api: add request/response interception API to Page (#306)

Expose the interception infrastructure that already existed internally but was
private, so embedders can intercept, observe, mock, or block JS fetch()/XHR
without forking or unsafe casts.

On obscura::Page:
- add_preload_script: inject JS before any page <script> (CDP
  Page.addScriptToEvaluateOnNewDocument). The field already ran the scripts;
  this only adds the method.
- enable_interception: returns a receiver for every JS fetch()/XHR request;
  resolve each through its resolver with InterceptResolution::{Continue,
  Fulfill, Fail} to pass, mock, or block (CDP Fetch.enable equivalent).
- on_request / on_response: passive callbacks. on_response is the main path for
  capturing API response payloads from SPAs.

on_request/on_response previously fired only for navigation; op_fetch_url now
invokes them for JS fetch()/XHR too (both reqwest and stealth paths), gated on
a non-empty callback vector so the common no-callback case pays nothing.

Re-export InterceptedRequest, InterceptResolution, RequestInfo, ResourceType,
RequestCallback, ResponseCallback, Response from the obscura crate.

Integration test tests/interception.rs exercises all three against a local echo
server.
This commit is contained in:
SGavrl committed 2026-06-24 21:36:01 +00:00
1 parent 12b927f476
commit 1f5066fc30
8 files changed
+261 -7

No files matched your search

+3
View File
@@ -20,3 +20,6 @@ thiserror = "1"
serde_json = "1"
serde = { version = "1", features = ["derive"] }
url = "2"
[dev-dependencies]
tokio = { version = "1", features = ["rt-multi-thread", "macros", "sync", "time"] }
+4
View File
@@ -26,3 +26,7 @@ pub use config::BrowserConfig;
pub use cookie::{Cookie, CookieStore};
pub use error::Error;
pub use page::Page;
// Request/response interception types (issue #306).
pub use obscura_browser::{InterceptedRequest, InterceptResolution};
pub use obscura_net::{RequestCallback, RequestInfo, ResourceType, Response, ResponseCallback};
+35 -1
View File
@@ -1,7 +1,8 @@
use std::time::Duration;
use obscura_browser::lifecycle::WaitUntil;
use obscura_browser::Page as InnerPage;
use obscura_browser::{InterceptedRequest, Page as InnerPage};
use obscura_net::{RequestCallback, ResponseCallback};
use serde_json::Value;
use crate::error::Error;
@@ -91,6 +92,39 @@ impl Page {
pub async fn settle(&mut self, max_ms: u64) {
self.inner.settle(max_ms).await
}
/// Register a script that runs before any of the page's own `<script>` tags,
/// equivalent to CDP `Page.addScriptToEvaluateOnNewDocument`. Runs on the next
/// `goto()` / navigation. Use it to install a fetch()/XHR interceptor or any
/// other page-init logic before the page's bootstrap runs.
pub fn add_preload_script(&mut self, script: &str) {
self.inner.add_preload_script(script);
}
/// Enable CDP-Fetch-style interception of every JS `fetch()`/XHR. Returns a
/// receiver yielding each request; resolve it through its `resolver` with
/// [`obscura::InterceptResolution`] (`Continue`, `Fulfill`, `Fail`) to pass,
/// mock, or block it. Works in stealth and non-stealth.
pub fn enable_interception(
&mut self,
) -> tokio::sync::mpsc::UnboundedReceiver<InterceptedRequest> {
self.inner.enable_interception()
}
/// Register a passive callback fired for every request the page makes
/// (navigation and JS `fetch()`/XHR), once its method/headers/body are known
/// and before it is sent. Non-blocking; use `enable_interception` to mutate
/// or block.
pub fn on_request(&mut self, cb: RequestCallback) {
self.inner.on_request(cb);
}
/// Register a passive callback fired with every response the page receives
/// (navigation and JS `fetch()`/XHR), including its body. Non-blocking. The
/// main path for capturing API response payloads from SPAs.
pub fn on_response(&mut self, cb: ResponseCallback) {
self.inner.on_response(cb);
}
}
/// Handle to a DOM element.
+107
View File
@@ -0,0 +1,107 @@
//! End-to-end check for the request/response interception API on `obscura::Page`
//! (issue #306): preload scripts, the interception channel, and the passive
//! on_request/on_response callbacks all work for JS-initiated fetch()/XHR.
use std::io::{Read, Write};
use std::sync::atomic::{AtomicU32, Ordering};
use std::sync::{Arc, Mutex};
use obscura::{Browser, InterceptResolution, ResourceType};
/// Minimal HTTP/1.1 server: `/` returns HTML that fires `fetch('/api')`; `/api`
/// returns JSON. Enough to exercise JS fetch() interception + the callbacks.
fn spawn_echo_server() -> String {
let listener = std::net::TcpListener::bind("127.0.0.1:0").unwrap();
let addr = listener.local_addr().unwrap();
std::thread::spawn(move || {
for incoming in listener.incoming() {
let mut s = match incoming {
Ok(s) => s,
Err(_) => continue,
};
let mut buf = [0u8; 2048];
let _ = s.read(&mut buf);
let req = std::str::from_utf8(&buf).unwrap_or("");
let path = req.split_whitespace().nth(1).unwrap_or("/");
let (ct, body) = if path.starts_with("/api") {
("application/json", "{\"hello\":\"world\"}".to_string())
} else {
("text/html", "<script>fetch('/api');</script>".to_string())
};
let resp = format!(
"HTTP/1.1 200 OK\r\nContent-Type: {}\r\nContent-Length: {}\r\nAccess-Control-Allow-Origin: *\r\nConnection: close\r\n\r\n{}",
ct,
body.len(),
body
);
let _ = s.write_all(resp.as_bytes());
let _ = s.shutdown(std::net::Shutdown::Both);
}
});
format!("http://{}", addr)
}
#[tokio::test]
async fn page_intercepts_and_observes_js_fetch() {
std::env::set_var("OBSCURA_ALLOW_PRIVATE_NETWORK", "1");
let base = spawn_echo_server();
let browser = Browser::new().unwrap();
let mut page = browser.new_page().await.unwrap();
// Passive on_request counter (fires for navigation + the JS fetch).
let req_count = Arc::new(AtomicU32::new(0));
let rc = req_count.clone();
page.on_request(Arc::new(move |_info| {
rc.fetch_add(1, Ordering::SeqCst);
}));
// Passive on_response: capture the /api (Fetch) body.
let captured = Arc::new(Mutex::new(String::new()));
let cap = captured.clone();
page.on_response(Arc::new(move |info, resp| {
if info.resource_type == ResourceType::Fetch {
*cap.lock().unwrap() = String::from_utf8_lossy(&resp.body).into_owned();
}
}));
// Channel-based interception: resolve every request Continue so the fetch
// is not blocked.
let mut rx = page.enable_interception();
tokio::spawn(async move {
while let Some(req) = rx.recv().await {
if req
.resolver
.send(InterceptResolution::Continue {
url: None,
method: None,
headers: None,
body: None,
})
.is_err()
{
break;
}
}
});
page.goto(&base).await.unwrap();
// Pump the event loop so the inline fetch('/api') resolves.
for _ in 0..20 {
page.settle(500).await;
if captured.lock().unwrap().contains("hello") {
break;
}
}
assert!(
req_count.load(Ordering::SeqCst) >= 1,
"on_request never fired for navigation or fetch"
);
let body = captured.lock().unwrap().clone();
assert!(
body.contains("hello"),
"on_response did not capture the fetch response body: {:?}",
body
);
}