Store all signal and improve structuce of signal line about c3

[nwittler]

I think there should be some discussion whether we want to store all of this in the object. We already do this in a number of places, the propagation too for example, and I suspect this bloats the runtime environment. Better practice might be to get good logging in general. And even then, storing absolutely everything is a debug situation and not required in production, or is it?

[alex-simm]

Looks to me like there are two separate things:

• allow obtaining the signal in between devices in the line: I'm guessing this would be mainly for debugging. For this, a callback that is called after each device might be the best option. It doesn't create any overhead and allows people to retrieve and store the signals if needed.
• improving the signal line structure to something more tree-like: instead of passing a large 'chains' dictionary to the generator, I would propose connecting the devices in an object-oriented way. Each device can store its successor (or predecessors) and the generator only needs to iterate through that tree.

[shaimach]

My two cents: - Devices do not need to know what they are connected to (store nodes and edges separately). - The network topology is a DAG (directed acyclic graph) (and if we ever encounter a cycle in real life - we simply run away as fast as we can) Shai

…

On Wed, 14 Jul 2021 at 15:36, Alex ***@***.***> wrote: Looks to me like there are two separate things: - allow obtaining the signal in between devices in the line: I'm guessing this would be mainly for debugging. For this, a callback that is called after each device might be the best option. It doesn't create any overhead and allows people to retrieve and store the signals if needed. - improving the signal line structure to something more tree-like: instead of passing a large 'chains' dictionary to the generator, I would propose connecting the devices in an object-oriented way. Each device can store its successor (or predecessors) and the generator only needs to iterate through that tree. — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#29 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAH3Q4INYEFRFPQXXI7TGWDTXWAFBANCNFSM4WY7RSLA> .