unsupported media type for "application/json; charset=utf-8"

[kikyomits]

Hi, my company has a rule that we mandatory set Content-Type: application.json; charset=utf-8 to all the request/response header. (I know the charset=utf-8 is probably redundant for json though.)
Since OpenAPI Validator drop charactors of Content-Type after the charactor of ; as below, our OpenAPI Validator cannot find the appropriate request body schema and respond unsupported media type application/json.

express-openapi-validator/src/middlewares/util.ts

Lines 6 to 17 in 04a84e5

	export function extractContentType(req: Request): string | null {
	let contentType = req.headers['content-type'];
	if (!contentType) {
	return null;
	}
	let end = contentType.indexOf(';');
	end = end === -1 ? contentType.length : end;
	if (contentType) {
	return contentType.substring(0, end);
	}
	return contentType;
	}

Sample api spec is described as below. Due to the issue above, requests which has Content-Type: application.json; charset=utf-8 cannot be associated with the schema $ref: '#/components/schemas/user'.

paths:
  /user/:userId:
    post:
# --- skip ---
      requestBody:
        content:
          application/json; charset=utf-8:
            schema:
              $ref: '#/components/schemas/user'
# --- skip ---

I think we should use original raw value of Content-Type in request header to find appropriate schema.

express-openapi-validator/src/middlewares/openapi.request.validator.ts

Line 196 in 04a84e5

const content = requestBody.content[contentType];

If you feel okay with my suggestion, let me create PR.

[cdimascio]

@Mk811 i agree, the raw should be used. a PR is greatly appreciated!
we likely also want to fall back to the media-type if the charset or boundary are not present

The syntax is as such:
e.g.

Content-Type: text/html; charset=UTF-8
Content-Type: multipart/form-data; boundary=something

we should fall back to the media-type if a charset or boundary is not provided

[kikyomits]

@cdimascio Sounds great. Thank you for the quick reply.

To sum up, if the Content-Type has charset or boundary in itself, we will use Content-Type to find appropriate request/response body from spec yaml, e.g. text/html; charset=UTF-8.
If not, we will use media-type instead, e.g. text/html.

Does it sound good for you?

[cdimascio]

Yes, that sounds good.

[cdimascio]

@Mk811 how are you making out? If u have any questions or want to discuss more freely, feel free to join us on gitter. https://gitter.im/cdimascio-oss/community

[cdimascio]

fix available in v2.15.1