improve exception handling for exceptions that implement to_hash
Fixed: rake handler failing when exception happens before Airbrake has been configured
Call `close` on the old body in UserInformer middleware
I think I stumbled across a bug in the user_informer - for me it always showed the same error id (always that of the first error to occur after an app restart), even if new errors were created. Reason: in user_informer.rb you're putting the replacement text into a global, which isn't a good thing imho as the middleware instance appears to be re-used across requests. Update: While the global might be fine (as long as no threads are involved), using it as a cache with ||= isn't. Happened to me in a Rails 2.3.14 / 1.8.7 / Passenger environment, and removing the global fixed it. Cheers, Jens
…read joining errors
… a deploy notification
…ted the cucumber feature to look for the right task name.
example: Savon::SOAP::Fault see: http://help.hoptoadapp.com/discussions/hoptoad-notifier-patches-and-discussion/67-exceptions-that-implement-to_hash-lose-all-of-their-useful-data This replays thoughtbot/hoptoad_notifier#25 for airbrake
… been configured (for example, when running an unknown Rake task)
Change Hoptoad references to Airbrake.