utility, or ETag one, which depend on static factory
methods...
Cheers, Sergey
Regards,
Stephan
-----Original Message-----
From: K Fung<[email protected]>
Reply-To: "[email protected]"<[email protected]>
Date: Tue, 31 Jan 2012 18:07:05 +0100
To: "[email protected]"<[email protected]>
Subject: Re: Jersey / CXF compatibility
Hi Sergey,
By sending an explicit Accept header, this will block HttpUrlConnection
from sending an invalid header to CXF. CXF was (correctly!) spitting
out a
406 when the HttpUrlConnection was sending:
text/html, image/gif, image/jpeg, *; q=.2, */*; q=.2
is it a single '*' which causes CXF to reply with 406 ? May be CXF
should be more tolerable to single "*" which I guess is a legacy
wildcard value...I think it should.
Yes, it's throwing the IllegalArgumentException on the single star
(it's expecting<something>/<something>). CXF could be more lenient on
that.
-kl
