Fix iPad pointer and trackpad input

[alantangok]

Summary

• use focused absolute hover for normal desktop pointer input and reserve GCMouse deltas/pointer lock for host-requested relative mode
• keep UIHoverGestureRecognizer as the sole absolute movement producer
• preserve raw trackpad button transitions without duplicate or premature multi-button chord edges
• refresh pointer mode after host cursor-mode, app, key-window, scene, view, and active-app PiP transitions

Verification

• unsigned Release iOS device build succeeded
• Xcode Release static analysis succeeded
• narrow source regression checks cover absolute/relative routing, focus gating, single movement ownership, and button-mask transition handling
• no project file, .checkpoint, or scripts/ changes

Device validation still required

• switch between desktop absolute input and a host application requesting relative/captured input
• leave and return to fullscreen/Stage Manager and confirm lock/hover recovery
• verify left+right and drag+secondary-button chord sequencing on physical mouse/trackpad

Split from #93.

[alantangok]

@SandiyosDev
Hi, when you have a chance, could you please review this PR first? I split the original work into separate PRs, and I’d like to address any feedback on this one before you review the others. Thanks!

[SandiyosDev]

@SandiyosDev
Hi, when you have a chance, could you please review this PR first? I split the original work into separate PRs, and I’d like to address any feedback on this one before you review the others. Thanks!

I'll review your changes in the next two weeks, if you'd like to help me get your changes merged faster, please test these PRs on your personal device as well

Thank you for your PRs

[alantangok]

Thanks for the update! I’ll test the relevant changes on my iPad and run regression checks on my iPhone, then post the device models, OS versions, and results on each PR to help streamline the review.

[SandiyosDev]

thanks for the split, reviewing this one first as you asked. One thing from my own testing before the code, on my iPad (iPadOS 16.6.1) a left click under pointer lock comes in as a direct touch with an empty buttonMask, not as an indirect pionter touch, so in relative mode the new recognizer won't see the click at all and it falls through to the finger path, that's the jump #88 was about

which iPadOS did you test on, and did you try a host that captures the cursor? If you can make the locked click case work, I'll test your builds on my iPad to see if the bug can be reproduced...

[SandiyosDev]

And a code review, the gate/mapper enums just return their argument, inline them, and pointerButtonTouchSupported is always true at our deployment target so the GameController button handlers are dead code now