> I guess if users call getInput directly instead of using a Guava wrapper they 
> will have to cope with the checked 
> IOException. Unfortunately the Payload implementors wrap IOException in an 
> unchecked RuntimeException. This 
> is unfortunate and I would really like to declare the IOException instead of 
> hiding it.

Nothing against that, but that would make it a breaking change, so something 
for 1.8.0 then? And I guess it's not possible, as a transitional step, to 
provide a backwards-compatible `getInput` that swallows the exception but 
immediately deprecate it and suggest that callers use Guava or the inherited 
version instead?

---
Reply to this email directly or view it on GitHub:
https://github.com/jclouds/jclouds/pull/233#issuecomment-30624157

Reply via email to