Full Rebuild (Sync) of February 2026: bump ros2-distro-mutex to 0.14.0 and build_number to 16#161
Conversation
|
macOS CI is failing with: This is probably a network flake, we can restart once the other jobs finish (to avoid removing the cache). |
|
I wonder why it takes over 6 hours to finish.. But few successful runs suggest that it might actually work. |
Windows is much slower. However, as all packages are saved in cache, restarting the CI after the 6 hours timeout make it progress from the last time.
Incredible! This was the fasted rebuild I ever saw. : ) |
|
Regarding the |
|
@traversaro Could you restart again? Though it is second or third time already, I hope the runs will succeed for actions except one failing already |
Restarted. It seems that macOS has some problems related to foxglove_bridge, while Windows I suspect it took many rebuilds as the cache is not being saved correctly on timeout. I can wait a few hours and manually kill the job to unsure that the cache is saved correctly. |
|
There is a minor issue, I guess trivial to fix: |
|
Windows is failing with: |
I am not sure what happened, but the |
|
I actually wanted to suggest to remove that from cache. Looks like macos was just fine. I think we are almost there! |
|
I realized that due to https://github.com/RoboStack/ros-jazzy/blob/main/pkg_additional_info.yaml#L156-L157 we need to bump the build number to 16 . However, if the build works fine, we can do that at the end without the need to wait a full rebuild to complete. |
|
For the last Windows failure we need to patch Wx out of https://github.com/foxglove/foxglove-sdk/blob/94d02861d09ae9a95b5c85d0d12e646895584a8b/cpp/CMakeLists.txt#L87 |
|
I wonder whether it makes sense to manually stop the windows run and then restart it again so we don't occasionally loose cache for all the built packages. It already runs for about 5.5 hours |
Yes, unfortunately I was sleeping. : ) Apparently there is something strange going on, as in:
However, I am monitor the Windows jobs and stopping it manually at 3/4 hours to avoid losing the cache. We should be a few hours far from the green CI on Windows as well. |
|
I also added |
This fails with: probably there are some patches in RoboStack/ros-humble#275 did not landed upstream, let disable it for now to unblock the merge. |
|
Thanks @lvjonok ! |
|
Wow, great! I am glad to help! The next question I will have is when do these packages actually arrive to robostack-jazzy channel? |
|
You can find the progress at https://github.com/RoboStack/ros-jazzy/actions Seems to be some issues for some platforms - sorry I don't have time to look into it, my best guess is that it might be related to the (very) recent urdfdom migration on conda-forge. |
No description provided.