Extracted from #246.
When unmarshalling a "known" problem type, we currently ignore the "status" property of the problem payload. The status code is hardcoded in the constructor. What if the "status" property in the problem payload was different? Could we log a warning?
What if the "status" property in the problem payload was different from the status code in the HTTP header? Could we log a warning?
For "unknown" problem types, they are unmarshalled to a DefaultProblem instance. For those, we do take the "status" property from the problem payload. If the problem payload doesn't contain a "status" property, it is considered to be 0. Should we log a warning when that happens?
And again, what if its "status" property in the problem payload was different from the status code in the HTTP header? Could we log a warning?
Extracted from #246.
When unmarshalling a "known" problem type, we currently ignore the "status" property of the problem payload. The status code is hardcoded in the constructor. What if the "status" property in the problem payload was different? Could we log a warning?
What if the "status" property in the problem payload was different from the status code in the HTTP header? Could we log a warning?
For "unknown" problem types, they are unmarshalled to a DefaultProblem instance. For those, we do take the "status" property from the problem payload. If the problem payload doesn't contain a "status" property, it is considered to be 0. Should we log a warning when that happens?
And again, what if its "status" property in the problem payload was different from the status code in the HTTP header? Could we log a warning?