Forum Two light switches for the same light. how to install relay swi…

Thanks for the reply — and for the concerns; they're the right ones to raise. - On the rewrite. plua's 400+ commits weren't thrown away. The QA-side Lua — quickapp, fibaro, the API behavior — has survived every version largely intact; what I've been re-drawing is the line between Lua and Python. hc3emu was 100% Lua, flua is ~80% Python. Each version answered one question: where does the complexity stop paying rent? plua answered that it keeps growing — Python processes, asyncio, libraries, cross-platform glue — and at some point that complexity became the ceiling for the very thing I care about: clear code that is enjoyable to work with... I had also built up the realization that it could be done simpler... flua is the cleanest model I've made, and the first one where I can see the end of the problem. Ive added more and more of Plua's features the last days and it still keeps within bounds. - On AI-generated code. You're describing a real failure mode, and I agree it's the biggest threat to the "small and simple" goal. My answer is that the discipline has to live in the repo, not in my mood: the architecture is documented, AGENTS.md tells any agent — from any framework — what belongs in flua and what doesn't, every behavior change lands with a test, and changes are reviewed commit by commit. I don't hand an agent an end goal and walk away; I describe the mechanism I want, in detail, and I understand all of the resulting code. The Lua side is ~90% hand-written, the Python ~50%

I really appreciate your work. Thanks to these emulators, I can work faster and with much less effort. I would not have been able to write them myself. I do, however, have one concern that I would like to share. This is already the third or fourth version of the emulator, and my question is: is it really worth writing it again from scratch? In your post, you basically explain why. I understand the motivation. On the other hand, looking at the previous version, there was quite a long period of development behind it, getting closer to HC3 compatibility and stability (>400 commits). I also have some concerns about the idea of developing flua with a large contribution from AI agents. Aren’t you worried that, over time, the code may become more general, defensive and abstract than the specific problem actually requires, even if it still works correctly? That is a well-known weakness of AI-generated code. One person will describe the problem very precisely, check the result, and question unnecessary mechanisms. Another person, after finding a bug, will simply tell the AI: “fix it” and submit a pull request. With many different users, different AI models, and different levels of control over the generated code, there is a risk that a project intended to stay small and simple will gradually start growing and becoming complex again. And then there is your role as the reviewer - I don’t think you are going to get bored. I have no practical experience with this, but I suppose that is th

Plua, the development tool for QuickApps, has evolved over time and has an internal architecture that has become increasingly complex over time. So, let me introduce flua — it's a simplified plua with an agent friendly architecture... Meaning that it's easy to help developing flua; find bugs or HC3 non-compliance, let your agent fix it and create a pull request and I add it to flua. Flua is a Lua engine for Fibaro HC3 QuickApps that runs on your Mac or PC. You write the same Lua you'd write on the HC3 — QuickApp, fibaro, api, net, mqtt — and flua provides the rest of the world. The same QA file runs offline in the emulator, online against a real HC3, or in proxy mode where a mirror QA on the HC3 renders the real UI and funnels every interaction back to your emulated QA. For QA developers The HC3 editor is tiny, and debugging there means print() statements. flua exists so you can develop in your real editor with real tooling, and upload when it works: Simulated HC3 — devices, global variables, scenes, alarms, profiles, refreshStates, multiple isolated QAs talking to each other. Virtual clock — run at 60x, jump instantly to timers, or let simulated time drive long-running logic. fibaro.sleep suspends only the calling QA. Real networking — HTTP, TCP, UDP, WebSocket and MQTT clients work against your real services. A proper debugger — breakpoints and stepping in VS Code (Lua MobDebug). UI development — describe your UI with --%%u: directives; the viewer renders it, and proxy mode

Any changes done in LAN settings? Please post the line 2231

Thanks for sharing this, i will soon test your app on a android tablet. I see some medium risk results from the assesment (like M1). You might want to look into that, since i think it aint minor risk's.

