Open andre-merzky opened 11 years ago
I don't see a way how to resolve this cleanly ...
One way is to try the next adaptor on BadParameter exceptions on object creation (or to introduce a new internal exception 'BadSchema').
(not hindered by any knowledge on the issue) Configuration?
What kind of configuration? I don't think we want to store hostnames for which the http URL us handled by adaptor A, and for other hostnames its handled by adaptor B?
On 14 Sep 2014, at 17:06 , Andre Merzky notifications@github.com wrote:
What kind of configuration? I don't think we want to store hostnames for which the http URL us handled by adaptor A, and for other hostnames its handled by adaptor B?
I can throw in another random word if you prefer :-)
Shouldn't we use different scheme's for these adaptors?=
Different schemes, or resource specific configurations. But the AWS adaptor is outdated in the first place... :/
Right now,
http://
URLs are routed towards the eucalyptus code path (for no particular reason) -- but it is not clear how to distinguish backend types based on the URL alone...