[1Wire] Fix not always showing 1-wire sensor & correct saving sensor - #5636
Conversation
|
And if you go then to another page and then load the task page again? |
|
maybe option value is the problem in ESP_Easy_mega_20260916_display_B_ESP32_4M316k changed to in Firmware ESP_Easy_mega_20240822_normal_ESP32_4M316k_LittleFS_ETH
Firmware ESP_Easy_mega_20260916_display_B_ESP32_4M316k
|
than it is loading normal and displaying the address - that is what i explained all the time in firmware without that problem option value is numbered -1 to 0, 1, 2, 3, ... New one is option value is numbered -1 to 2954339766987260677, 2954314937781322897, 2954279688984724647, ... |
|
but in older firmware there is for first recognized sensor and then the hex value (28-ff-10-0c-34-16-04-41) in new one this hex value is twice there and not the number of recognized sensor 0 (first), 1 (second), 2 (third), ... |
|
Hmm looking at this list you posted: I thought the order would be guaranteed to be the same, but I can't see any logical ordering in these numbers. At least not in the order they are posted here. |
Yep that's what I changed, the smaller values (0...x) were indices of the order the sensor was found. Now I add a 64-bit uint representation of that address as value, so I can literally decode the intended address from the value and there is no chance a new scan will result in a different ordered set of scan results. |
|
### response on ESP8266 ### on ESP32 |
|
Not 100% sure what you try to tell me with that HTML. Is the task even enabled? (or has been running for a while to collect at least 1 measurement) |
|
this part is missing in (browser) response at esp32 that is the difference to esp8266 the task is enabled at esp32 - esp8266 running a few minutes, esp32 running more than 1 hour after updating from 20260915 firmware |
|
Uptime: | 05h55m Task edit pressing submit reloading task by paste http://192.168.10.185/devices?index=14&page=1 to browser gives the addresses backsame thing Google Chrome, Firefox or Microsoft Edge
|
|
OK, I think this might be a thing that has unintentionally become timing-critical as the bus is now not scanned when the page is reloaded after submitting. When submitting a change on a task, the task is stopped and restarted again. I will have a look at it. |
|
yes, but resolution or error state can be changed and also only pressing submit clears device address reloading page did not change this but loading device and then this task again or filling task-address in browser address bar an load -> work this behavior is only on my unit´s esp32 - esp8266 work normal |
OK, this remark appears to be the actual clue to where the bug very likely is. The stored 1-wire addresses are in the It seems like there is a bug in this caching for storing exactly those values as used for the 1-wire sensor addresses. So will keep you posted and please keep making those 'casual remarks' as they may seem unimportant, but they can also be just the clue I need to find the bug :) |
|
sorry, for this late discovery |
|
Not totally happy with the fix, but at least on my test system it now works... |






Fixes: #5309