>Now you just need a test harness for your test harness. :) Ad infinitum!
It's not strictly a bug in the harness. The test harness implements one correct scenario for a concurrent "thrashing". It uses the Jersey client API to post requests, by constructing a new client each time. The only state maintained by the harness are atomic counts for the debugging I've been doing. The only states passed within the harness are either Strings or immutable collections of name/value pairs. There is no session state maintained on the server as each request is context complete. With both end-point Jersey resources in operation (each with completely distinct URLs), one resource gets fed duplicates of requests it has already received, the other resource (which I haven't audited) certainly gets some (if not all) of its requests (as it it solely responsible for persisting a specific type of data to the database). It's all rather bizarre and points to a thread-safety issue with the client. Either a bug or I'm using it incorrectly (though not according to the docs I've read). The Jersey client can have Apache's HttpClient 3 plugged in which may be worth a shot. There's an experimental extension for HttpClient 4. At any rate, I need to gather more info to pass on to the Jersey folks. Best download the Jersey source. Sorry if I've disturbed you here! -- View this message in context: http://old.nabble.com/Tomcat-appears-to-be-sending-duplicate-requests-tp32019702p32024536.html Sent from the Tomcat - User mailing list archive at Nabble.com. --------------------------------------------------------------------- To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org For additional commands, e-mail: users-h...@tomcat.apache.org
