Thomas Watson
Thomas Watson
It would also be good to know how to calculate the appropriate margin. Calling `lfs_fs_size` is relatively slow and it's unclear if every free block can be used for a...
> only one commit per library is allowed in final pr This is definitely not true and IMO reduces clarity.
As a big proponent of scripting I concur that I would like to see this remain a script instead of the overhead of flash and maintenance of C++ code. I...
The flight controller handles the protocol, the periph just sends and receives bytes. So I don't think the protocol should change, not sure if 255 is the best value to...
What would the protocol be used for? The client should not interpret it.
In general, dronecan devices should use dronecan messages rather than tunneled serial data. If the device itself wanted to parse the data, it would have to document a particular `serial_id`...
My goal with this PR was to minimize changes to the code and just document things as they are. Currently the results are indeed manually copied, and hopefully the steps...
> i think we should try to use one version of sympy for all generation, and check it generates good code, then assert sympy version instead of using nix-shell Moved...
Found a version of Sympy (1.9) that generates code identical to the current for both scripts, so only one version needs to be installed. Both scripts check for that version...
Drafted until those two PRs get merged and we have final hashes for them.