Bleat required regardless of BLE flag in CLI

[chalkers]

Problem:

On Windows, when a non-compatible USB Bluetooth 4.0 wasn't found, it caused a massive stage trace to appear regardless if you wanted to use ble flag.

Error: No compatible USB Bluetooth 4.0 device found!
    at BluetoothHciSocket.bindUser (C:\Users\andre\Projects\test_craig\node_modules\noble\node_modules\bluetooth-hci-socket\lib\usb.js:70:11)
    at BluetoothHciSocket.bindRaw (C:\Users\andre\Projects\test_craig\node_modules\noble\node_modules\bluetooth-hci-socket\lib\usb.js:28:8)
    at Hci.init (C:\Users\andre\Projects\test_craig\node_modules\noble\lib\hci-socket\hci.js:99:35)
    at NobleBindings.init (C:\Users\andre\Projects\test_craig\node_modules\noble\lib\hci-socket\bindings.js:83:13)
    at new Noble (C:\Users\andre\Projects\test_craig\node_modules\noble\lib\noble.js:50:18)
    at Object.<anonymous> (C:\Users\andre\Projects\test_craig\node_modules\noble\index.js:4:18)
    at Module._compile (module.js:556:32)
    at Object.Module._extensions..js (module.js:565:10)
    at Module.load (module.js:473:32)
    at tryModuleLoad (module.js:432:12)

The problem was that bleat was always being "tried" to be included outside of the checkInit() thus always being required and that the defaultValue was always true.

Solution:
Set defaultValue to be false and include bleat only when necessary.

I also did a version bump to 0.0.21.

[gfwilliams]

Can we not just trap this exception and carry on? This will get used a lot more with puck.js, and I'd rather not have to make people explicitly add --ble every time they use the ide or command-line tools

[chalkers]

The require('bleat') is already in a try so I don't know how you'd trap it. It also feels it's further down in the stack. The CLI (even though it works) exits with exit code 1.

If this isn't the desired behavior of the --ble flag then it should be removed from the CLI or have the option to switch it off. I tried --ble false and --ble=false and it doesn't do anything...

Either way the pull request fixes the behavior that's described in --help. I guess the question is, do you want a massive stack trace to appear for Windows users of the non-Puck.js users?

[gfwilliams]

Yeah, the --ble flag should ideally get modified to allow you to turn it BLE off.

So the CLI actually works fine? I'm not sure why the -1 error code, but that could be because it hit some other snag later down (like not finding a port?).

Are you sure it's not just the console.error that gets called with the exception in the catch block? That looks trivial to convert to a .log, which then gets hidden unless --verbose is called.

[gfwilliams]

Any thoughts on this? Is it likely to just be the console.error?

[chalkers]

I'll take a look soon! Been a little busy!

[chalkers]

Hi there,

There still was a stack trace without the console.error. If you check the last commit, I've changed the --ble to --no-ble now.

0267ec9

So doing --no-ble jumps over the initialization now.

[gfwilliams]

Great, thanks!

[gfwilliams]

I'll add an issue about the stack trace though - I'll have to look into why that's happening

[chalkers]

Can you ping me when you publish the module? :)

[gfwilliams]

Ping!