-
Notifications
You must be signed in to change notification settings - Fork 2
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
Use absolute paths to reference models in unit tests #168
Conversation
028e055
to
bdb1cd9
Compare
Apparently the modification of unit test paths to models used by the fixture (that was needed BTW) is not sufficient to make the CI pass. |
The error log that can be the symptom of something not working properly is gz-sim-yarp-plugins/libraries/device-registry/DeviceRegistry.cc Lines 50 to 56 in bdb1cd9
This could be an indication that the pointer of the ECM used as the indicator of the Gazebo instance in use could pose issues. What do you think @traversaro? |
From the ForceTorqueTest the error:
and from the ClockTest:
are making me think of something with YARP, but it's strange that these error started occurring the past week, also because the Yarp version has not changed AFAIK. |
This comment was marked as resolved.
This comment was marked as resolved.
Not sure, how did you solved? |
I've did two things so not sure which one is the real solution. First I updated my mamba version, then I copied to the txt file the entire output of
|
In this way I'm finally able to reproduce locally the failures of the tests. |
The error message |
The diff between sdformat 14.0.0 and 14.2.0 is gazebosim/sdformat@sdformat14_14.0.0...sdformat14_14.2.0 . |
This change may be related: gazebosim/sdformat#1239 . |
In particular this function: https://github.com/gazebosim/sdformat/pull/1239/files#diff-497b59a757dc6d26ca6ec70d1e804f60c934bd2c0f53ffbe8551a010dee6844dR46 . |
And all may boil down to gazebosim/sdformat#198 . I am not sure why conda's binary prefix relocation is not working anymore (see https://docs.conda.io/projects/conda-build/en/latest/resources/make-relocatable.html), but that is indeed a tricky machinery (that anyhow does not work on Windows), and that the reason why we switched to explicitly making the packages relocatable in gazebosim/gz-sim#626 . |
Ok thank you so much for unravelling this tangle! So to recap the issue could be due to the fact that when passing the path of the SDF model to the |
@traversaro how do you suggest to proceed? |
Not really. My hypothesis is that he is not able to find anymore some SDF files, that however are not related to the |
Can you try to check a standalone TestFixture example from gz-sim (for example https://gazebosim.org/api/sim/8/test_fixture.html) and see if it reproduces the problem? If yes, I would open an issue in https://github.com/conda-forge/libsdformat-feedstock and pin sdformat to 14.0.0 in the CI in the meanwhile. |
when running it I get the same error:
|
I compiled and ran the standalone tests in https://github.com/gazebosim/gz-sim/tree/gz-sim8_8.3.0/examples/standalone/gtest_setup They report the same errors we are experiencing. gravity_TEST:
command_TEST
|
Great, can you open an issue at https://github.com/conda-forge/libsdformat-feedstock ? Thanks! |
|
Given conda-forge/admin-requests#994 (comment), you can probably remove the pinning of sdformat . |
2fff1f0
to
22bd0e5
Compare
Finally the Conda CI is passing for all platforms. |
It seems that the apt CI is failing? |
Yep, that's addressed by #166, I would merge it after this one. |
Ok! |
Are you sure that disabling codecov will be sufficient? This CI on apt fails with error:
that is different from the error mentioned in that issue. |
22bd0e5
to
c93dcf3
Compare
Oh sorry I didn't see it! |
c93dcf3
to
076727a
Compare
Now the CI is passing. Merging |
Fixes #164