Remove oneof and ValueFailure from Datapoint and some proto naming improvements

[rafaeling]

No description provided.

[codecov]

Codecov Report

Attention: Patch coverage is 27.39726% with 53 lines in your changes missing coverage. Please review.

Please upload report for BASE (feature/databroker-api-v2@c9f1533). Learn more about missing BASE report.

Files with missing lines	Patch %	Lines
databroker/src/grpc/kuksa_val_v2/conversions.rs	10.00%	45 Missing ⚠️
databroker-cli/src/kuksa_cli.rs	0.00%	3 Missing ⚠️
databroker-cli/src/sdv_cli.rs	0.00%	3 Missing ⚠️
databroker/src/grpc/kuksa_val_v2/val.rs	88.23%	2 Missing ⚠️

Additional details and impacted files

@@                     Coverage Diff                      @@
##             feature/databroker-api-v2      #72   +/-   ##
============================================================
  Coverage                             ?   51.75%           
============================================================
  Files                                ?       33           
  Lines                                ?    12893           
  Branches                             ?        0           
============================================================
  Hits                                 ?     6673           
  Misses                               ?     6220           
  Partials                             ?        0           

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[erikbosch]

So what is now supposed to be returned if there is no value available, a value that is None?

Like if you do a get request, but no value available yet, should yet get a Datapoint back where value is None. If so, maybe we shall add a comment that value may be None to indicate no value available (but signal known)

Well, actually I see no need for change here if we agree that value may be None in some scenarios, then I can add a comment after rebasing my PR.

[BjoernAtBosch]

I accept this change, but comment as proposed by Erik should be added.
Remark:
If needed we could later re-add the oneof (with renamed fields) and still stay compatible with the current version:

message Datapoint {
  google.protobuf.Timestamp timestamp = 1;

  oneof value_state {
    Value value = 2;
    MissingReason missingValueReason = 3;
  }
}

enum MissingReason {
   ...
}

[erikbosch]

@rafaeling - will you update so we can have it merged (If my interpretation of when None shall be used is correct)

[rafaeling]

I accept this change, but comment as proposed by Erik should be added. Remark: If needed we could later re-add the oneof (with renamed fields) and still stay compatible with the current version:

message Datapoint {
  google.protobuf.Timestamp timestamp = 1;

  oneof value_state {
    Value value = 2;
    MissingReason missingValueReason = 3;
  }
}

enum MissingReason {
   ...
}

I added it also to the discussion topic list - https://github.com/eclipse-kuksa/kuksa-databroker/pull/21/files#diff-7ed9bb015da6a2a5d42536ce062cb6df04a6b8a9688a4dcfe15995a3ef300087R556
maybe if we use one day the design documentation we could also add the discussion link why it was decided to design the proto message that way.

[rafaeling]

@rafaeling - will you update so we can have it merged (If my interpretation of when None shall be used is correct)

Will update it

[BjoernAtBosch]

Proto files lgtm

[BjoernAtBosch]

I accept this change, but comment as proposed by Erik should be added.
Remark:
If needed we could later re-add the oneof (with renamed fields) and still stay compatible with the current version:

message Datapoint {
  google.protobuf.Timestamp timestamp = 1;

  oneof value_state {
    Value value = 2;
    MissingReason missingValueReason = 3;
  }
}

enum MissingReason {
   ...
}

[erikbosch]

LGTM