hc-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Kalnichevski, Oleg" <oleg.kalnichev...@bearingpoint.com>
Subject To all those who have been experiencing problems with 'expect: 100-continue' stuff
Date Thu, 06 Mar 2003 09:54:34 GMT
The 'expect: 100-continue' related code has been significantly refactored

- the overall design should have gotten simpler and cleaner
- CPU-bound response polling (busy wait) in WaitForResponse which somehow might have been
contributing to the problem has been removed in favour of a more elegant solution suggested
by Simon Roberts <simon.roberts@fifthweb.net> that relies on the socket timeout
- RFC 2616 compliance problem has been fixed. 1xx range status codes should never be returned
from the HttpClient#executeMethod. Unexpected 1xx status code are now ignored and discarded
as specified in the RFC 2616 

This said, sadly enough I still have no idea what have been causing 100 response to time out
in the very first place. I really need more input from you. I am kindly asking you to re-test
your applications with the newest CVS snapshot and provide me with additional feedback. As
far as I understand, the problem (if it has not been eliminated by the redesign and the busy
wait fix) should now be manifesting itself through warning messages in the log only. Please
watch for the following messages in the logs:

[INFO] HttpMethod - -100 (continue) read timeout. Resume sending the request
...
[INFO] HttpMethod - -Discarding unexpected response: HTTP/1.1 100 Continue

If you succeed in reproducing the problem, I would appreciate your sending me instructions
as to how the problem can be reliably reproduced. A test case would be a fantastic and much
appreciated contribution

Cheers

Oleg




Mime
View raw message