JSON schema

[n8willis]

Doing some bulk-analysis, I quickly decided that it would be helpful to have some description or documentation of the structure of the JSON output format.

Ideally, a real JSON schema would probably be the best thing, but for starters I wanted to go the human-readable discussion route.

From what Felipe said in casual conversation, the JSON is pretty much just straight-from-the-code output, so there might be some cases where a check or a profile goes against the norm. If anyone maintaining checks or working with large-scale libraries knows more, your input would be greatly appreciated.

So for here, to get things started, I'm just jotting down what I'm seeing.

Top level:

"result"
"sections": array

"result" object

(not finished): int
FAIL: int
INFO: int
PASS: int
SKIP: int
WARN: int

"sections" array: contains only "checks" array.

"checks" array element:

[
description: string
documentation: string? URI?
experimental: boolean
filename: string? src path?
key: array
logs: array
module: string?
proposal: array? single string? src identifier?
rationale: string
result: string?
severity: string
]

"checks" element "key" array:

XML "Section" element?
XML "FontBakeryCheck" element?
array?

"checks" element "logs" array:

message: object
status: string

"checks" element "logs" array "message" object:

code: string
message: string

"checks" element "proposal" array object:

string

[n8willis]

I'm not entirely clear whether some of the string-formatted entries have further constraints on them, and whether the logs and proposal have more structure to them in some (or any) checks/profiles. And not real clear how much difference there is in the other profiles yet. This is just from Universal....