a resource accepts different http verb, seems we are providing verbs but they are not actually used for resources, e.g. the verb is ignore (both GET, DELETE or PUT will work the same)

the verb is already in the configuration object, but does not currently seem to work as expected,
also for different verbs on a same resource/path

shutdown does not seems to be graceful in the current setup, seems necessary to force shutdown

specify override parameters in json body of response.

Simple example project showing how to mock out external references to build automated acceptance tests.

Early on it as very easy to call the "run" command in a test project and the library would run in a background thread for you. We should add back in that behavior for folks who want to the library more inline instead of a server.

At this point it might be easier to change .Run to do what the old default behavior was and create a new on for the server behavior?

// Would run and return
KestrelMock.Run(config);

// Would run the server
KestrelMock.RunServer(config)

Thoughts?

add .NET8 and AOT compilation to reduce image size? just an idea

Give a way to view recent requests and the data that was sent to enable more advanced mocking flows.

Original Request:
"With kestrel mock, we've got a use case where our service receives a message, translates and then PUTs to another service ...

So we'd set up a Kestrel Mock to return 200 on the PUT ...

but ideally we want to see that translated data that was made to the PUT ...

Is there a way in Kestrel Mocks to see was was sent to mocked API please?"

Having to to use the following seems odd and doesn't match the docs.

KestrelMock.KestrelMock.Run(config);

tired to pull and run from dockerhub the image we published, seems to have some problems. couldn't not get the mocked responses. (tried both localhost and 127.0.0.1)

docker run -p 5006:5000 -e ASPNETCORE_URLS=http://127.0.0.1:5000 jasonrowe/kestrelmock

i also tried inspect the image and required configuration files seems to be there (responses folder and appsetting.json),
and server "starts listening to 5000 correctly". Not sure what's wrong

Hosting environment: Production

Content root path: /app

Now listening on: http://127.0.0.1:5000

Application started. Press Ctrl+C to shut down.

Given a developer configures a sequence of body results and http status.
When a request to a mock endpoint matches that result for a 'n' number of times.
Then the nth result from the sequence body and status is returned for the request.

Eventually we could replace this feature with the ability to have control over the full response but keeping this simple for now.

Pseudo implementation:

Update the' Response' class to have list of responses something like 'SequenceBodies' and 'SequenceHttpStatus'

public List<string> SequenceBodies { get; set; }

public List<string> SequenceStatus { get; set; }

Then add a counter property to the' Response' class that is incremented when a response is used.

 public int SequenceUsageCount { get; set; }

Then use a new method to figure out the correct body and status sequence to use from the list.

public string GetStatus()
{
 //TODO similar to GetBody below
}

 public string GetBody()
        {
            if ( SequenceBodies != null && SequenceBodies.Any())
            {
                int total = SequenceBodies.Count();
                var index = SequenceUsageCount;

                if (index < total)
                {
                    return SequenceBodies[index];
                }
                else
                {
                    return SequenceBodies[index % total];
                }
            }
            else
            {
                return Body;
            }
        }

Would be nice to easilly configure the mock for aspnetcore integration testing, either using only the TestServer somehow, or else providing a way to configure it easilly

using var mock = new MockServer(5001); //or other fluent configuration ways
//from here mock-server is listening on a TestServer (or if not possible real loopback server) on port 5001, with all configured mocks listening in appsetting.json

using var client = new HttpClient();
var response = await client.GetAsync("http://localhost:5001/api/mocked");

TODO:

• 1. Fix up test that uses body.txt to work on windows and Linux or something else
• 2. look into critical error logs about port binding and see if we can avoid those.

