chore: dont special case json #169
Conversation
|
@udayvir-singh what do you think ? |
we special case json when checking the payload body via encoding/decoding it. This is done by checking the "content-type" in: rest-nvim#157 There doesn't seem to be a need for that other than possibly validate the payload ? in which case we should implement something more generic.
| @@ -67,23 +66,6 @@ local function get_body(bufnr, start_line, stop_line, has_json) | |||
| end | |||
| end | |||
|
|
|||
| local is_json, json_body = pcall(vim.fn.json_decode, body) | |||
|
|
|||
| if is_json then | |||
There was a problem hiding this comment.
@teto Don't delete these lines; this statement does two things:
- LINE 75: Minify json. ie. remove extra tabs and spaces to reduce bandwidth.
- LINE 78-83: Parse nested json into valid
application/x-www-form-urlencodedstring values. see fix parsing of nested tables and curl arguments #138
There was a problem hiding this comment.
I think 1 is non-relevant, also it means we dont send the exact same payload as specified by the user which is concerning (encoding/decoding floats can raise issues I suppose, same for numbers ?).
As for 2, application/x-www-form-urlencoded seems like a pain https://stackoverflow.com/questions/9870523/what-are-the-differences-between-application-json-and-application-x-www-form-url but indeed that's the curl default. One solution could be to default to application/json but that can be surprising too.
Maybe we could do is implement the same mechanism as for "formatter", but for inputs instead of outputs. have the ability to to preprocess the payload depending on content-type.
we special case json when checking the payload body via
encoding/decoding it.
This is done by checking the "content-type" in:
#157
There doesn't seem to be a need for that other than possibly validate
the payload ? in which case we should implement something more generic.