Ticket Created
over 5 years ago

WERETECH-8553

fenix 6 series firmware 6.0 causes watch face to crash

Hi, fenix 6 series users are reporting a crash while using the MagicDust watch face. (https://apps.garmin.com/en-US/apps/2851d38b-22a8-41b6-84bf-2b0d4a229869)

The crash report leads to the code logic where ActivityMonitor.getInfo().floorsClimbed returns an invalid value.

---
Error: Invalid Value
Details: 'Failed invoking <symbol>'
Firmware-Version: '6.00'
Language-Code: eng
ConnectIQ-Version: 3.1.7
Filename: A1VE3325 <- strange file name!?
Appname: MagicDust
---

Repro steps: just install the watch face, and trigger the high power mode with a rotation of your wrist. This seems to cause the crash.

I don't have a fenix 6 device, testing in simulator works perfectly.

On my fenix 5x also everything is fine.

Thanks, for your support.

Please verify this issue, thanks.

TheMagician

TheMagician
  • over 5 years ago in reply to TheMagician

    Check the FW version in the ERA crashes, as there was/is a bug where things are getting installed on devices with old FW that can cause some odd crashes.

  • over 5 years ago in reply to jim_m_58

    Null check and division by 0 is checked.

    Thanks for the tip with ERA (nice tool!)

    But the reported crashes are really strange e.g.:

    .onPowerBudgetExceeded: Too Many Arguments Error / Symbol not found

    .onUpdate (line where an arc is drawn with dc.drawArc(140, 140, 140, 1, 61, 30)): Too Many Arguments Error / Symbol not found

    wtf ;-)

    is this right? -> the Backtrace in ERA shows the line number of the crash?

    Do have any additional tip how to proceed with this issues?

  • over 5 years ago in reply to TheMagician

    One think to check is seldom used code paths, where you might not be seeing other reports of this, as it's something that's not common - a couple of examples is turning off activity tracking on the watch, or in this case, have a floors goal of 0.  I ran into the floors goal=0 maybe 6 months back, in a watchface that had been available for years.

    Also check the ERA logs, as that might give you a few hints, as to how many people are seeing this and on what devices, as not everyone that has a problem reports it.

  • over 5 years ago in reply to jim_m_58

    Hi, the stack trace location of the error is not listed in the debug.xml. So i used the next matching lines in the debug.xml to get the code location.

    The strange thing is that the lines of code are running stable since at least one year on several devices.

    It could also an issue of the new 3.1.7 SDK.

  • over 5 years ago

    When downloaded from the store, apps get assigned a unique file name, so that's normal.

    What are you seeing in the stack trace as far as the location of the error?  When installing from the store there are no debug symbols, so you need to use the corresponding debug.xml (it's in the .iq file) to translate the PC values.