chore: improve trezor dashboard

[ovitrif]

Description

This PR is intended to address review feedback without changing the Trezor dashboard behavior introduced by #792:

1. Addresses the remaining actionable review feedback from feat: trezor hardware support #792.
2. Propagates Trezor repository, reconnect, credential, and BLE failures instead of silently swallowing them.
3. Splits and tightens Trezor UI state, including immutable UI collections and a composite connected-device state.
4. Adds focused unit coverage for retry, reconnect, filtering, persistence, send guards, BLE close timeout reporting, and ViewModel flows.

This PR should be validated only against #792.

Preview

testing-3x.mp4

QA Notes

Base comparison:

• Base PR: feat: trezor hardware support #792, feat/trezor-hardware-support.
• Stacked PR: feat/trezor-hardware-support-fixes.
• This PR is built on tip of feat: trezor hardware support #792.
• Expected behavior: the Trezor dashboard should behave the same as feat: trezor hardware support #792 while carrying the review fixes in this stack.

Official emulator setup:

1. Use the official Trezor User Env from https://github.com/trezor/trezor-user-env.
2. Start it from ~/repos/github/trezor-user-env with ./run.sh --no-regtest.
3. If the Apple Silicon container cannot load SDL3, install the missing runtime packages once with docker exec trezor-user-env.mac sh -lc 'apt-get update && apt-get install -y libsdl3-0 libsdl3-image0'.
4. Start Bridge and a deterministic T2T1 emulator through the User Env controller:

python3 - <<'PY'
import asyncio, json, websockets

async def send(payload):
    async with websockets.connect('ws://localhost:9001') as ws:
        await ws.recv()
        await ws.send(json.dumps(payload))
        print(await ws.recv())

async def main():
    await send({'type': 'bridge-start', 'version': 'node-bridge', 'id': 1})
    await send({'type': 'emulator-start', 'model': 'T2T1', 'version': '2-main', 'wipe': True, 'id': 2})
    await send({
        'type': 'emulator-setup',
        'mnemonic': 'all all all all all all all all all all all all',
        'pin': '',
        'passphrase_protection': False,
        'label': 'Bitkit Test Trezor',
        'needs_backup': False,
        'id': 3,
    })
    await send({'type': 'background-check', 'id': 4})

asyncio.run(main())
PY

Physical phone install:

• adb reverse tcp:21325 tcp:21325
• TREZOR_BRIDGE=true TREZOR_BRIDGE_URL=http://127.0.0.1:21325 ./gradlew installDevDebug

Android emulator install:

• TREZOR_BRIDGE=true TREZOR_BRIDGE_URL=http://10.0.2.2:21325 ./gradlew installDevDebug

Validation done:

• Official Trezor User Env cloned at ~/repos/github/trezor-user-env and started with ./run.sh --no-regtest.
• Host Python websockets installed for controller commands.
• User Env controller ping passed.
• bridge-start with node-bridge passed.
• T2T1 2-main emulator starts after installing libsdl3-0 and libsdl3-image0 in the container.
• emulator-setup passed with the deterministic all all ... seed and label Bitkit Test Trezor.
• curl -X POST http://localhost:21325/enumerate returns the official Bridge device: path=1, id=127.0.0.1:21324, apiType=usb.
• adb reverse tcp:21325 tcp:21325 passed on the Pixel 10 Pro test phone.
• Bridge-enabled dev APK built and installed on the Pixel 10 Pro.
• ./gradlew compileDevDebugKotlin testDevDebugUnitTest detekt --rerun-tasks.

Regression checklist against #792:

• Scan shows the same Bridge device as feat: trezor hardware support #792.
• Connect succeeds and shows the same Trezor features as feat: trezor hardware support #792.
• Get address succeeds with the same script/network selection as feat: trezor hardware support #792.
• Get public key succeeds with the same script/network selection as feat: trezor hardware support #792.
• Sign and verify message behave the same as feat: trezor hardware support #792.
• Send/compose path produces the same funded or no-funds result as feat: trezor hardware support #792.
• Disconnect, reconnect, and forget-device cleanup behave the same as feat: trezor hardware support #792.

PR can be approved after code review and light smoke testing, without the need for extensive testing.

The dev validation approach, as described above, involved retesting same functionalities tested in #792 then comparing the results for equality in user-facing functional behavior.

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 1750d15a1d

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]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 1ecae3dcbc

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".