Startup is faster when the socket bridge uses 127.0.0.1 instead of localhost #1103

• Version: 0.5.2
• .NET: 10.0.302 SDK
• Target: win-x64 (Windows 11)

On some of our Windows machines, app startup was noticeably slow. Replacing localhost with 127.0.0.1 on both ends of the Electron ⇄ .NET socket bridge made it start visibly faster.

I don't have a confirmed root cause. My guess is that on those machines localhost goes through name resolution (and/or resolves to ::1 first, where nothing is listening), so each connection attempt costs a delay before falling back to IPv4.

Where the two ends disagree today

• src/ElectronNET.Host/main.js, in startSocketApiBridge (~line 264) — the bind address depends on whether the port was forced:
  const host = !port ? '127.0.0.1' : 'localhost';
• src/ElectronNET.API/Runtime/Services/SocketBridge/SocketBridgeService.cs, in the constructor — the client always uses the name:
  this.socketUrl = $"http://localhost:{this.socketPort}";

Steps to Reproduce:

1. On Windows 11, create (or take) any Electron.NET app targeting win-x64.
2. Run electronize start and time from launch to the first interactive window.
3. Edit main.js to const host = '127.0.0.1'; (unconditionally) and SocketBridgeService.cs to this.socketUrl = $"http://127.0.0.1:{this.socketPort}";.
4. Rebuild, run electronize start again, and compare the timing.

The difference was clear on the affected machines.

I don't know the consequences of hard-coding the literal — whether any supported scenario relies on localhost resolving to something other than 127.0.0.1, or on the host binding ::1. The conditional in main.js presumably exists for a reason I'm not aware of. Happy to open a PR if the approach seems acceptable.

Possibly related to #1024.