> I can't comment much on the internals of payload slicing (it would be 
> interesting if @adriancole or @andrewgaul had a chance to review this), but 
> is there any reason for introducing a new interface rather than extending the 
> existing PayloadSlicer?

It probably could, yes.  I'm not very familiar with the projects structure, nor 
the intended scope of PayloadSlicer, my thoughts were that this would be apply 
to multipart strategies, and so maybe that was made clearer be encapsulating 
that in it's own interface.

My convictions aren't very strong on this though.

> This would allow us to avoid the breaking change to the SMUStrategy 
> constructor.

Is it common for people to use this class directly?

---
Reply to this email directly or view it on GitHub:
https://github.com/jclouds/jclouds/pull/192#issuecomment-27094783

Reply via email to