Memory full, fifo error

I keep getting this errors (see pictures) about memory full or fifo.
I just did factory reset but it keeps coming back.
Black screen is before and white after factory reset.
Please help.

2 Attachments

If you have done a factory reset you will have reverted to an older versions of the Anton clock.

A factory reset probably reverts to the light theme. You would need to change it back to the dark theme.

Also a good idea to upgrade to the latest firmware v2. 13 if you can. Then make sure you ate using the latest versions of the apps.

I would also suggest running a few days using a different clock in case the problem is a bug in the click app

Finally what widgets are you using?

What other apps did you install after the factory reset? I can see you still have the pedometer after the factory reset (which shouldn't normally be there) so I guess you may have added some other things too?

Also I see Bluetooth is currently connected on the Bangle. What app/device are you connected to?

If something sends a lot of data over bluetooth (especially if it ignores flow control) then that can cause the FIFO_FULL message, and depending on what it sends that can also cause out of memory errors too.

I also get this quite regularly now. My wife too, but much less often.
Anything I can do to help find the cause before doing a factory reset?
I am on 2.12.126, she is on 2.13 which are the same thing if I remember right.
I guess its about RAM?! How can I get Stats about that?

>require("Storage").getStats()
={ totalBytes: 8388608, freeBytes: 7335192, fileBytes: 499652, fileCount: 133,
  trashBytes: 553640, trashCount: 310 }

Current state:

• clock face is shown only partly (sometimes completely black screen) and doesn't change anymore
  
• Longpress (both reboot or just to watch-face) removes problem for some hours (time is shown correctly then, so something was still working in the background).
  
• Repeat
  
  

{Update] Don't know if that helps:
Got into that situation while connected in WebIDE. Saw whole screen full of

Execution Interrupted during event processing.
Execution Interrupted
Execution Interrupted during event processing.
Execution Interrupted
Execution Interrupted during event processing.
Execution Interrupted during event processing.
Execution Interrupted
Button held down - interrupt JS execution...
Button held down - interrupt JS execution...
Execution Interrupted during event processing.
Execution Interrupted during event processing.
>

1 Attachment

Please can you run process.memory(1).usage in the Web IDE after loading your watch face, and do that a few times with some minutes between calls.

My guess is that the watch face you have (or any widgets - although I can't see any) may have a memory leak

Thanks a lot for that command @Gordon, enabled me to find the cause.
Showed a memory leak clearly with number going up by ca 300/min.
After deleting sleeplog and sleepphasealarm number keeps constant now.
So sorry to say @storm64, but I am pretty sure its one of those two causing the problem.

[Update] Did this test with current dev versions 0.04 (log) and 0.07 (alarm)

Great, thanks! Definitely good to know. Hopefully we'll be able to find out what the issue is with whichever one of those soon.

The last time I tracked down a memory leak, this tool was very helpful.

sleepphasealarm does not have any widget or bootloader hook that runs in the background.

Much much better, no problem the whole day.
But: as said above, my wife had the same problem but much less frequent. She does not have the sleep apps installed. So I checked back on memory on my watch (no apps used the whole day, only looked for time). It still looses memory, but now with approx 3/min (instead of 300 before).
Will continue to check (maybe this is expected behavior from the health app collecting data or something like this?)

maybe this is expected behavior from the health app collecting data or something like this?

No, it's definitely not... Maybe the first few minutes it might lose something, but after a few minutes no new memory should be getting used.

If you want to try and debug it yourself, you can do E.dumpVariables() (you'll have to enable logging via the IDE as it dumps a lot of stuff) then wait until the free memory goes down by a chunk, run it again, and 'diff' the two sets of data you get.

It might give you some clue (based on variable names) as to what's using the memory