timvahlbrock
commented
3 days ago The following pact-js example produces a single interaction in the contract.
import { LogLevel, PactV4, SpecificationVersion } from "@pact-foundation/pact";
import { HTTPMethods } from "@pact-foundation/pact/src/common/request";
describe("check if it generates two interactions if requests are same, but state is different", () => {
it("check if it generates two interactions if requests are same, but state is different", () => {
const pact = new PactV4({
consumer: "Consumer",
provider: "Provider",
spec: SpecificationVersion.SPECIFICATION_VERSION_V4,
logLevel: (process.env.LOG_LEVEL as LogLevel) || "warn",
});
pact.addInteraction()
.given("one state", {
value: "the first value"
})
.uponReceiving("a request")
.withRequest(HTTPMethods.GET, "/test")
.willRespondWith(200)
.executeTest(async (mockServer) => {
await fetch(mockServer.url + "/test")
})
pact.addInteraction()
.given("one state", {
value: "the second value"
})
.uponReceiving("a request")
.withRequest(HTTPMethods.GET, "/test")
.willRespondWith(200)
.executeTest(async (mockServer) => {
await fetch(mockServer.url + "/test")
})
})
})
YOU54F
I'm not sure if this is by design, but I haven't found anything on this so far. If two interactions have the same request and state description, but use different state parameter values, they are still treated the same and only one of them is added to the contract file. If this is by design, it may be helpful to get some sort of warning (if the architecture even allows to detect this), because I'm currently working on an adaption of contract testing in an existing project and just realized that two contract tests weren't saved to the contract file, as they had only differed in state parameter values.