Vivoactive 3 firmware 6.90 cause watch faces to crash

Knowing that two of mij widget crashed and having tried al solutions I could think of (including generating a new app-id), I first tried if an empty widget would work. I made a new project with another name in Eclipse and verified that id did not crash (and it didn't). Next i copied and pasted entire functions from the original sourcecode to the new project.(renamed classes as i went) and I copied resourcefiles.  In between I verified that it still worked (which by the way once did not when i copied an entire sourcefile to the new project). Of course I hoped that on the way I would discover which part of the source caused the crash, but none did.I ended up with a working widget that was almost identical to the original one. Only difference is I did not copy some unused resources and I used some other names for classes (and the entrypoint in the manifest). I did the same with another widget and got it working that way. Weird ... . Can't think of an explanation.

Knowing that two of mij widget crashed and having tried al solutions I could think of (including generating a new app-id), I first tried if an empty widget would work. I made a new project with another name in Eclipse and verified that id did not crash (and it didn't). Next i copied and pasted entire functions from the original sourcecode to the new project.(renamed classes as i went) and I copied resourcefiles.  In between I verified that it still worked (which by the way once did not when i copied an entire sourcefile to the new project). Of course I hoped that on the way I would discover which part of the source caused the crash, but none did.I ended up with a working widget that was almost identical to the original one. Only difference is I did not copy some unused resources and I used some other names for classes (and the entrypoint in the manifest). I did the same with another widget and got it working that way. Weird ... . Can't think of an explanation.

I know it makes no sense. Originally it was ment as a method for isolating the causing program-lines but it worked out the unsatisfying way I described.

Upgrade of the firmware was in my case the visible and instant cause of the crashes. Might of course be that the cause lies in the compiler and surfaces because of the firmware-upgrade but my guess is not fwiw.

You're right, the difference is that this time is just VA3, so it's probably an entirely different issue. But why reentering code helped?

It was exactly like that. Now that you say it, I remember also loading the source file up in n++ and saw no difference in line endings. At that time I always "resolved" the issue by going back in git history and writing the code in another way (e.g. on different lines as that actually would have mattered) so it wouldn't crash.

Hey guys, that makes no sense. We have to look at the final exported project. Only on the vivoactive 3 series the app/watch face doesn't work. All other supported devices are ok. I think Garmin have to look at the firmware or the compiler.

Was that also a silent crash with no entry in the logfile? I just loaded the crashing and non-crashing files in notepad+++ and saw no difference in line-endings