Testing SDL patch for FFB quality
Description of Changes
This (should be) a patch of SDL that implements my planned fix to the library. This should address cases where redundant flags were being set to DirectInput's SetParameters method, which appears to have been causing some wheels to misbehave in a way that was best described as "worse feeling/quality/definition" with various FFB effects.
Rationale behind Changes
This is a draft PR only so that I can get the CI bot to apply this patch to SDL. The final version should go into the SDL project itself.
Suggested Testing Steps
If you have a wheel which feels worse ever since PCSX2 switched to SDL, try this version and compare directly to see if things feel better.
For the rest of us, you can validate whether this has fixed the (presumed) low-level problem by opening Wireshark and peeking at USBPcap: If the patch worked correctly, you should only see one PID report being sent per update, instead of two, under most circumstances (e.g. when a racing game like GT4 implements its FFB by rapidly updating a constant force).
Try a Wireshark filter string like this:
usbhid.data && usbhid.data[0]!=0x41 && usb.dst != host
If that sounded convoluted, it's late and my brain is cooked after work.
Did you use AI to help find, test, or implement this issue or feature?
fuck no
@Mrlinkwii This should not be attached to any milestone.
This PR will never be merged--it's being used so that testers can access a version of PCSX2 that is directly bundled with an experimental SDL patch from a trusted source (e.g. not sending people .zip files directly).
When this work is finished, the final version will be merged into SDL itself.