hpx 2.5.25

High Performance HTTP Client
Documentation 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271

#![allow(missing_docs)]
//! Regression tests for large (~120 KB) response bodies.
//!
//! A user reported that responses around 120 KB could not be handled by
//! hpx. These tests pin down the expected behaviour across the transfer
//! encodings and compression codecs a real client may encounter.

mod support;

use std::io::Write;

use flate2::{Compression, write::GzEncoder};
use futures_util::stream::StreamExt;
use support::server;

/// ~120 KB payload size used across every scenario.
const PAYLOAD_LEN: usize = 120 * 1024;
const CHUNK_SIZE: usize = 4096;

fn make_payload(len: usize) -> Vec<u8> {
    // Deterministic pseudo-random bytes so compression cannot collapse them.
    let mut payload = Vec::with_capacity(len);
    let mut seed = 0x1234_5678_u32;
    while payload.len() < len {
        seed = seed.wrapping_mul(1_664_525).wrapping_add(1_013_904_223);
        payload.push((seed >> 24) as u8);
    }
    payload
}

fn assert_payload_matches(actual: &[u8]) {
    let expected = make_payload(PAYLOAD_LEN);
    assert_eq!(
        actual.len(),
        expected.len(),
        "body length mismatch: got {}, want {}",
        actual.len(),
        expected.len()
    );
    assert_eq!(actual, expected, "body content mismatch");
}

async fn collect_bytes(uri: String) -> bytes::Bytes {
    let res = hpx::get(uri).send().await.expect("response");
    assert_eq!(res.status(), 200);
    res.bytes().await.expect("body")
}

/// Serve `payload` in fixed-size chunks as a streaming body.
///
/// With a `content-length` header this exercises the framed decoder; without
/// one hyper switches to chunked transfer encoding.
fn chunked_body(payload: Vec<u8>, with_content_length: bool) -> http::Response<hpx::Body> {
    let stream = futures_util::stream::unfold((payload, 0), move |(payload, pos)| async move {
        let chunk = payload.chunks(CHUNK_SIZE).nth(pos)?.to_vec();
        Some((chunk, (payload, pos + 1)))
    })
    .map(Ok::<_, std::convert::Infallible>);

    let mut builder = http::Response::builder();
    if with_content_length {
        builder = builder.header("content-length", PAYLOAD_LEN.to_string());
    } else {
        builder = builder.header("transfer-encoding", "chunked");
    }
    builder.body(hpx::Body::wrap_stream(stream)).unwrap()
}

#[tokio::test]
async fn h1_content_length_large_body() {
    let server = server::http(move |_req| {
        let body = chunked_body(make_payload(PAYLOAD_LEN), true);
        async move { body }
    });

    let body = collect_bytes(format!("http://{}/big", server.addr())).await;
    assert_payload_matches(&body);
}

#[tokio::test]
async fn h1_chunked_large_body() {
    let server = server::http(move |_req| {
        let body = chunked_body(make_payload(PAYLOAD_LEN), false);
        async move { body }
    });

    let body = collect_bytes(format!("http://{}/chunked", server.addr())).await;
    assert_payload_matches(&body);
}

#[tokio::test]
async fn gzip_large_body() {
    let server = server::http(move |req| async move {
        assert!(
            req.headers()["accept-encoding"]
                .to_str()
                .unwrap()
                .contains("gzip"),
            "client should advertise gzip support"
        );

        let mut encoder = GzEncoder::new(Vec::new(), Compression::default());
        encoder.write_all(&make_payload(PAYLOAD_LEN)).unwrap();
        let compressed = encoder.finish().unwrap();

        let stream = futures_util::stream::unfold((compressed, 0), move |(c, pos)| async move {
            let chunk = c.chunks(CHUNK_SIZE).nth(pos)?.to_vec();
            Some((chunk, (c, pos + 1)))
        })
        .map(Ok::<_, std::convert::Infallible>);

        http::Response::builder()
            .header("content-encoding", "gzip")
            .body(hpx::Body::wrap_stream(stream))
            .unwrap()
    });

    let body = collect_bytes(format!("http://{}/gzipped", server.addr())).await;
    assert_payload_matches(&body);
}

