Is anybody developing a MeshCore-Node in near Future?
To deal with the "exec"-Command is quite painful...
Forgive my ignorance, but what would a MeshCore-Node do?
Can you expand on this, with an example showing where it hurts?
For example, on a Raspberry Pi:
To send Messages over meshcore, you have to install the meshcore-cli (GitHub - meshcore-dev/meshcore-cli: Command line interface to MeshCore node · GitHub) and send the meshcore-cli-commands over an exec-Node.
E.g.: to send a message, exec-Code will be:
/home/pi/.local/bin/meshcli -s /dev/ttyACM2 chan 1 "yourtext1"
Your post is confusing
Are you having a problem with the "exec"-Command, as in it doesn't work correctly and you need some assistance?
Or are you saying that using it is not the easiest thing in the world. If so, in what way?
I am not sure where I stand on having lots of special purpose nodes.
For example there is (probably) an Openweathermap node. Or there is a general purpose HTML-request node.
A special purpose node might avoid the need to install any code outside of Node-red (eg meshcli), so I see the value there.
Anyway, the exec approach:
It can be a bit of a faff to construct the input to the exec node, but it should be possible to do it just once and use link nodes to call the exec.
I don't have meshcli so this example is speculative.
And to call it (these need not be on the same editor tab)
Assuming that the only variable in the command is "yourtext", the exec node might be like this
And if eg "chan 1" is also variable, it's not too hard to compose the entire command line in a template node to pass to exec.
Thank you, jbudd, for your explanations,
this is the point where I'm right now in my project.
And because to deal with the exec-Node (at least in my scenario) is much more fiddling than with an dedicated node (like the openweathermap-node, for instance) I've made my request: to make the communication to MeshCore-Nodes easier in order to get rid of these exec-nodes.
Perhaps you did not take on board that my example gets rid of all of the exec nodes except one, which you can call from any flow, passing the intended notification as msg.payload?
I do agree with you that minimising use of exec nodes results in a cleaner Node-red setup, the above suggestion was just a work-around in the absence of a specialised node..
Incidentally, is /home/pi/.local/bin/meshcli a script, bash or python maybe?
