[jira] Commented: (HTTPCORE-132) In NIO, there doesn't seem to be a simple way to set a custom HttpRequestFactory

["Rachel Greenham (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/HTTPCORE-132?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#action_12552159 ] 

Rachel Greenham commented on HTTPCORE-132:
------------------------------------------

Cool beans. :-) For my local change I did start off by doing it as a protected factory method, but then changed it to taking a factory in the constructor *primarily* to keep symmetry with the BufferingHttpServiceHandler constructor in 4.0alpha6. But I'm happy with it either way.

I'll wait for 4.0beta1 to actually appear before upgrading to it. :-)


> In NIO, there doesn't seem to be a simple way to set a custom HttpRequestFactory
> --------------------------------------------------------------------------------
>
>                 Key: HTTPCORE-132
>                 URL: https://issues.apache.org/jira/browse/HTTPCORE-132
>             Project: HttpComponents Core
>          Issue Type: Wish
>          Components: HttpCore NIO
>    Affects Versions: 4.0-alpha6
>         Environment: Java 5. Dev platform is OS X Leopard but I doubt that's relevant.
>            Reporter: Rachel Greenham
>            Priority: Minor
>             Fix For: 4.0-beta1
>
>         Attachments: ServerIOEventDispatch.java
>
>
> It appears that for developing a server application using NIO, you can't easily define a replacement HttpRequestFactory. It's hardcoded to create a new DefaultHttpRequestFactory in DefaultServerIOEventDispatch, in a method that can't be overridden because the subclass method also references a *private* data member (the allocator) of DefaultServerIOEventDispatch. It could be done by also instantiating and supplying an external ByteBufferAllocator in the overridden method, which "wastes" the instance created in the constructor. (I notice the constructor that lets you supply an allocator in 4.0-alpha6 is gone from later versions - but I'm not even interested in replacing the allocator, I just want to replace the request factory.)
> The way I am doing it is by actually *copying* the current source of DefaultServerIOEventDispatch to my project (with a rename) to make my own modified version to implement IOEventDispatch - which seems excessive for the small change I actually want. The change I made was simply to have a constructor that allows me to supply an already-instantiated HttpRequestFactory, in a similar fashion to the constructor of BufferingHttpServiceHandler takes an instantiated HttpResponseFactory.
> This also stops this IOEventDispatch from creating a new instance of the *factory* for each hit, which it seems to be doing at the moment...

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@hc.apache.org
For additional commands, e-mail: dev-help@hc.apache.org