#[tokio::test]
async fn h2_prior_knowledge_large_body() {
    // hyper-util's auto builder serves h2c when the client speaks the
    // preface; `http2_only` makes hpx open the connection with prior
    // knowledge instead of upgrading from HTTP/1.
    let server = server::http(move |_req| {
        let body = chunked_body(make_payload(PAYLOAD_LEN), true);
        async move { body }
    });

    let res = hpx::Client::builder()
        .http2_only()
        .build()
        .expect("client")
        .get(format!("http://{}/big", server.addr()))
        .send()
        .await
        .expect("response");
    assert_eq!(res.version(), hpx::Version::HTTP_2);
    let body = res.bytes().await.expect("body");
    assert_payload_matches(&body);
}

#[tokio::test]
async fn keep_alive_reuse_after_large_body() {
    // A large response followed by a second request on the same connection:
    // leftover buffered bytes must not corrupt the next response.
    let server = server::http(move |req| {
        let path = req.uri().path().to_string();
        async move {
            if path == "/big" {
                chunked_body(make_payload(PAYLOAD_LEN), true)
            } else {
                http::Response::builder()
                    .header("content-length", 5)
                    .body(hpx::Body::from("small"))
                    .unwrap()
            }
        }
    });

    let base = format!("http://{}", server.addr());
    let big = collect_bytes(format!("{base}/big")).await;
    assert_payload_matches(&big);

    // Same client => same pooled connection.
    let small = hpx::get(format!("{base}/small"))
        .send()
        .await
        .expect("second response");
    assert_eq!(small.status(), 200);
    assert_eq!(small.bytes().await.expect("body").as_ref(), b"small");
}

#[tokio::test]
async fn bytes_stream_large_body() {
    use futures_util::StreamExt;

    let server = server::http(move |_req| {
        let body = chunked_body(make_payload(PAYLOAD_LEN), true);
        async move { body }
    });

    let res = hpx::get(format!("http://{}/stream", server.addr()))
        .send()
        .await
        .expect("response");
    let mut stream = res.bytes_stream();
    let mut collected = Vec::with_capacity(PAYLOAD_LEN);
    while let Some(chunk) = stream.next().await {
        collected.extend_from_slice(&chunk.expect("chunk"));
    }
    assert_payload_matches(&collected);
}

#[tokio::test]
async fn large_response_headers() {
    // ~120 KB of response headers spread over FEW headers (under the 100
    // header count limit): a single huge value must still parse.
    let server = server::http(move |_req| async move {
        let big_value = "x".repeat(120 * 1024);
        http::Response::builder()
            .header("content-length", 5)
            .header("x-big-data", big_value.as_str())
            .body(hpx::Body::from("small"))
            .unwrap()
    });

    let res = hpx::get(format!("http://{}/big-header", server.addr()))
        .send()
        .await
        .expect("response with ~120KB single header");
    assert_eq!(res.status(), 200);
    assert_eq!(
        res.headers()["x-big-data"].len(),
        120 * 1024,
        "large single header value must survive"
    );
}

#[tokio::test]
async fn many_response_headers_over_default_limit() {
    // More than the parser's default 100-header budget: servers behind CDNs
    // or auth gateways routinely emit hundreds of Set-Cookie lines. The
    // parser must escalate its header budget automatically instead of
    // failing with `Parse(TooLarge)`.
    let cookie_count = 600;
    let server = server::http(move |_req| async move {
        let mut builder = http::Response::builder().header("content-length", 5);
        for i in 0..cookie_count {
            builder = builder.header(
                "set-cookie",
                format!("session_{i:03}=abcdef0123456789abcdef0123456789; Path=/; HttpOnly"),
            );
        }
        builder.body(hpx::Body::from("small")).unwrap()
    });

    // Default client: the parser escalates its budget transparently.
    let res = hpx::get(format!("http://{}/cookies", server.addr()))
        .send()
        .await
        .expect("header-heavy response must parse without extra configuration");
    assert_eq!(res.status(), 200);
    assert_eq!(
        res.headers().get_all("set-cookie").iter().count(),
        cookie_count
    );

    // An explicitly pinned cap still works and stays respected.
    let client = hpx::Client::builder()
        .http1_options(
            hpx::http1::Http1Options::builder()
                .max_headers(cookie_count * 2)
                .build(),
        )
        .build()
        .expect("client");
    let res = client
        .get(format!("http://{}/cookies", server.addr()))
        .send()
        .await
        .expect("response with raised max_headers");
    assert_eq!(res.status(), 200);
    assert_eq!(
        res.headers().get_all("set-cookie").iter().count(),
        cookie_count,
        "all cookies must survive"
    );
}