Revert "Reset RMWCount when DEALLOC rmw storage of wait set"#210
Revert "Reset RMWCount when DEALLOC rmw storage of wait set"#210
Conversation
This reverts commit 1d35c60.
|
I know what issue is, RMWCount should be passed as parameter. I'll re-submit a PR |
|
@dirk-thomas @wjwwood Sorry about that, I should have run and test it locally, but somehow neglect. I'll be more careful later on. On the other hand, it remind me that if ROS2 CI can provide a quick pre-test (ament test) mechanism for each new PR? |
|
@jwang11 normally we run CI, but since it takes so long, we sometimes try to avoid it on trivial changes, but then again it bites us sometimes. |
|
@wjwwood Formal CI is fine, which can be triggered manually when PR is approved, like you did now. what I mean is another kind of lightweight CI test, something like "ament test compoent_path". It can be auto-triggered when new PR arrive and be done in 10 minutes. |
|
Yeah, we're working on improving the CI, especially for external contributions. The problem is that if your change causes downstream changes it still bites us on our nightlies and larger changes which need "formal CI". It would help us with more trivial changes like this as a trip wire though. |
Reverts #209
I believe this is a regression based on http://ci.ros2.org/job/ci_osx/3214/consoleFull. Since there was no CI (that I saw on the origin pr) I'll assume that's the case and revert until it can be tested separately.