Hi all, We've recently upgraded to restlet 2.0 and are facing a number of odd 'hangs' during our junit testcases.
As far as we can tell now, we're facing some variant of what is mentioned here: http://restlet-discuss.1400322.n2.nabble.com/Suspended-thread-doing-Client-put-td4479317.html And the answer there is that representations (actually the streams in there) should be "exhausted" for sure. Up to now we were more down the path of following the advise of http://restlet-discuss.1400322.n2.nabble.com/Problem-with-HttpClient-extension-org-restlet-ext-httpclient-and-unreleased-connections-during-404s-td5284804.html#a5284804 Which points out representations should be "released" properly. However in our case this doesn't seem to be enough any more. Quickly scanning the code the biggest difference seems to be that streams are completely read-out before being closed in the exhaust(). My questions: * Should both advises be combined into "assure exhaust _and_ release"? * (For my interest) Why is reading the stream till the end making a difference anyway. What additional resource is cleaned up at the end of the stream that isn't/can't be freed upon close()? * Is there a reason why release() is not automatically including exhaust() regards, -marc= ------------------------------------------------------ http://restlet.tigris.org/ds/viewMessage.do?dsForumId=4447&dsMessageId=2647412
