WCF 4.0 Rest Service setting content type
I've just finished my first WCF 4.0 Rest service and don't understand why the Content-Type of the data being returned changes between calling the service via Fiddler and Firefox. Here's my service contract:
[ServiceContract]
public interface IProjectService
{
[OperationContract]
[WebGet(UriTemplate = "project/{id}/json", BodyStyle = WebMessageBodyStyle.Bare, ResponseFormat = WebMessageFormat.Json)]
ProjectDataContract GetProjectJson(string id);
[OperationContract]
[WebGet(UriTemplate = "project/{id}/xml", BodyStyle = WebMessageBodyStyle.Bare, ResponseFormat = WebMessageFormat.Xml)]
ProjectDataContract GetProjectXml(string id);
[OperationContract]
[WebGet(UriTemplate = "userprojects/{userKey}/json", BodyStyle = WebMessageBodyStyle.Bare, ResponseFormat = WebMessageFormat.Json)]
List<ProjectDataContract> GetProjectsByUserJson(string userKey);
[OperationContract]
[WebGet(UriTemplate = "userprojects/{userKey}/xml", BodyStyle = WebMessageBodyStyle.Bare, ResponseFormat = WebMessageFormat.Xml)]
List<ProjectDataContract> GetProjectsByUserXml(string userKey);
}
As you can see, I'm setting the response format for each operation. If the request ends with "/json" then I'm returning json data. If the request ends with "/xml", then xml data is returned. At least that is what my intentions are.
When I make a call to http://localhost:5050/ProjectServiceLibrary/project/27/xml in Firefox, I can see the content-type is set to "text/html" whereas the same request invoked in fiddler shows the correct content type of "application/xml". Same thing happens for a call to a "/json" suffixed request - "text/html" in firefox and "application/json" in fiddler.
So, why is this happening? Which one do I trust? I downloaded JSONView Firefox add-on, but that makes everything look like json. It treats the XML as JSON.
I'm sure I'm missi开发者_如何学编程ng something obvious. Any help would be greatly appreciated.
This is related to the Accept header in the request sent by the client. The Accept header contains a prioritized list of MIME types. Accept headers are defined by the client (Firefox, Fiddler), and tell the server which content-types it is capable of receiving. The server will use the best match based on priority and compatibility.
Accept headers generated by FireFox give text/html a higher priority - telling the server to send text/html if it is possible. You will probably find that Fiddler does the opposite, giving application/xml the higher priority - this explains what you are seeing.
There's more detailed information about request headers on Kris Jordans blog.
精彩评论