Fix wireguard integrations for 4.8.x

[Hertzole]

This fixes the WireGuard calls to the correct endpoints as of version 4.8.1. This is a breaking change and will no longer support anything below 4.8. While it's possible to support older versions, it would add more code to maintain to route each call to the correct version.

This can be used to fix HarvsG/ha-glinet4-integration#57, which I also have a fix for once this package gets a new updated published.

[Hertzole]

@HarvsG I have added versioning by passing a version string, usually gathered from the get_info call.
I do not have a pre 4.8 device to test with, but I assume that it will work since I just took the old code and put it back.

I've also added tests that use the new API and tests for starting and stopping. Anyone with an old device could run these tests to make sure everything works. I can confirm that it works on 4.8.1.

[HarvsG]

Just a code review. I hope to test it tomorrow.
Lots of linting errors

[HarvsG]

Ill need to confirm this

[Hertzole]

I'm not sure what to do about most of the linting errors about mixed line endings. I fixed some whitespace issues and unnecessary else statements though.

[HarvsG]

I'm not sure what to do about most of the linting errors about mixed line endings. I fixed some whitespace issues and unnecessary else statements though.

I'll fix them when I review tomorrow

[HarvsG]

Can you check if this is true for modern firmware - as it seems you are using tunnel_id and ned peer_id

[Hertzole]

What exactly do you want me to verify? Just so I understand.
But in-case you're referring to the comment being outdated, you'd be right.
Here's the full output of calling that endpoint on 4.8.1 (with some redactions)

{
	"id": 7,
	"jsonrpc": "2.0",
	"result": {
		"status_list": [
			{
				"enabled": true,
				"type": "wireguard",
				"tx_bytes": 24953488020,
				"peer_name": "PEER NAME HERE",
				"domain": [
					"XXX.XXX.XXX.XXX"
				],
				"group_id": 1234,
				"via": "wgclient1",
				"port": 51820,
				"tunnel_id": 10,
				"rx_bytes": 49725356924,
				"peer_id": 5678,
				"status": 1,
				"name": "Primary Tunnel",
				"ipv4": "XXX.XXX.XXX.XXX"
			}
		],
		"mode": 0
	}
}

Should I update the comment with the new response?

[HarvsG]

does 4.8 support more that one connection at once?

[Hertzole]

does 4.8 support more that one connection at once?

According to their releases, yes: "Added VPN multi-instance, allowing support for enabling multiple VPN clients at the same time."
I just personally can't verify it because I don't have another wireguard server to connect to.

[HarvsG]

Ok so HarvsG/ha-glinet4-integration#66 now works with these API changes and hopefully adds support for multiple wg clients simultaneously.

I've done quite an ugly change to the WG api calls and to use peer_id and tunnel_id interchangeably

Can you test your system with this PR and the above PR and confirm it works

[Hertzole]

I am getting a seemingly unrelated issue with an index out of bounds error on this line which stops me from testing. Printing the output of dev_info I only get 3 items, so no wonder 5 is out of range. Why 5 specifically?

[HarvsG]

I am getting a seemingly unrelated issue with an index out of bounds error on this line which stops me from testing. Printing the output of dev_info I only get 3 items, so no wonder 5 is out of range. Why 5 specifically?

list(DeviceInterfaceType) turns the strenum defined at the top into a list with 7 items.

Which means it's not the 5 that is out of bounds but dev_info.get("type") (aka dev_info['type']) that is too large. Debug dev_info and see.

It may be that your device has a new interface type that has been added to the API
https://github.com/HarvsG/ha-glinet4-integration/blob/fc9327adf7a89f3c5e40487c725bafd398e3eb43/custom_components/glinet/router.py#L51

[Hertzole]

Ah, I see now. This router has a 6Ghz network, which is not listed. It returns as type 11. There's also WiFi 7, which I have disabled (and believe was disabled by default?)

I guess my phone just didn't use the 6Ghz network before when I set this up? Because now it does and that causes the issue.

Update: I enabled WiFi 7 / MLO WiFi (which I honestly don't know how it works), and that returns type 9. I do not know what type 8 and 10 are.

Update 2: I went ahead and tested and other networks. I can not find network type 8, but found 9 (MLO), 10 (MLO Guest), 11 (6Ghz), and 12 (6Ghz Guest). I can make a separate pull request for that later. I can now test these actual changes and will post on the PR you linked.

[Hertzole]

Can you upload the diff or create your own pr on this repo

The diff for what exactly? I'm sorry, but I'm not quite following what you mean.

If you're talking about the network interfaces, I made a pull request to the other repo. HarvsG/ha-glinet4-integration#67