- name: Test
      run: dotnet test --no-restore --verbosity normal

#4

we want to be able to simply generate random values in response body (and possibly response header fields).
an inspiration could be value placeholders here:
https://marketplace.visualstudio.com/items?itemName=humao.rest-client
(system dynamic variables section)

for example

     "myField" : "{{$guid}}"

From Uri: Uri template (path only)

We should add query string parameter and have a complete valid request parameters parsing from URI,

UriTemplate is not yet available in dotnetcore atm:
https://docs.microsoft.com/en-us/dotnet/api/system.uritemplate?view=netframework-4.8

would be nice to be able to specify json path for body rewrites, so that body rewrite can be precise ,
or a similar mechanism

https://jsonpath.com/

https://www.newtonsoft.com/json/help/html/QueryJsonSelectTokenJsonPath.htm

sample body response

{
  'Stores': [
    'Lambton Quay',
    'Willis Street'
  ],
  'Manufacturers': [
    {
      'Name': 'Acme Co',
      'Products': [
        {
          'Name': 'Anvil',
          'Price': 50
        }
      ]
    },
    {
      'Name': 'Contoso',
      'Products': [
        {
          'Name': 'Elbow Grease',
          'Price': 99.95
        },
        {
          'Name': 'Headlight Fluid',
          'Price': 4
        }
      ]
    }
  ]
}

and e.g. in the mock settings have

"Replace": {
          "UriTemplate": "test/{store}/{productName}",
          "UriJsonPathReplacements": {
            "$.Stores[0]": "{store}", //bodyValue: uriValue
            "$.Manufacturers[1].Products[0].Name": "{productName}"
          }
        }

Early on it as very easy to call the "run" command in a test project and the library would run in a background thread for you. We should add back in that behavior for folks who want to the library more inline instead of a server.

At this point it might be easier to change .Run to do what the old default behavior was and create a new on for the server behavior?

// Would run and return
KestrelMock.Run(config);

// Would run the server
KestrelMock.RunServer(config)

Thoughts?

Setup API that can be used to add or remove a mocks. This will allow mocks to be setup and removed on the fly during testing. This will give another option for how people can configure tests.

• Remove mock endpoint should take an id to lookup and remove the mock.
• Add mock endpoint should take a HttpMockSetting and return a id.
• View mock endpoint should return a list of mocks and ids.

This will require us to add ids to our HttpMockSetting object during creation. If we don't find an id in the appsetting we can auto create one.

Might be this lines actually close the stream in some context and dont allow further middleware to re-read the stream again eventually

we might neet to "reqind it" or re-seek to zero or somehow manage to make it readeable again
--> #46.

The following command fails

docker run -p 5006:5000 -e ASPNETCORE_URLS=http://*:5000 jasonrowe/kestrelmock

Cannot use file stream for [/app/KestrelMockServer.deps.json]: No such file or directory
A fatal error was encountered. The library 'libhostpolicy.so' required to execute the application was not found in '/app/'.
Failed to run as a self-contained app.

• The application was run as a self-contained app because '/app/KestrelMockServer.runtimeconfig.json' was not found.
• If this should be a framework-dependent app, add the '/app/KestrelMockServer.runtimeconfig.json' file and specify the appropriate framework.

Create a helper class for creating mocks via the API.

Include methods such as AddObservableMock, AddMock, DeleteMock, and GetMock.

Maybe we could also add the image to registry via this , @JasonRowe , what do you think about it ?
maybe we also miss a way to add meta-data to the image? not sure

https://github.com/marketplace/actions/build-and-push-docker-images

add docker image to run the mock component as standalone

moving to NET6 and NET7 could give potential benefits and add new capabilities to the mock..

JFrog Xray alerted me to the following second-order vulnerabilities in my solution based on my use of KestrelMock for testing:

In my application, i call various Rest endpoints using a third party framework called Refit
https://github.com/reactiveui/refit

When i try to mock that api using KestralMock , Upon execution the api call remains in the
Result = "Not Yet Computed" and status "WaitingForActivation" state
and hence the code throws an exception when trying to retrieve the response data.

Where as if i try to create an HttpClient similar to samples given in your code it works .

mock-server doesnt support files, so atm this solution is better in this regard! :)

mock-server/mockserver#875

Would be very cool to display a dynamic swagger/swashbuckle (openapi) page based on currently configured appsettings.json, quite challenging but would be awesome

https://docs.microsoft.com/en-us/samples/aspnet/aspnetcore.docs/getstarted-swashbuckle-aspnetcore/?tabs=visual-studio

idea for improvement 👍