python3 -m venv venv
source ./venv/bin/activate
pip3 install catkin_pkg lark-parser empy colcon-common-extensions
git clone --recursive https://github.com/micro-ROS/micro_ros_azure_rtos_app
cd micro_ros_azure_rtos_app
cmake -Bbuild -GNinja
cmake --build build # run this three times
Expected behavior
Successful build/no errors
Actual behavior
The first two build runs produce errors, but there are fixed by subsequent runs as the required byproducts are produced.
The third and all further runs result in the errors:
[...]
micro_ros_azure_rtos_app/app/microros_netx_transports.c:1:10: fatal error: uxr/client/transport.h: No such file or directory
1 | #include <uxr/client/transport.h>
| ^~~~~~~~~~~~~~~~~~~~~~~~
compilation terminated.
[...]
micro_ros_azure_rtos_app/app/main.c:9:10: fatal error: rclc/rclc.h: No such file or directory
9 | #include <rclc/rclc.h>
| ^~~~~~~~~~~~~
compilation terminated.
(Build command output has been removed for brevity)
Additional information
The repository has been successfully cloned. The two files rclc.h and transport.h can be found in their respective paths on disk.
The same error occurs with the previous commit (f83776e).
I was also a little suspect of Ninja, but compiling with cmake -Bbuild -G "Unix Makefiles" yields the same result.
This seems to either be some sort of build system issue or an environment problem.
Perhaps the readme is missing some implicit steps I'm unaware of?
Steps to reproduce the issue
Compile as desribed in the redame:
Expected behavior
Successful build/no errors
Actual behavior
The first two build runs produce errors, but there are fixed by subsequent runs as the required byproducts are produced. The third and all further runs result in the errors:
(Build command output has been removed for brevity)
Additional information
The repository has been successfully cloned. The two files
rclc.h
andtransport.h
can be found in their respective paths on disk. The same error occurs with the previous commit (f83776e). I was also a little suspect of Ninja, but compiling withcmake -Bbuild -G "Unix Makefiles"
yields the same result.This seems to either be some sort of build system issue or an environment problem. Perhaps the readme is missing some implicit steps I'm unaware of?