Conversation
Signed-off-by: Gaël Écorchard <gael.ecorchard@cvut.cz>
henningkayser
left a comment
There was a problem hiding this comment.
@galou thanks for your PR! I agree, it makes perfect sense to use steady time here. I bet there are other cases in the codebase where we should fix this as well. Could you target main with your PR? We can backport it to foxy.
| bool KDLKinematicsPlugin::timedOut(const rclcpp::Time& start_time, double duration) const | ||
| { | ||
| return ((node_->now() - start_time).seconds() >= duration); | ||
| static rclcpp::Clock clock{ RCL_STEADY_TIME }; |
There was a problem hiding this comment.
Can you make clock a class member steady_clock_?
There was a problem hiding this comment.
OK for both but I cannot compile main so this will be a blind PR.
There was a problem hiding this comment.
main only builds for galactic or rolling, I presume that is why it does not build for you. If it is something else please let us know and we are happy to help. Again, thank you for this change.
Codecov Report
@@ Coverage Diff @@
## foxy #793 +/- ##
==========================================
- Coverage 46.81% 46.80% -0.00%
==========================================
Files 185 185
Lines 19693 19695 +2
==========================================
- Hits 9217 9216 -1
- Misses 10476 10479 +3
Continue to review full report at Codecov.
|
Signed-off-by: Gaël Écorchard <gael.ecorchard@cvut.cz>
|
I have posted my report in the new PR ... see here #795 (comment) |
Signed-off-by: Gaël Écorchard gael.ecorchard@cvut.cz
Description
The
searchPositionIKcan hang forever when using a simulated time because the time is stuck while waiting for this function to return. For this simple reason the function never times out. This was reported by @jmarsik, I'll let him write his full report here. Most of the work, i.e. isolating the cause, was done by him.Checklist
I don't know how to write a test for this. It doesn't appear to be worth it.