After a month of relentless troubleshooting, we have finally managed to pinpoint the cause and implement a fix for crashes that affected some Android users while trying to open the game! If you were affected by this problem, be sure to download the newest build of the game from our official website! We have uploaded a patched versions of both patron-only release (Bo, Part One) and public build (Hanako, Part Two).
If you're curious about specifics, we will provide a detailed breakdown of the exact reason behind the crashes down below. Additionally, we would like to thank everyone who assisted us by providing detailed reports and information about their devices, and especially the user who shared their successful attempt at solving the crashes on their end by changing renderer settings!
Why did it crash for some players, but not others?
As it turns out, there were two separate issues that led to crashes and/or freezes on startup. The first one was obvious - phones sometimes ran out of RAM while the game attempted to load the assets. Therefore, we split the asset preloading into smaller chunks. After doing internal testing, the performance impact during gameplay was negligible, but there is a possibility users on older or less powerful devices could experience slowdown or freezes while loading the gallery. Since we are already working on updating the gallery as a whole, this issue is temporary and should be resolved once we implement new scene sorting system.
The second problem, which affected users seemingly at random regardless of how old or new their phone was, turned out to be caused by RenPy's internal renderer settings. One of the custom shaders we used for blurring background images during dialogues utilized a different renderer as opposed to the one we use for scene pictures or character sprites. Unfortunately, said renderer isn't supported by certain mobile GPUs, which led to crashes. And even more unfortunately, these crashes didn't leave any crash reports in the game's folder and didn't affect any of our own devices we used for internal testing. We went over dozen phones, some nearly decade old, and others basically modern flagships, but none had experienced same problems which made it significantly harder to test potential fixes. Luckily, one of the users who encountered a similar issue had found a workaround by manually switching renderer setting on their device, and after further investigation on our end, we have confirmed that this was indeed the cause behind the crashes. Funnily enough, the actual fix afterwards took us only a couple of minutes.
Again, big thank you to everyone who took part in this bug-fixing adventure! Without your input, it would've taken us at least another month of scratching our head and screaming into the void.
we have finally managed to pinpoint the cause and implement a fix for crashes that affected somewe have finally managed to pinpoint the cause and implement a fix for crashes that affected somewe have finally managed to pinpoint the cause and implement a fix for crashes that affected some