Tundra-Fetch

Tundra-fetch is Javascript-based client for the Tundra server that helps with offline test data management - specifically for projects using the fetch http library.

Installing the Client

npm install --save tundra-fetch

Using the Client

There are two main uses for the client:

Intercepting Fetch Calls

This client is built to intercept all uses of the fetch library and forward them to the Tundra server to be recorded for offline use. To do this, we need to initialize the interceptor when the application starts up and direct it to our server:

function someAppInitFunction() {
    require('tundra-fetch').interceptFetchCalls();
}

Intercepted calls are sent to http://localhost:9091 by default but the port is configurable by specifying it in the first parameter - e.g.interceptFetchCalls(9092).

The idea is to only enable this code before running a manual test of the app, so ensure that you remove this before going to production or is toggled off for a production build

Replay

After a profile has been captured, you now have an offline data store that can be replayed during an end-to-end test.

When and where you decide to load the profile is up to the specific needs of your application. However, getting the traffic from the profile to replay during the test execution is as simple as:

import {replayProfile} from 'tundra-fetch'
...
loadProfile() {
    replayProfile(require("./fixtures/profiles/scenario1.json"));
}

Customizing Replay Behavior

For finer-grained control over replay functionality, you may specify an optional config parameter. For example:

const config = {
  repeatMode: 'last'
}

replayProfile(require("./fixtures/profiles/scenario1.json", config));

Config options:

Option	Description
headersToOmit	An array of header keys to ignore (ex. ['Accept-Encoding', 'User-Agent'])
repeatMode	errorAfterLast (default): Replay exactly as recorded but throw an error for further requests first: Always repeat the first recorded response last: Replay exactly as recorded and then repeat the last request forever
debuggingEnabled	default: true - All profile replay activity will be sent to the Tundra server and visible in the Tundra console. (Only available using tundra-cli 2.0.0 or higher)
debugPort	default: 9091 - The port number of the Tundra server to send request debug data.

Using Wildcards to Match Requests

There are times when an application will use dynamic data and the body, headers or url will not be exactly the same each time (ex. including the current date in the request body).

For these scenarios, Tundra provides a special wildcard syntax to allow requests to be fuzzy-matched:

Matching the Body

For example, let's say you have a profile with the following layout after capturing a request:

[
  {
    "request": {
      "url": "https://www.someapi.com/user",
      "headers": { ... }
      },
      "method": "POST",
      "content": "{\"name\":\"John Doe\",\"created\":1533654607}"
    },
    "response": { ... }
  }
]

In this case, the created property in the content is just the current date, so this will change from request to request. To get around this, you can use {{*}}, which acts like a traditional wildcard matcher:

...
"content": "{\"name\":\"John Doe\",\"created\":{{*}}}"
...

Now the request will be matched, no matter what the value of created is.

Matching the URL

...
"url": "https://www.someapi.com/user?created={{*}}&name=John"
...

Matching Headers

...
"headers": {
  "SomeHeader": "Wildcards can be {{*}} anywhere"
  ...
 }
...

Working Example and More

The best way to understand this tool is to see is used in context. For a full, working example look at react-native-tundra (coming soon).

Additionally, it may be helpful to read this article on offline testing for the philosophy behind why Tundra was created: Test your mobile app offline for lightning fast, reliable automation