fix(rclcpp_action): Fix sleep of expire thread in case of canceled timer

[jmachowinski]

This fixes a bug, that the expire action thread would not sleep as, the sleep duration was not computed correctly.

Fixes #2799

[jmachowinski]

Pulls: #2800
Gist: https://gist.githubusercontent.com/jmachowinski/b788276df668933ea120184f152756af/raw/79f8aa943bf792cecb497d0c0234b9a887e5bf27/ros2.repos
BUILD args:
TEST args:
ROS Distro: jazzy
Job: ci_launcher
ci_launcher ran: https://ci.ros2.org/job/ci_launcher/15652

• Linux
• Linux-aarch64
• Linux-rhel
• Windows

[jmachowinski]

Pulls: #2800
Gist: https://gist.githubusercontent.com/jmachowinski/3eb0cc44967b6a96ae007084db7be5b1/raw/79f8aa943bf792cecb497d0c0234b9a887e5bf27/ros2.repos
BUILD args:
TEST args:
ROS Distro: jazzy
Job: ci_launcher
ci_launcher ran: https://ci.ros2.org/job/ci_launcher/15659

• Linux
• Linux-aarch64
• Linux-rhel
• Windows

[Timple]

@marcoag if this fix is merged, would a quick re-sync for Jazzy be possible?

[jmachowinski]

Test failures are unrelated.
test_clock : multiple_threads_wait_on_one_clock is missing mutexes. We should make a good first issue ticket from this one...

[jmachowinski]

FYI #2804

[marcoag]

@marcoag if this fix is merged, would a quick re-sync for Jazzy be possible?

Will try to get one out asap, so we can put the fix out, thanks @Timple!

[Timple]

Cool, what would asap look like? As a typical cadence is 3 weeks and we're past week 2 😉

[marcoag]

Cool, what would asap look like? As a typical cadence is 3 weeks and we're past week 2 😉

I'll put the hold today and get everything ready. I was waiting for a fix on the current regression but there doesn't seem much movement there, I'll give them one more day before I sync.