M. Mei

The pull request for resolving this issue has been opened at https://github.com/ros2/rmw_connext/pull/391.

> When this is done, the tests need to be fixed in `rclpy` to actually test things: > > https://github.com/ros2/rclpy/pull/454/files#diff-960e40f792ccd92281ee56dfb65726bcR197 This change is in pull request https://github.com/ros2/rclpy/pull/511.

I have a feeling that every `ASSERT_`/`EXPECT_` statement we use in our tests contribute to missed branch coverage, because the failure branches are never hit since our tests are always...

If the user for some reason sets the measurement period to be greater than the publishing period, then all bets are off.

Managed to workaround this by adding to the source file: ``` #include #include int main() { throw std::runtime_error("HELLO WORLD!!"); return 0; } extern "C" { void * __cxa_allocate_exception(size_t /*thrown_size*/) {...

Hi, What type of media output is coming out of the `usb_cam` node? If you're trying to see media playback on the Kinesis console, it supports H.264 (see "Q: What...

Hi @java-Xs and @ngthanhtin, do you mean that you're usually encountering the `0x52000011` error but it is successful once in a while? Can you describe approximately what percentage of the...

While this pull request is code complete, the `test_rolling_recorder_upload.test` test is for some reason particularly much flakier on GitHub Actions than on my local machine.

Additionally, I think changing implementation is less risky than pivoting on the design/architecture of the system.

After some non-trivial modifications to `recorder.cpp` (fixing cyclical `shared_ptr` references, increasing `NodeHandle` lifetimes, etc.), I believe we have a working solution. EDIT: Please see https://github.com/aws-robotics/rosbag-uploader-ros1/pull/76. Lots of code and CMake...