[Bug 53391] CometEvent.close() doesn't close socket when called from different thread

https://issues.apache.org/bugzilla/show_bug.cgi?id=53391 --- Comment #10 from Mark Thomas --- Glad to hear that all is OK. It wasn't a waste of time - the new test cases I wrote for Comet are good to have. -- You are receiving this mail because: You are the assignee for the bug. --

[Bug 53391] CometEvent.close() doesn't close socket when called from different thread

https://issues.apache.org/bugzilla/show_bug.cgi?id=53391 --- Comment #9 from Lawrence Kesteloot --- Hi Mark, I made a test servlet and it worked fine: the response was properly ended when I called event.close() from another thread. I was thrown off for hours by bug 51881, and finally when I upgra

[Bug 53391] CometEvent.close() doesn't close socket when called from different thread

https://issues.apache.org/bugzilla/show_bug.cgi?id=53391 --- Comment #8 from Mark Thomas --- Again, Tomcat does correctly signal the end of the response as per RFC2616. If you have a test case where it doesn't please re-open this bug. -- You are receiving this mail because: You are the assignee

[Bug 53391] CometEvent.close() doesn't close socket when called from different thread

https://issues.apache.org/bugzilla/show_bug.cgi?id=53391 --- Comment #7 from Lawrence Kesteloot --- Got it. You were right that this is an invalid bug. Other async web servers like Tornado must transparently buffer the output and set the content length themselves, so "it just works", and I expect

[Bug 53391] CometEvent.close() doesn't close socket when called from different thread

https://issues.apache.org/bugzilla/show_bug.cgi?id=53391 --- Comment #6 from Mark Thomas --- You don't mean close connection either. You mean the client is unable to determine the end of the response. This is usually signalled by one of the following: - content length header plus that many bytes

[Bug 53391] CometEvent.close() doesn't close socket when called from different thread

https://issues.apache.org/bugzilla/show_bug.cgi?id=53391 --- Comment #5 from Lawrence Kesteloot --- That should say "NOT from within the event() method". -- You are receiving this mail because: You are the assignee for the bug. --

[Bug 53391] CometEvent.close() doesn't close socket when called from different thread

https://issues.apache.org/bugzilla/show_bug.cgi?id=53391 --- Comment #4 from Lawrence Kesteloot --- I think I said "close socket" when I really meant "close connection". Here's the behavior I'm seeing: If I do "curl URL" from the command line, and in BEGIN I respond and ComentEvent.close(), the c

[Bug 53391] CometEvent.close() doesn't close socket when called from different thread

https://issues.apache.org/bugzilla/show_bug.cgi?id=53391 Mark Thomas changed: What|Removed |Added Status|NEW |RESOLVED Resolution|---

[Bug 53391] CometEvent.close() doesn't close socket when called from different thread

https://issues.apache.org/bugzilla/show_bug.cgi?id=53391 --- Comment #2 from Mark Thomas --- I'm not convinced that the behaviour you are seeing is a bug. Currently: a) END if called from within event() closes the connection b) END if called from a separate thread does not close the connection

[Bug 53391] CometEvent.close() doesn't close socket when called from different thread

https://issues.apache.org/bugzilla/show_bug.cgi?id=53391 --- Comment #1 from Lawrence Kesteloot --- The first paragraph of this message from 2008 is what I'm seeing: http://bill.burkecentral.com/2008/11/12/buggybroken-tomcat-6-comet-nio-apis/ -- You are receiving this mail because: You are the