Feat/objectarium instance

[ErikssonJoakim]

This PR finishes the following issue by implementing a objectarium handler for indexing the initialization of objectarium smart contracts:

• 💿 Index objectarium contract #7

As of date there seems to be a problem filtering the code id through the manifest so the contract filtering is implemented in the handler.

objectaria query including config and limits

{
  "data": {
    "query": {
      "objectaria": {
        "id": "okp41jg5y3m2u9ac50g5jsd7mksh3y9nvp3pyuf5fj8entsk9tv2tq4gsv9r25d",
        "owner": "okp41zn4x0geu5eemk7ul05ld7n9e4kzs7cxx4al3fs",
        "config": {
          "hashAlgorithm": "sha224",
          "acceptedCompressionAlgorithms": [
            "snappy"
          ]
        },
        "limits": {
          "maxObjects": "20",
          "maxTotalSize": "1000",
          "maxObjectPins": "1",
          "maxObjectSize": "50"
        },
        "objectariumObjects": {
          "nodes": []
        },
        "messages": {
          "nodes": [
            {
              "block": {
                "height": "4004814"
              }
            }
          ]
        }
      }
    }
  }
}

[ccamel]

LGTM! 👍

[fredericvilcot]

As a general remark, errors are handled nowhere in code and I see no relevant test suite.
Are there plans to implement it in the near future?

[ccamel]

As a general remark, errors are handled nowhere in code and I see no relevant test suite. Are there plans to implement it in the near future?

That's a good question. AFAIU, the code for indexing (i.e. mapping) is ran in a virtual machine (required for decentralisation) driving the process. That's why the best approach here is the "let's propagate" and "fail fast" pattern, and just let the node runner do what it takes to process properly.

Test suites are considered, but it's currently difficult to achieve. You can get more info here: https://academy.subquery.network/build/testing.html

[fredericvilcot]

As a general remark, errors are handled nowhere in code and I see no relevant test suite. Are there plans to implement it in the near future?

That's a good question. AFAIU, the code for indexing (i.e. mapping) is ran in a virtual machine (required for decentralisation) driving the process. That's why the best approach here is the "let's propagate" and "fail fast" pattern, and just let the node runner do what it takes to process properly.

Test suites are considered, but it's currently difficult to achieve. You can get more info here: https://academy.subquery.network/build/testing.html

@ccamel
Thanks for these explanations, I see now the point much clearer.
I agree that failing fast is a great move but it must not justify the non-handing of errors in the whole codebase.
We will especially lack the immediate context of the error, and lose the execution stack trace, IMHO. And thus even more if the error is not well handled in the library..

[amimart]

LGTM!

Following remarks about error handling, I agree with the failing fast mechanism and to go further, I think this is something that should be took into account at the operation level and be monitored :)

[ErikssonJoakim]

@amimart for the monitoring SubQuery does offer a management service where it's possible to receive alerts on the node health regarding indexer stalling, service health and deployment status

[fredericvilcot]

LGTM

[bot-anik]

🎉 This PR is included in version 1.0.0 🎉

The release is available on GitHub release

Your semantic-release bot 📦🚀