Shawn Silverman

Please split unrelated changes into different PRs.

So currently, none of the examples will validate. Here's why (excerpt from the new "Notes on the examples" section in the README): --- ### Manufacturer ID zero All the example...

I’ll add: I don’t love that none of the examples validate. Maybe a workaround is to temporarily change the minimum allowed manufacturer ID to zero during the validation step. But...

I'm familiar with squash merges, I just wanted to give you an opportunity to create your own commit message. If you don't want to, I'll just create a commit and...

See: https://github.com/ssilverman/rdm-schema/pull/6#issuecomment-888343344

If someone wants to write a schema with "get_request_subdevice_range" and "set_request_subdevice_range", but without corresponding "get_request" and "set_request", that's up to them. These are fields that effectively annotate "get_request" and "set_request",...

Another way of putting it: My principle here is that annotations shouldn't require the thing they're annotating, at least in the schema validation level.

I’m with you on this one. I think the range should be the entire 0-0xffff, especially since it’s defined as a 16-bit value only. I should probably change it back...

I don’t think of those dynamic IDs as just a bit being set. I think of them as being >= 0x8000. That’s why I included them.

> I also meant to say, on a more positive front, I think there would be some ways to provide both bits of functionality with subschemas etc, so there could...