-
Notifications
You must be signed in to change notification settings - Fork 939
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
kinematics tests #1272
kinematics tests #1272
Conversation
Note, that the unit test fails due to #1271. |
17b9ca0
to
23526f0
Compare
23526f0
to
ad83b7f
Compare
@rhaschke thanks for reviving this PR, do you plan on making further changes or can it now be reviewed? |
I'm still working on it. Changed the title accordingly. |
f91b05a
to
36d6a99
Compare
@jrgnicho This is ready for reviewing now. |
@jrgnicho: The test getNearestIKSolution checks that |
moved and adapted from https://github.com/ros-planning/moveit_kinematics_tests
61abe24
to
bfabb75
Compare
- read parameters only once and share them between all tests - reduce code duplication, introducing expectNear() - remove timing (done by gtest as well) - simplify getNearestIKSolution()
@jrgnicho @davetcoleman I addressed your comments. Please have another look. |
a967df4
to
afcf6b7
Compare
Other than the linking error it looks good to me. |
The linking issue seems to be caused in the linking step to libmoveit_test_utils |
The linking issue has the following underlying reason:
I searched for a while how to convince clang to link as lazy as gcc does. No solution so far. |
2525a54
to
44ad02d
Compare
I reverted the last changes (moving some testing functions to libmoveit_test_utils. |
44ad02d
to
b4d247e
Compare
b4d247e
to
bed7cc2
Compare
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
You never officially got an approval for this PR. I understand we were taking longer than you'd prefer to review, but notably I was busy in China representing MoveIt!. Please ping us more next time rather than breaking our review policy
|
||
public: | ||
testing::AssertionResult isNear(const char* expr1, const char* expr2, const char* abs_error_expr, | ||
const geometry_msgs::Point& val1, const geometry_msgs::Point& val2, double abs_error) |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Thanks for trying this.
Why can't we install gtest? Testing is important to software and shouldn't be held back due to the desire to reduce dependencies
I considered Jorge's comment in #1272 (comment) as an approval and I pinged you in #1272 (comment). |
Because it's only for testing. This shouldn't be installed. Actually, it's Google's philosophy that each and every software package builds gtest on its own from source. If it would be installed, you could even trigger conflicts with other source packages. |
* Small changes to make the action server more performant * Move the threads to member variables * Add a simple cancellation demo * Do not use unique_ptr * Lint * Launch file cleanup * When a new HP goal is received, give the previous some time to cancel * A unique callback group for each action server
This is another attempt for #1147, addressing #4. Similarly to moveit/moveit_kinematics_tests#7, I tried to separate the tests from robot-specific code: The tests as designed by @jrgnicho are robot- and ik-plugin-agnostic and can be run (in an appropriate environment) for any robot and plugin.
Additionally to the initial feasibility tests, I added some real validation to FK tests: RobotState's FK and KinematicsPlugin's one should coincide. A failure there pointed me to #1271.
All original tests are random feasibility tests that turned out to be flaky, which is why I disabled them.
Instead of adding additional robots to moveit_resources, I'm running those tests on Fanuc + Panda.
I'm planning to add a
ikfast
test for these robots as well.