[undertow-dev] undertow is redirecting?

[Jason Greene]

It's not just content root's that need to be redirected, but also all subdirectories. Check out the rules in 10.10

On Aug 7, 2013, at 2:36 AM, Stuart Douglas <sdouglas at redhat.com> wrote:

> 
> This is an issue that will be fixed once the latest Undertow release is merged into WF core. It came about as a result of fixing
> 
> https://issues.jboss.org/browse/UNDERTOW-89
> 
> With the latest code it only redirects if the target is the war root without a / and the target servlet is registered as the default servlet.
> 
> I have done a Undertow release with this fix, I just need to sort out some of the development mode stuff before it can be merged into core. 
> 
> Stuart
> 
>> 
>> 
>> On Aug 6, 2013, at 6:55 PM, Bill Burke <bburke at redhat.com> wrote:
>> 
>>> Client is doing a get request:
>>> 
>>> GET /jaxrs_spec_resource_requestmatching_web HTTP/1.1
>>> 
>>> 
>>> Response is is a 302.  Notice that there is no trailing / for the path.
>>> Is this part of tthe servlet spec or something?
>> 
>> Is this a WAR root or subdirectory (and not a servlet)? if so then yes, take
>> a look at section 10.10 of the servlet spec. It just uses the term
>> "redirect" and in all other places in the spec they refer to 302 with the
>> word redirect, and never 301.
>> 
>> 
>>> Did Undertow used to
>>> send a 301, or just invoke on the war's root?
>> 
>> I think it used to forward.
>> 
>>> 
>>> I have a JAX-RS TCK test that is regressing.  It is expecting a 200 or
>>> 301 response back from the server.  I need to know how to word the
>>> challenge to allow 302 as well, or Undertow needs to be fixed.
>> 
>> I am not sure which makes more sense. 301 would be more efficient but I can
>> think of some contrived scenarios where you change the layout of your war
>> and the client misbehaves.
>> 
>> e.g you have foo.war/blah/ as a directory but then you later map blah to a
>> servlet, so then the servlet gets the trailing slash which it somehow
>> mishandles.
>> 
>> Hmm. As I say that I am leaning more towards 301.
>> 
>> IMO the TCK should accept any redirect status code, since it is not
>> restricted in the spec.
>> 
>> 
>>> 
>>> Thanks,
>>> 
>>> Bill
>>> 
>>> --
>>> Bill Burke
>>> JBoss, a division of Red Hat
>>> http://bill.burkecentral.com
>>> _______________________________________________
>>> undertow-dev mailing list
>>> undertow-dev at lists.jboss.org
>>> https://lists.jboss.org/mailman/listinfo/undertow-dev
>> 
>> _______________________________________________
>> undertow-dev mailing list
>> undertow-dev at lists.jboss.org
>> https://lists.jboss.org/mailman/listinfo/undertow-dev
>>