Describe the bug
Using microRTPS agent with PX4 and a RaspberryPI4 connected via serial port The flushing operation at the start of the micrortps_agent might last forever if messages are received faster than the flushing operation (a pose of 1 ms has been set between each flush).
Serial bandwidth was set to 921600.
To Reproduce
Steps to reproduce the behavior:
microRTPS client started on PX4 (-b 921600)
microRTPS agent started on Raspberry PI 4 'micrortps_agent -b 921600 -d /dev/ttyS0'
It probably depends a lot on the hardware setup.
Expected behavior
The flush shouldn't loop forever. I have no idea of the best approach to solve the issue. With printing debug, I was able to be ensured that the code remains stuck in the flushing loop (see microRTPS_transport.cpp, from line 374). There is nothing to prevent flushing forever if you still receive messages at high rate.
**Additional context***
Messages exchanged in the RTPS bridge
Describe the bug Using microRTPS agent with PX4 and a RaspberryPI4 connected via serial port The flushing operation at the start of the micrortps_agent might last forever if messages are received faster than the flushing operation (a pose of 1 ms has been set between each flush). Serial bandwidth was set to 921600.
To Reproduce Steps to reproduce the behavior:
Expected behavior The flush shouldn't loop forever. I have no idea of the best approach to solve the issue. With printing debug, I was able to be ensured that the code remains stuck in the flushing loop (see microRTPS_transport.cpp, from line 374). There is nothing to prevent flushing forever if you still receive messages at high rate.
**Additional context*** Messages exchanged in the RTPS bridge