Mark Thomas
Mark Thomas
+1 to a PR,
I agree the TCK challenge is valid. We'll need to exclude this test for now. Exactly how we fix this in the TCK needs some thought.
The current Tomcat behaviour during async IO can be summarised as: - any async IO errors are delivered to WL (if configured) and RL (if configured) - If RL.isReady() triggers...
I think I am OK with the direction this is heading in but I'd like to review the changes once the PR has been rebased to take account of the...
The spec and the Javadoc are not consistent. From the spec > ...the HttpServletMapping is not available for servlets that have been obtained with a call to ServletContext.getNamedDispatcher(). and the...
Hmm. Something doesn't add up here and I need some time to dig into it. I believe that Tomcat implements the behaviour described in the updated Javadoc but Tomcat also...
Figured it out. Tomcat's behaviour is configurable. The default behaviour is consistent with the updated text agreed here but when running the TCK Tomcat is put into "strict spec compliance...
[1] Given we have RequestDispatcher.INCLUDE_QUERY_STRING and RequestDispatcher.FORWARD_QUERY_STRING I think this should be target for consistency. Same for ASYNC. [2] has been fixed to be target [3] is not an issue...
Sure. I'll add a copy of the table with what I think it should be then we can iterate on the differences and hopefully reach consensus.
| REQUEST | FORWARD | NAMED FORWARD | INCLUDE | NAMED INCLUDE | ASYNC | ERROR -- | -- | -- | -- | -- | -- | -- |...