This is in response to Greg Wilkins's blog about the Servlet 3.0 Public Draft.
We have discussed some of the use cases on the EG for things you have said do not have use cases. However, they are not in the Servlet spec and I think they should be. Also, I think your criticism of the JCP should be separate from your arguments about the Servlet spec. They are two different issues and when they are put together it just invites more defensiveness on the part of folks who have a higher interest in making JCP successful. Fixing the JCP is not something that's going to be fixed before we decide on a Final Draft, so I hope we can simply focus on the servlet issues at hand.
I'm not personally a huge fan of using forward semantics in the async case especially because it clears out any content already written which seems like a prime candidate for a reason to keep the wrappers. However, I do think it is better than not having any redispatch ability at all.
I still agree there should be a mechanism for excluding/including jars you need to scan for framework pluggability. However, that is not a showstopper for me and can certainly be a value add feature of any app server.
9 months ago