Move timer objects to the end of member list #5
Conversation
In high CPU load it can happen, that Timer thread wakes up, is swapped out, and when application resumes it does not resume with timer thread, but transport thread, that decides to do failover, destroys `InactivityMonitor` object. When its members are getting destroyed, they arrive at `Timer` object that joins, timer thread resumes in `writeCheck` or `readCheck` method, only to find out `asyncTasks` is already destroyed because it comes later in member declaration list so it was destroyed first. Moving `Timer` object to the end of the list, guarantees that they will be destroyed first, so their threads will get joined and allowed to finish, before the rest of the object is destroyed. Closes AMQCPP-626
|
Guys. Is there anything that needs to happen, so this PR gets some (any) attention? |
|
Just wait and keep patient, or maybe you would like to send an email to the master. |
|
Needing to wait almost three weeks for a patch for a serious, crash-causing, race-condition to receive any attention, doesn't exactly build trust in a library. I have sent 2 emails to dev list, as well as logged JIRA. Not sure what (who?) the "master" is, but I feel like I've done enough. Thank you for the tip tho. I'll just wait and keep being patient. |
|
@Pingmin thank you for trying, but after one month with no reaction whatsoever, and more than a year since last release, I should probably come to a conclusion that this project is effectively abandoned. |
|
@unjello I don't think so and we don't want to see that. Maybe He is too busy in these months... |
|
Many years later. Any news? |
|
Good luck @unjello ! I was about to do a PR to update zlib and get rid of a CVE alert we have. I don't think I'll waste time with that. |
|
I'm an ActiveMQ committer and would be able to help modernize. My C++ is a bit rusty, so I'd need an assist. I think first step would be to modernize the tool chain (add git action and jenkins file to get some automated build testing going. Would either of you be able to assist in PRs for that? |
|
The patch code is incomplete. The initialization list in the constructor needs modification as well. |
|
Hi, don’t want to come across rude, but 7 years later I barely use C++ anymore and don’t have any use for AMQ either. I won’t have time to assist with this, sorry. |
In high CPU load it can happen, that Timer thread wakes up,
is swapped out, and when application resumes it does not resume with timer
thread, but transport thread, that decides to do failover, destroys
InactivityMonitorobject. When its members are getting destroyed,they arrive at
Timerobject that joins, timer thread resumes inwriteCheckorreadCheckmethod, only to find outasyncTasksis alreadydestroyed because it comes later in member declaration list so it was
destroyed first. Moving
Timerobject to the end of the list, guaranteesthat they will be destroyed first, so their threads will get joined and
allowed to finish, before the rest of the object is destroyed.
Closes AMQCPP-626