Align Error messages from validator and parser

[brianpos]

Is your feature request related to a problem? Please describe.
There are some errors that the parser is able to check that the validator also checks, these should really be aligned so that they use the same error code, and hopefully return the same location. (so they can be de-duplicated when needed)

e.g.

{
    "resourceType": "OperationOutcome",
    "issue": [
        {
            "severity": "error",
            "code": "code-invalid",
            "details": {
                "coding": [
                    {
                        "system": "http://hl7.org/fhir/dotnet-api-operation-outcome",
                        "code": "6007"
                    }
                ],
                "text": "Code 'grab' does not exist in the value set 'PublicationStatus' (http://hl7.org/fhir/ValueSet/publication-status)"
            },
            "expression": [
                "SearchParameter.status[0]"
            ]
        },
        {
            "severity": "error",
            "code": "code-invalid",
            "details": {
                "coding": [
                    {
                        "system": "http://firely.com/CodeSystem/ErrorMessages",
                        "code": "PVAL116",
                        "display": "Invalid code"
                    }
                ],
                "text": "Value 'grab' is not a correct code for valueset 'PublicationStatus'."
            },
            "diagnostics": "Value 'grab' is not a correct code for valueset 'PublicationStatus'. At SearchParameter.status, line 15, position 21",
            "expression": [
                "SearchParameter.status"
            ]
        }
    ]
}

Describe the solution you'd like
PVAL116 and 6007 are the same error, shouldn't they be the same coding?
Alternatively provide another routine to be able to cross reference the codes, and or locations. The validator looks like it doesn't use the short form of the path where it doesn't consider every item an array, and only include the positional accessors when it is a collection.

Describe alternatives you've considered
I'm contemplating writing a routine to manually do this processing.

[mmsmits]

We should make sure the property validator and the parser should not double errors.
Don't know how yet, but it's a good idea.

[mmsmits]

Maybe not run attribute validation when an element has reported a parse error.

[brianpos]

There is a flag on the serializer for that already.
I think it should be just changing the message that is being reported to use the same coding, and ensure that it doesn't re-add it for the same location.