We've had now, in our testing, three cases where a screen has gotten somewhat garbled, all non-reproducible (of course), and all similar. I'm wondering if anyone else has seen the same thing.
The three cases occurred in the same application, along two quite different paths, though there are some similarities. In all cases what happened is that, on a complex screen (with lots going on), part of the background did not get refreshed properly, with the result that the screen "behind" the current one "shows through". Specifically, in all cases, a background image (basically a white box with rounded corners) that covered about 80% of the background apparently did not get drawn.
Like I said, two cases occurred along one path, in essentially the same place both times, and the other occurred in an entirely different place in the code -- logically, temporally, and with regard to who wrote the code and what language facilities might have been employed. (Possibly, however, the actual background image was the same in both scenarios -- I'll have to check into that.)
One curious thing is that for one of these scenarios it is possible to back all the way out of that path of code, such that the widgets for the garbled screens should have been destroyed, but on reentry the screens are still garbled. (This might, I suppose, suggest that access to the background image file had been lost for some reason, but how do you "lose" the ability to access a PNG in a resource file?)
Of course, exiting and reentering the application resets everything.
So has anyone seen anything like this? Any ideas?
[Pre-posting update. Writing this made me think, and on the phone I have in this failure mode right now I checked several other screens that uses the same background image -- they too are failing, though the effect isn't nearly as spectacular. So it appears that a PNG image in a resource file has become inaccessible for some reason, and without obviously impacting any other "nearby" images. Any ideas?]