-
Notifications
You must be signed in to change notification settings - Fork 36
Fix compilation issues in moveit2 (#224). #225
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
Conversation
@Bckempa please don't put this in this sprint/release yet. Let's see if we can get it to minimally work and we can decide if this can be considered for this release. |
ec74c78
to
a56f922
Compare
4933dc4
to
7d6b0e8
Compare
Tagging @EzraBrooks @asimonov @eholum @Bckempa @mkhansenbot There seems to be some issue with this image due to
The problem seems to be here: I tried bringing Any clues as to how to make this compile? Could disabling the tests for |
i have seen the same error and stopped. is there not a jazzy version of moveit2 yet? there would be a dockerfile for it with relevant dependencies we can use as inspiration? |
Sorry I was attempting to upgrade my FPGA toolchain at the same time and that never goes smoothly. Back on this now.
Yep, for about 9 months.
The one I found first just uses rolling as a base, but I'm digging deeper. |
Good news is I got this working, the bad is it involves like 4 separate issues. I need some more time to make a coherent patch set from this, but it should also improve the build time and size of the moveit2 image in the process |
@Bckempa excellent. Feel free to tag me to review any other PRs to close this quickly, and to push changes to this branch if anything in this partial solution still applies. We can rebase/squash/credit commits later as needed before this PR is approved. |
Bug with the tests not correctly declaring their dependency on stomp opened upstream: moveit/moveit2#3436 @EzraBrooks can you triage that ASAP? If it's a same-day fix on the Picknik side (I'm happy to contribute a PR if that helps) that would be preferable, otherwise I can add a patch step to the Dockerfile the disable building those two tests. I'm also woking on an issue with |
Would it make sense to temporarily make the Dockerfile just insert the necessary Is it just adding these lines:
I would separate it, yes. That optimization is not preventing installation. It'd be two different issues. |
Fix for stomp merged to main upstream: moveit/moveit2#3437 After that we weed to add one more missing dep here to get to green, standby |
The moveit2 image fails to build due to some dependencies not being found during compilation. This commit adds the missing dependencies to the list of packages needed by moveit2. Co-authored-by: Alexey Simonov <[email protected]> Co-authored-by: Ivan Perez <[email protected]>
7d6b0e8
to
64c42fe
Compare
Pushed the following fix-ups, should be good to go if build is green:
|
@Bckempa No space left in Space ROS build.
|
Ooft, that's an @EzraBrooks question. Good news is it fixed the moveit phase of the build, but we don't know if that broke Space Robots. I'll test this at home tonight (where if builds much faster) to at least give us a heads-up on if we can expect space robots to work as well. |
It's extra annoying that we have to restart the whole job and do the 1.75 hour build that worked again. Will that cache be warm even if the job failed? |
It should be (https://github.com/space-ros/docker/actions/runs/14454616962/job/40534921075#step:6:20571), but if you have to slim the moveit2 image (since this seems to be a space issue), then it won't matter as much. |
I'm confused about why this line attempts to load an exported image that was created in the previous step. https://github.com/space-ros/docker/actions/runs/14454616962/job/40534921075#step:6:20571 |
@Bckempa since the result of building moveit2 was cached, I tried hitting the re-build button and it went on. This makes me suspect that running (some of) the space cleaning actions at the beginning of the CI job again after building and exporting the moveit2 image may help with the space issues. The next error once the build continues is in the compilation of https://github.com/space-ros/docker/actions/runs/14454616962/job/40544607040?pr=225#step:7:1655 possibly due to missing includes of |
6f6789d
to
6ea80e3
Compare
We've gone beyond the scope of the issue and should either rename the issue and PR to reflect fixing the compilation issues for the whole space_robots workflow, or roll back to the point that moveit was fixed, merge, and open a separate issue to work on the space_robots demo container |
Not quite there yet:
|
Why are we using a hand-coded |
Ok there is something really wrong here. I tired generating the repos and it is listing things that are clearly in the overlay and were picked up in the moveit layer. I strongly recommend we close this with just the moveit changes and attack the jazzification of space_robots fresh, it's going to require more work. |
2f945bf
to
6b79bc8
Compare
I think this was @dsobek's finding during the SBIR also |
I'm not sure how else we can free up space, the default runners are pretty wimpy and we don't have funding to supply larger runners. Maybe we need to tweak the optimization level of these builds. |
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.
Too bad it doesn't fix the full build, but getting moveit2
up to jazzy
is a good start
Add missing dependency to see if the moveit2 image compiles that way.