Feature/refactor set

[SebastianSchildt]

This is work in progress.

Replace old set code with streamlined code supporting "/" paths.

Advanced features such as multiset in a later step, when migrating msg semantic to Gen2

[wenwenchenbosch]

I like, that many code is removed. I will test the code now

[wenwenchenbosch]

Issue 1:

VSS Client> setValue Vehicle.Speed 200
{
    "action": "set", 
    "requestId": "0bc7f8af-7e9d-46c9-8de3-21d778b1d168", 
    "timestamp": 1613404411366
}

VSS Client> setValue Vehicle.Speed2 200
{
    "action": "set", 
    "requestId": "02eeebf2-f6fe-44f2-aefb-4dbd8ba30493", 
    "timestamp": 1613404413198
}

VSS Client> setValue bla 200
{
    "action": "set", 
    "requestId": "0a8883b9-546b-4e59-8ed6-4871ca7fb4e4", 
    "timestamp": 1613404429551
}

[wenwenchenbosch]

expected behavior:

VSS Client> setValue Vehicle.Speed2 200
{
    "action": "set", 
    "error": {
        "message": "I can not find Vehicle.Speed2 in my db", 
        "number": 404, 
        "reason": "Path not found"
    }, 
    "requestId": "63ca27e8-1afe-457c-a9cf-56c2d246c38c", 
    "timestamp": 1613404486744
}

[wenwenchenbosch]

issue 2: mqtt publisher does not work any more :(

[SebastianSchildt]

Now non-existing paths can not be set anymore. Fresh unittestests still missing....

[wenwenchenbosch]

@SebastianSchildt You looped setPairs three times!

[SebastianSchildt]

I was (that is what I told myself) going for readability instead of performance and deeper nesting. I hope the datastructure with the vector of tuples thing access compiles down to just moving a pointer, but I do not know 😕
Maybe we add an enhancement issue to look into that later?

[wenwenchenbosch]

you are right with readability. But in this case, I do not think, that putting everything inside one loop is less readable than using 3 loops. If you like to improve the readability, better refact the method with multiple functions. It is always best practices to have smaller function/method, better than a function with more than 100 lines of code. But If you like, you can create an issue for this, instead of fixing it now.

[SebastianSchildt]

#124

[wenwenchenbosch]

what is UUT?

[SebastianSchildt]

I took the style form the existing tests, i assume "Unit under test"

[wenwenchenbosch]

why this change necessary?

[SebastianSchildt]

That was an old token (see https://jwt.io/), the old code just never checked it (there are ealier and additional access checks now, in a place that is not mocked in this test unit. as the test already provided a token (albeit it was irrelevant) I replaced it with a validone, so the test goes through)