@andreaturli, no offense, but take your time to check that yourself :)

What you propose would cause every 500 error response in that api call to be 
captured and transforned to a null return value. Is that the desired behavior? 
By definition, 500 errors should be server side errors that should disappear 
after q certain amount of time. In other words, retrying the request should 
succeed at some point (that is why jclouds provides the default retry handlers 
for 5xx errors).

This case seems to be a bit different and to require a more custom error 
handling, to convert the return value to null only if the error is the one you 
identified, and otherwise propagate the error so the retry logic is executed.

Does this sound good?

---
Reply to this email directly or view it on GitHub:
https://github.com/jclouds/jclouds/pull/199#issuecomment-28220891

Reply via email to