Fix HiDPI mode filtering for portrait displays

[shaw-baobao]

What & why

Rotated portrait displays could lose their usable HiDPI modes because the native aspect ratio was inferred from the largest backing buffer, which could belong to a landscape HiDPI mode. The resolution menu also used landscape-only minimum dimensions.

This change:

• derives the panel aspect from its largest unscaled timing before falling back to scaled modes;
• applies orientation-independent minimum dimensions;
• keeps opposite-orientation modes out of the resolution menu and smooth-scaling slider;
• adds focused geometry regression tests and a make test entry point.

How tested

• make test — 3 tests passed
• make compile
• ./scripts/release.sh v0.0.0-ci — universal arm64 + x86_64 DMG built
• Physical 90° external display: 12 eligible portrait HiDPI modes enumerated; switched from 1440×2560 to 1080×1920 with a 2160×3840 backing buffer, then restored the original mode

Checklist

• Builds locally (./dev.sh, or ./scripts/release.sh v0.0.0-ci for the full release build)
• Added/updated Crisp/Resources/Localizable.xcstrings for any new user-facing strings (N/A: no new strings)
• Screenshot or short clip for any UI change (N/A: no visual or layout change)

[didriksg]

Hardware-verified on a physically rotated Dell U2412M (1200x1920 portrait) alongside an AOC Q27G3XMN (2560x1440 landscape) and the built-in display:

• Before: the Dell's Resolution section showed "No display modes available".
• After: portrait modes are listed (1200x1920 default, 1000x1600 / 840x1344 / 800x1280 in both HiDPI and non-HiDPI groups), and mode switching works.
• No regression on the landscape AOC or the built-in display.

Thanks for the fix!