Hi @cag014, My HC2 on 4.630 AOS 11.0 got the fatal error beneath and hanged :

I really appreciate everyone's time in responding and helping me to find a resolution here. I've read and attempted all of your suggestions. It has been a while since I posted this, but life's hectic and I'd been trying to keep up with juggling life and trying to sort this issue out, however over a couple of updates, the issue seems to have sorted itself out. I have had software issues since installation and all kinds of crashes etc since moving into our new home just over 4 years ago. I could not understand how it could be a scene or quickapp issue as I've gone through everything that I can possibly think of. As I said, everything works as it should now.....no more crashes or needing to reboot everyday. Fibaro must have corrected their faulty software. I still have Z-wave device issues though. Roller blinds and other devices just randomly lose connection and now I need to delete those devices and re-add them, which means going through every scene, every users devices permissions etc. I love and hate this technology.

@ppeterr thank you, Dependabot alerts dont see a problems, below report for app where i will keep http or not blocking widgets HubControl-Security-Assessment-0.4.0.pdf

Hello Tedee Plugin is not working today. Seems to appear just after the last bridge update (v 2.2.48018). It sais I should request a new key, but even with this new key, plug stay stuck with the message : Unknown error. Generate a new authorization code in the „Advanced” tab. Any idea ?

Great initiative! I think any form of dashboard represents a fantastic step forward. So, kudos on your first post here on the forum-- and a warm welcome. It might be a good idea to simply share the code and source files so that people with some expertise can review them. I’m very interested myself, but I wouldn't just install AI compiled code without it having been properly vetted first. Again-- welcome, great first post!

EU is very active if we talk about smarthome devices. 1) RED 18301 https://single-market-economy.ec.europa.eu/sectors/electrical-and-electronic-engineering-industries-eei/radio-equipment-directive-red_en where we added support for it in 5.180 like encryptions of data, abandon old standard like TLS 1.0 and 1.1, min password, complex of it, no more easy first paswword like PSTI for UK (goodbay admin, admin) and many more 2) Data act - possibile to made of copy what data what we keep in hub or cloud 3)The Cyber Resilience Act from this month option to monitor vulnerabilities and option to report them to us https://www.niceforyou.com/en/cra-policy as part 1, and part 2 from December 2027 with part 2. And this part 2 will be more problematic like no more sell of product who dont meet new CRA regulation, paid CE certification for some group of products, a lot of more documentation and proces of creating devices. Like Łukasz said under CRA will be all products, aplications services witch connect to smarthome hub (mobile aplication, Yubii id, end devices,

**[Beta testers wanted] HubControl – Android phone, tablet, Wear OS watch & Android Auto client for HC3 / Yubii Home** Hi everyone, I've been working on **HubControl**, an Android client for **Fibaro Home Center 3 and Yubii Home** (firmware 5.200 or newer). It's a dark-themed app meant for everyday use: see what's going on at home in a couple of seconds and control it from your phone, tablet, wrist or car. It's in **early alpha (0.1.0)**. I'd love your feedback, and I'm looking for people who'd like to test it on their own systems (details at the end). I use AI to create app --- **Platforms** - **Phone** – Android 15 or newer - **Tablet** – its own layout for large screens (navigation rail, dashboard, list + detail view) - **Home screen widgets** – home summary, favourites, a single device, a single scene, a list of scenes - **Wear OS watch** – Wear OS 5 or newer (tested with Galaxy Watch), including Tiles and a watch-face complication - **Android Auto** – runs on the car screen through the phone --- **What it can do** *Home at a glance* - Home status summary: lights on, blinds and gates open, open doors and windows, motion, locks, climate (temperature, humidity, light level), HVAC set-points, alarm zones, low batteries, dead devices - "To check" list: anything that needs your attention (unlocked locks, open windows, alarms, dead devices), sorted with the most urgent first - Energy panel: PV production vs. consumption, grid import and export, self-consumption, costs, consumpt