[RSDK-12899] Fix set device properties

[SebastianMunozP]

This PR fixes a crash on the get_device_properties do_command, caused by a missing continue statement on get_device_properties handling.

This issue happens when the user calls get_device_properties with a non-struct payload, such as this:

{
  "set_device_properties": ""
}

With this fix, these kind of calls are correctly handled producing the following error

{
  "error": "calling set_device_properties on non-struct properties"
}

We are also checking that all settings being modified can actually be modified before doing it

This works OK:

{
  "set_device_properties":   
  {"OB_PROP_DEPTH_NOISE_REMOVAL_FILTER_MAX_SPECKLE_SIZE_INT": {
    "name": "OB_PROP_DEPTH_NOISE_REMOVAL_FILTER_MAX_SPECKLE_SIZE_INT",
    "type": "int",
    "max": 1000,
    "current": 1000,
    "id": 41,
    "default": 1000,
    "min": 1,
    "permission": "read_write"
  },
  "OB_PROP_DEPTH_ALIGN_HARDWARE_MODE_INT": {
    "current": 2,
    "min": 0,
    "max": 5,
    "permission": "read_write",
    "default": 0,
    "name": "OB_PROP_DEPTH_ALIGN_HARDWARE_MODE_INT",
    "id": 63,
    "type": "int"
  }
  }

This does not (missing current field), no fields are modified

{
  "set_device_properties":   
  {"OB_PROP_DEPTH_NOISE_REMOVAL_FILTER_MAX_SPECKLE_SIZE_INT": {
    "name": "OB_PROP_DEPTH_NOISE_REMOVAL_FILTER_MAX_SPECKLE_SIZE_INT",
    "type": "int",
    "max": 1000,    "id": 41,
    "default": 1000,
    "min": 1,
    "permission": "read_write"
  },
  "OB_PROP_DEPTH_ALIGN_HARDWARE_MODE_INT": {
    "current": 5,
    "min": 0,
    "max": 5,
    "permission": "read_write",
    "default": 0,
    "name": "OB_PROP_DEPTH_ALIGN_HARDWARE_MODE_INT",
    "id": 63,
    "type": "int"
  }
  }
}

[hexbabe]

PR title is inaccurate. This fix is for set device properties

[hexbabe]

this is dead code btw

[seanavery]

Do we want to skip with continue or error out here?

[SebastianMunozP]

Good catch, changing this to error out, also, since this is intended to modify multiple settings of the camera at the same time, I'm adding a 2 phase approach, in which first we check if all changes can be applied, and only when we are sure, we apply them,

[seanavery]

LGTM!

Do we want to add a test to repro the mentioned error->fix in the description?