Making Requests¶
Contents | Previous: Installation and the First Request | Next: Request Bodies
View in another language — C++ · C#/.NET · Java · Kotlin · Node/TypeScript
After reading this chapter
You can choose an HTTP method and path, then shape one request for an external API with queries, headers, and a per-request timeout.
A request builder is created when a method is selected from the client. It accumulates the request shape until a terminator asks for a response. A terminator is the final call that finishes the builder, submits the request, and selects the response form. A path starts with / and follows the base URL.
1. Methods and Paths¶
The HTTP client provides the seven methods GET, POST, PUT, DELETE, PATCH, HEAD, and OPTIONS. GET reads a resource; POST starts work or creates one. C++ spells DELETE as delete_ because of its keyword.
A path that does not start with / ends with protocolError before a request is sent. One request keeps its base URL and path separate. Another base URL uses a separate client or a one-shot.
2. Queries, Headers, and a Timeout¶
A query accumulates names and values after URL encoding. A header applies only to that request and overrides a client default header with the same name. A per-request timeout replaces the client's default timeout only for that request.
val status =
client
.get("/ops/nodes/game-server-1/status")
.header("x-trace-id", "tutorial-1")
.timeout(Duration.ofSeconds(5))
.await<NodeStatus>()
// The admin route has another base URL, so this example uses a separate client.
val admin = zlinkHttpClient("http://127.0.0.1:5381") { basicAuth("ops", "tutorial-admin") }
try {
val weight =
admin
.post("/admin/channels/profile/weight")
.query("value", "2")
.await<WeightChanged>()
println("request shaping: status ${status.status} weight ${weight.body.weight}")
} finally {
admin.close()
}
The code uses separate base URLs for the regular and administrative APIs, adds a per-request header, timeout, and query, then reads the status and changed weight from typed responses.
3. Values That Change Per Request¶
Request-specific credentials, trace IDs, queries, and timeouts belong on the request builder. Headers and credentials that apply to every request belong on the client. A timeout is the time boundary for one attempt, and retry applies it to each attempt. Retry and response-status rules are covered in Redirect, Retry, and Cookies and Response Rules.
Next Chapter¶
Request Bodies adds JSON, form, and multipart bodies.