Request-response service best practice?
Background
- New to ROS 1/2. Working with ROS2. Playing with Python.
ROS2 is installed on PC and also cross-compiled and installed on embedded system.
- PC: Kubuntu 16.04, just followed ROS2 instructions - got Ardent, also installed colcon
- The embedded system is Elphel 393 camera system Xilinx Zynq (32bit ARM) based with GigE network. Yocto + meta-ros got ros2 recipes built fine.
Demos AddTwoInts and Talker-Listener work fine in pc-camera, camera-pc
There's a lighttpd webserver that is running on the camera and the parameters are normally changed with http requests like:
wget http://ip/somescript.php?p1=v1&p2=v2
The response is an xml status string with updated pars values:
<pars> <p1>v1</p1> <p2>v2</p2> </pars>
More Info (unrelated)
Question
Now I would like to implement a similar behavior (i.e. use a single service to send commands and get status replies as bools or preferably strings) in ROS2 - with a service. But after having a brief look into several repos (ros2_intel_realsense, axis_camera and inertial_sense_ros) - they don’t seem to use services and *.srv files a lot.
What did I miss? Is it implemented via parameters? What's the best practice?
Thanks.