I didn't think resetting the old resolution like that would fix it. My original plan was to schedule a reset by setting Option<Scheduler> there but I'll test this later today and if it works out great I'll merge it.

Thanks 🙏

Yeah this isn't working, it's not downsampling back to normal resolution but I think I know of a solution now.

No matter what I try to do something else breaks so yeah this logic is just aids and will probably require me to redo the dynamic sampling logic again. I'm pretty sure if I don't and we just hack it quick, another weird breakage in the logic will pop up.

I'll return to solve this on Saturday.

Okay I got a better grasp of the issue. I don't expect anyone to understand this yap below, I just need to post this for my own self documentation and sanity.

The actual issue isn't that the image doesn't up-sample a second time. The real issue is that when you zoom out to original zoom factor (or hit the reset button R) it tends to schedule a resize / dynamic sample for a much higher resolution instead of the original res that the image should be when zoom factor is at original scale (1.0). Basically my logic is fucked broken.

Now sometimes the "much higher resolution" is set exactly the same as the last dynamic sampling resolution (like in that image above) and we have logic that disallows resizing when that happens. So when we zoom in the second time it get's confused and shit starts acting sussy wussy.

The "second up-sample" issue is a side-effect of the real issue, being Roseate scheduling the incorrect resolution because of if zoom_pan.zoom_factor <= 1.0 { prohibiting dynamic_sampling_update from scheduling anymore resolutions when the zoom factor reaches 1.0.

annddd this probably makes zero sense to you does it...
fuck it let me stop the yap and just work out a fix for this already