Hi Team, +1 to rely on HTTP status code instead of the code in the response body.
This will take our burden of enforcing governance to define/maintain additional code in the response body which will unlikely to effectively override the HTTP status code. However - if possible - enriching the response of these APIs to adopt HATEOAS would be progressive immensely. Regards, Shyamal On Mon, Jul 26, 2021, 07:20 Zhiyuan Ju <juzhiy...@apache.org> wrote: > Hi, > > Users have 2 ways to use our product: Using Dashboard or Call API directly. > > We have JSONSchema checker on the backend, if users call API directly and > check failed, it will return the JSONSchema error; if users use Dashboard, > why not validate data on Web? So I think there has no need to add a > specific login code in the Response Body :) > > Best Regards! > @ Zhiyuan Ju <https://github.com/juzhiyuan> > > > Chao Zhang <zchao1...@gmail.com> 于2021年7月26日周一 上午9:41写道： > > > HTTP status code will be fine if we don’t care the very specific error > > type. > > > > Chao Zhang > > https://github.com/tokers > > > > On July 25, 2021 at 17:34:40, Jintao Zhang (zhangjintao9...@gmail.com) > > wrote: > > > > I agree! > > At the same time, I also think that HTTP Status code should be used > instead > > of the logical code field. > > > > Zhiyuan Ju <juzhiy...@apache.org> 于2021年7月25日周日 下午3:29写道： > > > > > Hi, > > > > > > Any further discussion on this mail? I'm going to list all APIs and > check > > > which part we should modify. > > > > > > Best Regards! > > > @ Zhiyuan Ju <https://github.com/juzhiyuan> > > > > > > > > > Zhiyuan Ju <juzhiy...@apache.org> 于2021年7月22日周四 下午2:47写道： > > > > > > > I would prefer relying on Status Code instead of `code` (actually > it's > > a > > > > manual logical and extendable code). > > > > > > > > Why not list all API cases then have a choice? > > > > > > > > Best Regards! > > > > @ Zhiyuan Ju <https://github.com/juzhiyuan> > > > > > > > > > > > > Ming Wen <wenm...@apache.org> 于2021年7月22日周四 下午12:36写道： > > > > > > > >> please give an example about i18n for a better understanding > > > >> > > > >> Thanks, > > > >> Ming Wen, Apache APISIX PMC Chair > > > >> Twitter: _WenMing > > > >> > > > >> > > > >> JunXu Chen <chenju...@apache.org> 于2021年7月22日周四 上午11:36写道： > > > >> > > > >> > Agree +1 > > > >> > > > > >> > > > > >> > I think we still need to keep the `code` field. > > > >> > > > > >> > The FE needs to implement i18n according to it. > > > >> > > > > >> > Unless `message` is semantic and can be used as a key of i18n. > > > >> > > > > >> > > > > >> > > > > >> > On Thu, 22 Jul 2021 at 00:02, Ming Wen <wenm...@apache.org> > wrote: > > > >> > > > > >> > > I don't think the `code` filed is useful, HTTP response code is > > > >> enough. > > > >> > > > > > >> > > Thanks, > > > >> > > Ming Wen, Apache APISIX PMC Chair > > > >> > > Twitter: _WenMing > > > >> > > > > > >> > > > > > >> > > Peter Zhu <sta...@apache.org> 于2021年7月21日周三 下午11:18写道： > > > >> > > > > > >> > > > Agree +1. > > > >> > > > And I think we should maintain the `code` filed and maintain > the > > > >> doc of > > > >> > > > API. > > > >> > > > > > > >> > > > > > >> > > > > >> > > > > > > > > > >
