fix(gfn): align keyboard mapping with official client for non-US layouts

[zortos293]

This PR aligns OpenNOW keyboard input mapping with the official GFN web client to fix keypresses on non-US keyboard layouts.

Renderer (opennow-stable/src/renderer/src/gfn/inputProtocol.ts):

• Remove layout-dependent keyCode fallback; always prefer physical code-based VKs
• Remove shouldUsePhysicalOemVirtualKey and physicalOemVirtualKeyCodes logic
• Update special VK mappings: PrintScreen (0x2A), IntlRo (0xC2), IntlYen (0xC1)
• Add missing mappings: OSLeft/Right, KanaMode, Lang1/2, Convert/NonConvert, NumpadClear

Native (native/opennow-streamer/src/input.rs, native/opennow-streamer/src/gstreamer_input.rs):

• Add layout_mapped_keyboard_keycode mapping Set 1 scancode → official position-dependent VK
• Apply keycode mapping before encoding Windows keyboard input
• Scancode remains zeroed to match browser client behavior

Tests:

• Update renderer tests to verify physical code preference for German layout
• Add native test covering scancode→VK mapping for official position keys

---

Note

Medium Risk
Touches all keyboard input encoding for GFN (renderer + native Windows); wrong VK tables would break typing on non-US layouts but scope is limited to mapping logic with added unit tests.

Overview
Aligns GFN keyboard packets with the official web client so non-US layouts (e.g. QWERTZ) send physical key-position virtual keys with scancode 0, instead of the browser’s layout-dependent keyCode.

Renderer (inputProtocol.ts): virtualKeyFromEvent now prefers event.code → VK when available; the layout-aware OEM override (shouldUsePhysicalOemVirtualKey / physicalOemVirtualKeyCodes) is removed. Several code→VK entries are corrected or added (PrintScreen, IntlRo/Yen, OS keys, Japanese IME keys, numpad clear). Tests expect physical-code VKs on German punctuation and Y/Z swaps.

Native streamer (input.rs, gstreamer_input.rs): Windows key events run through new layout_mapped_keyboard_keycode (Set 1 scancode → official position VK, fallback to OS keycode) while scancodes stay zeroed, mirroring the browser path.

^{Reviewed by Cursor Bugbot for commit b306f60. Configure here.}

[cursor]

Bugbot couldn't run - usage limit reached

Bugbot is counted against Cursor usage for this user or team, and this run hit a usage or spend limit.

A user or team admin can review and increase usage limits in the Cursor dashboard.

(requestId: serverGenReqId_ff0cbfea-8c8c-46c7-97fc-ffade4bd0ae5)

[coderabbitai]

Important

Review skipped

Auto reviews are disabled on base/target branches other than the default branch.

Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Pro Plus

Run ID: fc359473-33a7-47c4-a8f7-f0f64f3c1501

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• 🔍 Trigger review

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch capy/fix-gfn-keyboard-non

---

_{Comment @coderabbitai help to get the list of available commands.}

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: b306f60be0

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[chatgpt-codex-connector]

Preserve NumLock instead of remapping raw 0x45 to Pause

In Windows native input capture, if NumLock comes from Raw Input as make_code 0x45 without RI_KEY_E0, normalize_scancode() leaves it as 0x0045; this new table then converts it to Pause (0x13) instead of preserving the incoming VK_NUMLOCK (0x90). That means native fullscreen sessions can send/toggle Pause when the user presses NumLock, so handle the 0x45 ambiguity with the fallback keycode or normalize NumLock to the existing E045 entry.

Useful? React with 👍 / 👎.

[chatgpt-codex-connector]

Add native mappings for JP language keys

This international-key section covers JP IntlRo/IntlYen, but it skips the other physical JP 106 keys whose raw scancodes are 0x70 (Hiragana/Katakana), 0x79 (Henkan/Convert), and 0x7B (Muhenkan/NonConvert). In the native path those keys fall through to Windows fallback VKs while the renderer path added the official 0xE9/0xEA/0xEB codes in this commit, so Japanese-layout users get different behavior depending on whether native input capture is active.

Useful? React with 👍 / 👎.