3.5b1 Reset reason: StuckInSpinLoop?

[dc42]

@Diamondback I suspect it's an interaction between DWC polling the object model (which involves reading some values from the heap) and garbage collection of the heap taking place because string or array variables are being reassigned.

[Diamondback]

@dc42 I played around with the DWC polling time this morning but didn't get any immediate results, I'll play around with it some more and see if I can maybe make it go away by setting it much higher?

[dc42]

@Diamondback my next step will be to disconnect DWC completely and then see if I can provoke the reset by sending the command to run the macro over USB.

[Diamondback]

@dc42 Hm, I guess I can try the same by having the issue triggered by daemon.g and not having any browser connected to it? Or does DWC still do stuff in the background then somehow?

[dc42]

Confirmed, if I send the command over USB with DWC not connected, it doesn't reset. If I connect DWC and send the command over USB, it resets.

[dc42]

@Diamondback please try the firmware at https://www.dropbox.com/sh/z9f9v517t9txqh9/AACG2Dycr_EghlN4p1HsYVUNa?dl=0. It fixes the case that I was able to reproduce. The change is quite extensive, so please use it with caution.

[Diamondback]

@dc42 Will do! This includes the previous array memory leak fix?

[dc42]

@Diamondback yes it does. If you get any resets, please post a subsequent M122 report.

[Diamondback]

@dc42 Looking promising so far, no more resets atm.
I also yesterday played around more with the DWC polling frequency between 50ms and 60s and as you assumed, with lower values it was happening much quicker.

[Diamondback]

@dc42 Finally working
https://www.dropbox.com/s/fyguad75k8g2ffe/VID_20230124_102416.mp4?dl=0

[Diamondback]

@dc42 happy to report, no further resets over the day Many thanks!