Expect 100 Continue failure recovery should include immediately trying the same request again

[http://blog.nerdbank.net/]

The StandardWebRequestHandler class detects 417 errors and changes the ServicePoint to not send Expect 100 Continue headers to that server any more. However it does not try the request again, causing the first request to fail, potentially causing a login failure.

Since we have the capacity to clone requests we might want to retry the request immediately in these cases, or perhaps check for this specific error while forming associations and try it again at that level.

See Andrew's first response in the thread:  http://groups.google.com/group/dotnetopenid/browse_thread/thread/8d691fb6aacc4544

[http://blog.nerdbank.net/]

v3.0 70dcf0ab0a29ab