Closed nashant closed 3 years ago
Hold up, not quite ready I've realised
Ok, now it's ready
Hey nashant. I tried to build your fork but when I connect the core is panic'ing and the ESP32 restarts. Andriensis also added some bugfixes so I created branch 0.2 so we can work there together on the next version. What I could see from the new design and changes it looks really awesome. Thanks for adding it.
I think I might be leaking memory somewhere. When reading pointers from the json doc the data is corrupted and the memory addresses are, I think, fairly close to the end of the address space. I think. But again, this is the first c++ I've ever touched.
On Wed, 25 Aug 2021, 21:09 HeimdallMidgard, @.***> wrote:
Hey nashant. I tried to build your fork but when I connect the core is panic'ing and the ESP32 restarts. Andriensis also added some bugfixes so I created branch 0.2 so we can work there together on the next version. What I could see from the new design and changes it looks really awesome. Thanks for adding it.
— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/HeimdallMidgard/ESP-32-BLE-Scanner/pull/7#issuecomment-905838292, or unsubscribe https://github.com/notifications/unsubscribe-auth/AATNTOF34KQJOS5UMTRIOJLT6VEV3ANCNFSM5CZALCRQ . Triage notifications on the go with GitHub Mobile for iOS https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675 or Android https://play.google.com/store/apps/details?id=com.github.android&utm_campaign=notification-email .
Issue was the delay. I increased the delay time. Issue is gone now. I also changed some little things after reading your code updates. I would like to work in 0.2 branch and close this pull request so we can work together on it. Is this okay for you?
Absolutely, sounds like a great idea. I found increasing the interval and window sorted things out. I think that by the time the devices were being processed they had already been replaced in the scan (with duplicates set to be filtered out).
Just done another large push splitting things out and having a better way of initialising settings. Still need to sort out migration of devices a bit more.
Tbh I don't think this needs migration. Settings can just be done new. I don't think so many ppl are using this right now. Main problem I see at the moment is that the scanner doesn't work (didnt check your update right now). Can you please push updates inside the 0.2 than I don't need to do this handy. Thanks.
I've re-targeted to 0.2, feel free to merge into 0.2 and test
Adding the telemetry has been useful. Using a global mqtt_msg
is leaking memory. It takes ~20mins for the heap to drop to ~50k free. I've switched it to using a local scoped char instead and the heap is holding strong at ~150k. Need to solve the problem of reporting to the webpage though, I'm thinking maybe an async websocket
Currently { "results_last_scan": "3", "free_heap": "147216", "uptime": "29243" }
Websocket done
settings.json doesn't seem to contain bluetooth key for bluetooth settings (scan_time, scan_interval)
Doesn't need to. If you check all the initSettingsBlah
functions you'll see how they get created. In fact you don't need settings.json
or devices.json
at all, they'll both get initialised on first boot. Also means that one doesn't need to re-flash the data if new settings are added
At the moment the scanner function is broken. Core is panicing.
Mostly changes based around UI.
Adds MiniCSS for css and AlpineJS as a minimal javascript framework, new devices and settings pages. New settings format to allow for this and easy extension to settings, migrations from old settings format included. Allow adding of unlimited devices.