openapi-to-graphql Don't set the content-type when there is no payload

Hi,

I can't use the Mock Web Server https://github.com/stoplightio/prism because the GET operations set a content-type & content-length in the http headers. I think it's an error. a content-type & content-length must be set when there is a payload with operations like 'PUT' or 'POST'...

Web server like https://github.com/stoplightio/prism (mock of openapi) rejects the calls because it validate the request and a content-type has nothing to do with GET operations.

Exemple of a call made by openapi-to-graphql to my openapi web server:

GET http://127.0.0.1:4010/RESTMethod HTTP/1.1 content-type: application/json (bad) accept: application/json host: 127.0.0.1:4010 content-length: 0 (bad) Connection: close

Jun 12 '20 15:06 lionelschiepers

Hi, I'm not a specialist of node.js/javascript.

I tried to find the reason why payloadRequired was not correctly initialized during the unit test but didn't find the reason. That's why I changed the logic by checking the method instead of payloadRequired. Do you have any advise?

Lionel.

Jun 15 '20 11:06 lionelschiepers

@lionelschiepers I apologize for the late reply! A lot of things have been going on at my end.

We set the content-type here. This will be probably a super simple fix, we just need to check for the operation.method to see if it is a GET request or not. I'm not sure where we set the content-length, however. I assume that's set by the request.js library by default when we do not include a payload.

Let me give it a shot...

Jun 26 '20 22:06 Alan-Cha

Ah sorry, I got a bit ahead of myself. It seems that you have already made the changes.

This is the way I would have done as well. I do not think we should make things more complicated by checking if the payload is required.

Do these changes address your issue? Do you have any other problems with OtG?

Jun 26 '20 22:06 Alan-Cha

Hi,

Yes. This pullrequest solves my problem. If every body agree on the change I would be the most happiest man in the world if the pullrequest is merged :-)

Thanks, Lionel.

Jul 01 '20 12:07 lionelschiepers

@lionelschiepers I can try to get this checked in later today. I think we need a test case which I will look into.

Jul 01 '20 13:07 Alan-Cha