So for some reason, at random times when using Retroarch, game focus mode will activate along with other bizarre changes to Retroarch’s functionality, like the FPS counter showing up, gamepad and mouse buttons doing different things from what they should, and often, the game being played gets reset because of one of said buttons being pressed (including the d-pad).
That’s a weird thing that a number of people have reported, but we haven’t been able to reproduce it, much less pin down the cause. What we know is: it’s Windows-only, you can make it happen reliably by killing explorer.exe, and it sometimes can be related to virtual gamepad software. I don’t think changing your input/gamepad drivers will fix it, but you could try it.
Alright, I’ll try using the SDL2 gamepad driver instead of Xinput and see if that stops it from happening. On the topic of “virtual gamepad software”, does AntiMicro count as a possible culprit? Because I use that with RetroArch to map keyboard shortcuts to the gamepad without having to mess with RetroArch’s config.
It’s possible, sure! I have antimicro running on my Windows box, but I don’t use RetroArch on it much, so if it’s causing problems, I haven’t experienced it personally.
Alright, if it keeps happening with the SDL2 driver, then I’ll try disabling AntiMicro and see if that works.
So it did indeed keep happening with the SDL2 driver, so I went to plan B: switched back to Xinput and used RetroArch without AntiMicro. So far it seems to be working, as I’ve been playing CD games on the Mednafen PCE emulator, which seems to have a particularly high incident rate with the random game focus bug for some reason, and it just had a couple noticeable performance hitches like the kind that would happen just before the bug would take effect, but so far, no further incidents.
e: Okay, there was another instance of the random game focus bug, though it would seem that not having AntiMicro active at the same time as RetroArch reduces the likelihood of it happening.
I hate to triple-post on a topic, but I must mention that even though the random bug happens less often now that I no longer use AntiMicro with Retroarch, it still happens often enough that I’ve noticed a pattern in its behavior: namely, it seems to happen most often in moments where a lot of button presses are happening in a short period of time, especially in combination with the rewind feature, so maybe that’s related somehow.
Also, I wonder if using Vulkan as my Retroarch graphics driver has anything to do with it, since in the debug logs, there’s always a long string of Vulkan debug messages about creating frame buffers at some point prior to the bug happening.