https://bz.apache.org/bugzilla/show_bug.cgi?id=63191
--- Comment #19 from Boris Petrov ---
Still happening on Tomcat 9.0.54 and CometD 6.0.4.
--
You are receiving this mail because:
You are the assignee for the bug.
-
To
https://bz.apache.org/bugzilla/show_bug.cgi?id=63191
--- Comment #18 from Boris Petrov ---
Still happening on Tomcat 9.0.48 and CometD 6.0.3. Here is a blocked thread's
stack:
java.lang.Thread.State: WAITING (parking)
at jdk.internal.misc.Unsafe.park(java.base@16.0.1/Native Method)
https://bz.apache.org/bugzilla/show_bug.cgi?id=63191
--- Comment #17 from Boris Petrov ---
Still happening on Tomcat 9.0.39 and CometD 5.0.1. It seemed to work for a day
or two (being accessed mostly from Chrome and Firefox) but then a connection
from Safari broke it again. Bot sure how this
https://bz.apache.org/bugzilla/show_bug.cgi?id=63191
--- Comment #16 from Boris Petrov ---
Just a heads up. The same still happens on Tomcat 9.0.34 and CometD 5 (on 4
too). Doesn't happen *at all* when using CometD 3 (which uses blocking IO).
--
You are receiving this mail because:
You are the
https://bz.apache.org/bugzilla/show_bug.cgi?id=63191
--- Comment #15 from Boris Petrov ---
So, actually the bug is still here even on `9.0.29`, although maybe less often,
yes. Actually it happens mostly (only?) with Safari on Mac OS X, not that it
should matter. This is the stacktrace when
https://bz.apache.org/bugzilla/show_bug.cgi?id=63191
Boris Petrov changed:
What|Removed |Added
Status|NEEDINFO|RESOLVED
Resolution|---
https://bz.apache.org/bugzilla/show_bug.cgi?id=63191
--- Comment #13 from Mark Thomas ---
I read through the changelog since 9.0.14 for WebSocket and Coyote and there
were a handful of things that could have impacted this. They might have fixed a
bug, they might just have changed the timing
https://bz.apache.org/bugzilla/show_bug.cgi?id=63191
--- Comment #12 from Boris Petrov ---
Mark, I can't be completely certain but this issue seems to be resolved in
9.0.27 (probably in some of the previous ones too). Have there been any/many
changes between the last broken version until today
https://bz.apache.org/bugzilla/show_bug.cgi?id=63191
--- Comment #11 from Mark Thomas ---
Going back to the original thread dump with the "long" and the "short" stack
traces...
It is possible that the short stack trace is Tomcat trying to send the rest of
the message sent via Async.sendText()
https://bz.apache.org/bugzilla/show_bug.cgi?id=63191
--- Comment #10 from Boris Petrov ---
(In reply to Mark Thomas from comment #9)
> It is not clear that Tomcat not calling the callback for Async.sendText() is
> the cause of CometD failing to complete a Promise rather than the other way
>
https://bz.apache.org/bugzilla/show_bug.cgi?id=63191
--- Comment #9 from Mark Thomas ---
It is not clear that Tomcat not calling the callback for Async.sendText() is
the cause of CometD failing to complete a Promise rather than the other way
around.
With Callbacks triggering the sending of new
https://bz.apache.org/bugzilla/show_bug.cgi?id=63191
--- Comment #8 from Boris Petrov ---
java.lang.Exception
at
org.cometd.websocket.server.WebSocketEndPoint$Delegate.send(WebSocketEndPoint.java:102)
at
https://bz.apache.org/bugzilla/show_bug.cgi?id=63191
--- Comment #7 from Boris Petrov ---
In the GitHub thread I also explained that I tried putting try/catch around the
sendText call exactly because sometimes the callback is not called but rather
an exception is thrown:
https://bz.apache.org/bugzilla/show_bug.cgi?id=63191
--- Comment #6 from Mark Thomas ---
There are some error conditions that (as per the spec) mean that the
javax.websocket.RemoteEndpoint.Async.sendText callback is never called. So yes,
it is perfectly possible for the bug to be elsewhere.
https://bz.apache.org/bugzilla/show_bug.cgi?id=63191
--- Comment #5 from Boris Petrov ---
I'm not sure I understand.
If you do believe my claim that a call to
javax.websocket.RemoteEndpoint.Async.sendText never calls the callback, is it
possible the bug to be anywhere else other than Tomcat? I
https://bz.apache.org/bugzilla/show_bug.cgi?id=63191
--- Comment #4 from Mark Thomas ---
So far, you haven't provided any evidence to back up a claim of a Tomcat bug.
That the changes between CometD 3 and CometD 4 trigger the issue look to be
valid based on the evidence provided. The claim that
https://bz.apache.org/bugzilla/show_bug.cgi?id=63191
--- Comment #3 from Boris Petrov ---
Thank you for the detailed answer.
I'll start with the end of your message.
In the GitHub issue I answered a couple of days ago with the following message:
I downgraded to CometD 3.1.8 and the problem
https://bz.apache.org/bugzilla/show_bug.cgi?id=63191
Mark Thomas changed:
What|Removed |Added
Status|NEW |NEEDINFO
--- Comment #2 from Mark
https://bz.apache.org/bugzilla/show_bug.cgi?id=63191
--- Comment #1 from Boris Petrov ---
When it happens, there are two specific threads in a stack dump:
---
java.lang.Thread.State: BLOCKED (on object monitor)
at
19 matches
Mail list logo
