[BUG] State machine process shutdown with coredump with Ctrl+\(SIGQUIT).

[ZhenshengLee]

Describe the bug
State machine process shutdown with coredump with Ctrl+(SIGQUIT).

To Reproduce
Steps to reproduce the behavior:

simply execute the program

~/zs_ws/ga_ros/colcon/install/sm_guardian/lib/sm_guardian$ ./sm_guardian_node

Expected behavior
program canbe shutdown gracefully.

Screenshots

Environment (please complete the following information):

ROS DETAILS:

• OS: Ubuntu 20.04 LTS
• Version: Humble

REPO DETAILS:

• Repository: [e.g. robosoft-ai/SMACC2]

┆Issue is synchronized with this Jira Task by Unito

[pabloinigoblasco]

I have been testing this on galactic with sm_atomic.
I also get that same error.

However, debugging I do not detect any raised exception or crash in our code so I am not sure if this is our fault.
I think that is something expected using SIGQUIT. I tested that with other programs and I get the same "core dumped" error.

for example, try with "sleep 1000"

Nonetheless I am still analyzing this. I opened a PR subscribing to the SIGQUIT signal to see if we can get more insights about the issue.

#377

[ZhenshengLee]

From https://ftp.gnu.org/old-gnu/Manuals/glibc-2.2.3/html_node/libc_462.html I believe there will always a coredump if the SIGQUIT Handler is not installed,

after #377 merged, this canbe closed.

[pabloinigoblasco]

I am afraid that it keeps happening even if we install SIGQUIT handler :S

I would not worry too much about this, if that message happen with all linux tools (ls, watch, sleep, etc) , this is not a problem or a bad design.

Nonetheless, I think it is not a bad idea to subscribe to some of this signals in the signalDetector file as we have in #377 so we will merge this anyways.

[pabloinigoblasco]

Merged to humble and galactic.
I close the issue.