Skip to main content

mini_static/
server.rs

1use std::convert::Infallible;
2use std::fs;
3use std::net::SocketAddr;
4use std::path::{Path, PathBuf};
5use std::pin::Pin;
6use std::sync::Arc;
7use std::task::{Context, Poll};
8use std::time::{Duration, SystemTime};
9
10use bytes::Bytes;
11use http_body_util::Full;
12use hyper::body::Incoming;
13use hyper::http::response::Builder;
14use hyper::service::service_fn;
15use hyper::{HeaderMap, Method, Request, Response, StatusCode};
16use hyper_util::rt::TokioExecutor;
17use hyper_util::rt::TokioIo;
18use hyper_util::server::conn::auto::Builder as AutoBuilder;
19use tokio::fs::File;
20use tokio::io::{AsyncRead, AsyncReadExt, AsyncSeekExt, AsyncWrite, ReadBuf};
21use tokio::net::{TcpListener, TcpStream};
22use tokio::sync::{OwnedSemaphorePermit, Semaphore};
23use tokio::time::timeout;
24
25use crate::css::{CssOptions, CssTool};
26use crate::error::StaticError;
27use crate::handler::{FileBody, ResponseBody};
28use crate::js::{JsOptions, JsTool};
29use crate::reload::{self, SseBody};
30use crate::resolve;
31use crate::source::SourcePipeline;
32use crate::spa::{self, SpaTransition};
33use crate::tool;
34use crate::watcher::{start_watching, Broadcaster};
35
36const ACCEPT_BACKOFF_INITIAL: Duration = Duration::from_millis(10);
37const ACCEPT_BACKOFF_MAX: Duration = Duration::from_secs(1);
38
39/// Default maximum concurrent connections, overridable via `Server::with_max_connections`.
40const DEFAULT_MAX_CONNECTIONS: usize = 1024;
41
42/// A source of accepted TCP connections. Abstracted so the accept-error backoff below
43/// can be exercised against a listener that fails on demand, without needing to provoke
44/// real OS-level accept errors (e.g. EMFILE) in tests.
45trait TcpAccept {
46    async fn accept(&self) -> std::io::Result<(TcpStream, SocketAddr)>;
47}
48
49impl TcpAccept for TcpListener {
50    async fn accept(&self) -> std::io::Result<(TcpStream, SocketAddr)> {
51        TcpListener::accept(self).await
52    }
53}
54
55/// Accept a connection and reserve it a connection-limit permit.
56///
57/// `backoff` retries a failed `accept()` after an exponentially growing delay (reset on
58/// the next success, capped at `ACCEPT_BACKOFF_MAX`) instead of ending the accept loop,
59/// so a sustained failure — the process being out of file descriptors, say — degrades
60/// into periodic retries rather than a CPU-bound busy spin or a permanently dead server.
61///
62/// Returns `None` only if the semaphore itself has been closed (never happens in normal
63/// operation, since nothing ever calls `close()` on it — handled so a caller can still
64/// fail safely rather than panic).
65async fn accept_and_permit<L: TcpAccept>(
66    listener: &L,
67    backoff: &mut Duration,
68    semaphore: &Arc<Semaphore>,
69) -> Option<(TcpStream, OwnedSemaphorePermit)> {
70    loop {
71        let stream = match listener.accept().await {
72            Ok((stream, _)) => {
73                *backoff = ACCEPT_BACKOFF_INITIAL;
74                stream
75            }
76            Err(_) => {
77                tokio::time::sleep(*backoff).await;
78                *backoff = (*backoff * 2).min(ACCEPT_BACKOFF_MAX);
79                continue;
80            }
81        };
82        return semaphore
83            .clone()
84            .acquire_owned()
85            .await
86            .ok()
87            .map(|permit| (stream, permit));
88    }
89}
90
91/// A predicate deciding whether a resolved file path should get an immutable cache
92/// policy; see [`Server::with_immutable_assets`].
93type ImmutablePredicate = Arc<dyn Fn(&Path) -> bool + Send + Sync>;
94
95/// A static file server for serving files securely from a root directory.
96///
97/// `Server` canonicalizes the root directory once at creation time and uses the
98/// canonical form for all subsequent requests, avoiding repeated filesystem calls.
99///
100/// # Security
101///
102/// The server protects against:
103/// - Path traversal attacks (e.g., `../../etc/passwd`)
104/// - Accessing files outside the root via symlinks
105/// - Disclosing filesystem structure (traversal and missing files both return 404)
106///
107/// # Cloning
108///
109/// `Server` is cheap to clone: a `PathBuf`, a couple of primitives, and an `Arc`'d
110/// predicate closure. Multiple clones can be used concurrently in async tasks without
111/// synchronization overhead.
