New Performance Checkup App errors

[RKBoss6]

Affected hardware version

Bangle 2

Your firmware version

2v29.32

The bug

I stumbled across the new Performance checkup app today, and tried it out. However, I after checking storage times, the first step, the app loader threw an error when trying to check clock fastload support:

And would not continue with the check. Might be random, but `hum` is an object in the `weather` app's data, which this might be referring to as a long shot, but I'm really not sure what's going on, and am keen to actually use this handy app and see performance bottlenecks on my watch :)

Installed apps

taglaunch (0.09), messageicons (0.17), messagegui (0.92), notify (0.14), clock_info (0.27), messages (0.70), mylocation (0.11), smpltmr (0.13), weather (0.31), widbtstates (0.03), wid_a_battery_widget (0.05), locale (0.20), clkinfomsg (0.01), loadingscreen (0.01), themes (0.01), calculator (0.08), fileman (0.04), f9lander (0.05), stacker (0.03), stlap (0.04), imc (0.01), inspire (0.03), ultrainfo (0.02), trex (0.05), boot (0.69), ios (0.25), clkinfosunrise (0.06), kineticscroll (0.05), autoreset (0.02), qmsched (0.10), worldclkinfo (0.02), meridian (0.02), clockbg (0.10), sleeplog (0.26), myprofile (0.03), calories (0.01), health (0.42), kbmulti (0.09), compass (0.09), clkinfostopw (0.03), alarm (0.53), reply (0.03), sched (0.38), sleepsummary (0.01), widlock (0.09), widmessages (0.08), setting (0.84), pebblepp (0.15)

[thyttan]

Does it persist if you update to latest cutting edge firmware?

[RKBoss6]

Yes, it does, but I now suspect this is related to my iOS companion app sending data while the transfer is going on 🤦‍♂️

I'll keep you all posted, does it work for everyone else in the meantime?

[RKBoss6]

Alright, I looked into this and it looks like anything the watch prints to the console gets intercepted by the app while its running, and causes it to not work as expected. For me in particular, the main reason is the prints of

 ____                 _
|  __|___ ___ ___ _ _|_|___ ___
|  __|_ -| . |  _| | | |   | . |
|____|___|  _|_| |___|_|_|_|___|
         |_| espruino.com
 2v29.76 (c) 2026 G.Williams

Which causes the json parsing to fail midway through (in the test clock loading times)... Any way we can disable all these prints while the app starts and then restores them on finish?

[gfwilliams]

That's odd - there's no way to disable the prints (that happens when the watch is reset), but it should usually deal with it fine. Does it always happen for you?

And is it always the "hum":"5" error you get?

I know your watch in particular seems to take a long time to boot, and I wonder whether it's just taking so long to load that the text that's written at boot is getting interpreted as the result of the command that's sent.

I wonder whether for code like this:

https://github.com/espruino/BangleApps/blob/master/apps/performance/custom.html#L51-L52

We need a ~1s delay between the load() command being completed and then trying to evaluate the variable? Do you think you'd be able to try that and see if it fixes at least that step?