https://bz.apache.org/bugzilla/show_bug.cgi?id=66676
--- Comment #4 from Georg F. <magicb...@gmx.de> --- (In reply to Mark Thomas from comment #3) > There are two code paths to this line. > The path via WsSession.doClose() will have set WsSession.state to > OUTPUT_CLOSING. > The path via WsSession.onClose() will have set WsSession.state to CLOSING. Right - my fault: I've mixed up "doClose" and "onClose". Sorry. :( Then I don't have a real idea of how this happened. I'm working in a multithreaded environment, so I'm wondering if there is some way of having a race condition here? Or is everything secured by some locking mechanism? What I can tell is, that I first see that a send on "thread A" fails due to a broken connection to the client. Shortly after "thread B" calls the close() method and shows this stacktrace. Might be that thread B even calls close() while the sending fails - can't really tell that from the logs I have. Or would that already be an illegal procedure? -- You are receiving this mail because: You are the assignee for the bug. --------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscr...@tomcat.apache.org For additional commands, e-mail: dev-h...@tomcat.apache.org
