[Migrated] Bug: cocli corim display

[deeglaze]

Source: veraison/corim#69, confirmed still an issue

cocli corim display on a signed corim creates a json and prints "Meta:" and "CoRIM:" in https://github.com/veraison/corim/blob/main/cocli/cmd/corimDisplay.go#L88 and later.

This causes a problem when parsing the JSON file output through python or command line, and appears to be a non-conforming JSON.
Can you confirm this is a bug? and perhaps fix it?

Similarly cocli comid display also outputs the file name with >> at the top of the JSON file and makes it difficult to parse.

[ravjot07]

@deeglaze can we provide a --raw (or similarly named) flag to control whether extra labels/headings are displayed, giving users the choice of purely raw JSON vs. human-friendly headings.

[deeglaze]

I think it is difficult to say if this can be resolved fully, as CoRIM does not have a JSON-friendly CDDL definition.

Anything we say is --json would need to be proposed as an enhancement I-D to the CoRIM spec I think, since the CBOR serialization to JSON takes a design decision that isn't standardized (see nlohmann/json#1129), and the CDDL that binds names to constants is not intended to be a basis of JSON serialization. I don't want the reference implementation of the RFC to also prescribe JSON without a spec.

The display tooling is meant purely for human diagnostics. I'm tempted to close as won't fix, or label as enhancement rather than bug.

[ravjot07]

I think it is difficult to say if this can be resolved fully, as CoRIM does not have a JSON-friendly CDDL definition.

Anything we say is --json would need to be proposed as an enhancement I-D to the CoRIM spec I think, since the CBOR serialization to JSON takes a design decision that isn't standardized (see nlohmann/json#1129), and the CDDL that binds names to constants is not intended to be a basis of JSON serialization. I don't want the reference implementation of the RFC to also prescribe JSON without a spec.

The display tooling is meant purely for human diagnostics. I'm tempted to close as won't fix, or label as enhancement rather than bug.

@deeglaze yaa you are right that's the whole point of display tooling is for human diagnostic, however should we consider adding something like - raw tag? As some of the user faces this issue as I have seen one more issue that had mentioned same problem in corim repo

[deeglaze]

I wouldn't reject it outright. I would need to see a PR and see if the other maintainers share my stance.

Edit: If we say it's rendering in the template format, I would be okay with it. I just don't know how the template format handles binary strings.