Inconsistent naming case using --json

[jcamiel]

--json is producing a file which use camelCase and snakeCase :

          "request": {
            "cookies": [],
            "headers": [
              {
                "name": "Host",
                "value": "google.fr"
              },
              {
                "name": "Accept",
                "value": "*/*"
              },
              {
                "name": "User-Agent",
                "value": "hurl/5.0.0-SNAPSHOT"
              }
            ],
            "method": "GET",
            "queryString": [],
            "url": "https://google.fr/"
          },

And

            "certificate": {
              "expire_date": "2024-07-29 14:54:21 UTC",
              "issue": "C = US, O = Google Trust Services LLC, CN = GTS CA 1C3",
              "serial_number": "f3:eb:d0:e9:49:3e:8c:3e:10:8a:74:c4:33:e2:66:2f",
              "start_date": "2024-05-06 14:54:22 UTC",
              "subject": "CN = *.google.fr"
            },

          "timings": {
            "app_connect": 59653,
            "begin_call": "2024-05-19T15:09:07.805338Z",
            "connect": 24655,
            "end_call": "2024-05-19T15:09:07.897452Z",
            "name_lookup": 3617,
            "pre_transfer": 59883,
            "start_transfer": 91551,
            "total": 91608
          }

We should choose a convention and stick to it.

(personal, but not strong, preference is snake_case)

[jcamiel]

@lambrospetrou we'll certainly fix some field names for --json and for --report-json for the next Hurl version (once we choose a casing)

[lambrospetrou]

Thank you for letting me know :)

[fabricereix]

camelCase names (such as queryString) came originally from the HAR format (http://www.softwareishard.com/blog/har-12-spec).
It makes indeed sense to stick to only one format, I also prefer the snake_case one.