Stephen Early

@solick No, it's nothing to do with local socket addresses. By default Mastodon won't connect out to addresses it considers to be "private". In IPv4 world this is ranges like...

@ClearlyClaire may have a different opinion, but I think it is. Although I brought up the issue in the context of an IPv6-only system connecting to systems with both IPv4...

Yes, that should do it. (You can see it as part of the [Commissioning sequence](https://github.com/sde1000/python-dali/blob/6923d79cbb1c98f3503da006bff278e1cd673e77/dali/sequences.py#L175).)

Could you describe your setup in more detail?

That should do it. The two sends of 0x7f 0x80 must occur within 100ms of each other, with nothing else being sent in between — are you managing this?

They look correct, and 15ms between the two repetitions looks normal. What do you get for `QueryControlGearPresent` on the short address? And if you broadcast `QueryMissingShortAddress` do you get a...

Hang on, I might be being stupid. `SetShortAddress(Broadcast())` should be 0xff 0x80 not 0x7f 0x80! 0x7f 0x80 is `SetShortAddress(GearShort(63))`. So you should look into how the wrong frame is being...

It sounds like you might have both units set to the same short address? Or there might be some other hardware issue interfering with communication. What happens if you disconnect...

Oh! Probably worth opening a new issue for that and mentioning @sl-wallace — he wrote the LUBA driver. (I don't have one so can't test it.)

More generally, fixing mistakes once a delivery is confirmed is still needed. Perhaps introduce a delay between delivery confirmation and sending to the accounting system? Maybe send in a batch...