112///
113/// # Example
114///
115/// ```no_run
116/// # async fn example() -> Result<(), Box<dyn std::error::Error>> {
117/// use mini_static::Server;
118/// use std::path::Path;
119/// use std::time::Duration;
120///
121/// let server = Server::new(Path::new("./public"))?;
122/// let (port, _handle) = server.run(Duration::from_secs(30)).await?;
123/// println!("Server running on port {}", port);
124/// # Ok(())
125/// # }
126/// ```
127#[derive(Clone)]
128pub struct Server {
129    root_canon: PathBuf,
130    bundle_roots: Vec<PathBuf>,
131    max_connections: usize,
132    live_reload: bool,
133    broadcaster: Option<Broadcaster>,
134    spa_mode: bool,
135    spa_root: Option<String>,
136    spa_transition: SpaTransition,
137    immutable_predicate: Option<ImmutablePredicate>,
138    source_folders: Vec<PathBuf>,
139    asset_folders: Vec<PathBuf>,
140    output_dir: PathBuf,
141    css_tool: Option<(CssTool, CssOptions)>,
142    js_tool: Option<(JsTool, JsOptions)>,
143    prune_output: bool,
144}
145
146/// True when two canonical paths are the same path or one contains the other.
147///
148/// Used by the source/output overlap check: `Path::starts_with` compares component-wise, so
149/// `/a/b` neither equals nor contains their sibling `/a/b/c` itself. This is the single
150/// place the overlap rule is defined, so both `with_source_folder` and `with_output_dir`
151/// cannot drift apart.
152fn paths_overlap(a: &Path, b: &Path) -> bool {
153    a.starts_with(b) || b.starts_with(a)
154}
155
156impl Server {
157    /// Create a new server with the given root directory.
158    ///
159    /// Canonicalizes the root once at startup. All subsequent requests use the
160    /// canonical root without re-canonicalizing it, making this suitable for long-lived servers.
161    ///
162    /// # Errors
163    ///
164    /// Returns `Err(StaticError::Io)` if the root cannot be canonicalized (e.g., doesn't exist,
165    /// no read permissions).
166    pub fn new(root: &Path) -> Result<Self, StaticError> {
167        let root_canon = root.canonicalize().map_err(StaticError::Io)?;
168        let output_dir = root_canon.clone();
169        Ok(Server {
170            root_canon,
171            bundle_roots: Vec::new(),
172            max_connections: DEFAULT_MAX_CONNECTIONS,
173            live_reload: false,
174            broadcaster: None,
175            spa_mode: false,
176            spa_root: None,
177            spa_transition: SpaTransition::default(),
178            immutable_predicate: None,
179            source_folders: Vec::new(),
180            asset_folders: Vec::new(),
181            output_dir,
182            css_tool: None,
183            js_tool: None,
184            prune_output: false,
185        })
186    }
187
188    /// Set the maximum number of connections served concurrently (default 1024).
189    ///
190    /// Once this many connections are in flight, `run()`'s accept loop stops accepting
191    /// new ones — without pausing the accept loop, a client that opens a connection and
192    /// sends nothing (see the header-read timeout docs on [`Server::run_on`]) could
193    /// otherwise be used, in enough parallel copies, to exhaust the process's file
194    /// descriptors or memory with no bound at all.
195    pub fn with_max_connections(mut self, max: usize) -> Self {
196        self.max_connections = max;
197        self
198    }
199
200    /// Enable live-reload for this server (disabled by default).
201    ///
202    /// Once enabled, the `run*` methods start a background watcher (mtime polling,
203    /// bounded 500ms interval — see [`crate::start_watching`]) over the server's root
204    /// the first time the server actually starts accepting connections, and:
205    ///
206    /// - serve a live-reload SSE stream at [`crate::LIVE_RELOAD_PATH`], broadcasting a
207    ///   change event (with [`crate::ChangeType`]) whenever a served file is added,
208    ///   modified, or removed;
209    /// - inject a small `<script>` into every served `text/html` response that connects
210    ///   to that stream and reloads the page (or hot-swaps stylesheet `<link>`s, for CSS
211    ///   changes) — no manual client wiring required.
212    ///
213    /// This is meant for local development, not production: leave it disabled (the
214    /// default) for any server serving real traffic. A typical call site gates it behind
215    /// `#[cfg(debug_assertions)]` so a release build never pays for the watcher or the
216    /// injected script.
217    ///
218    /// # Example
219    ///
220    /// ```no_run
221    /// # fn example() -> Result<(), Box<dyn std::error::Error>> {
222    /// use mini_static::Server;
223    /// use std::path::Path;
224    ///
225    /// let server = Server::new(Path::new("./public"))?;
226    /// #[cfg(debug_assertions)]
227    /// let server = server.with_live_reload();
228    /// # Ok(())
229    /// # }
230    /// ```
231    pub fn with_live_reload(mut self) -> Self {
232        self.live_reload = true;
233        self
234    }
235
236    /// Enable spa-mode navigation for this server, swapping `document.body` on
237    /// each navigation (disabled by default).
238    ///
239    /// Once enabled, every served `text/html` response gets a small `<script>`
240    /// injected (see [`Server::with_spa_root`] for what it does) that treats
241    /// `document.body` as the swap target. Calling this after
242    /// [`Server::with_spa_root`] does not clear a previously configured root
243    /// selector — the two methods set independent fields, so
244    /// `.with_spa_root(sel).with_spa_mode()` and
245    /// `.with_spa_mode().with_spa_root(sel)` both end up with spa-mode on and
246    /// root `sel`. Use this one alone when there's no persistent chrome to
247    /// preserve across navigations.
248    ///
249    /// # Example
250    ///
251    /// ```no_run
252    /// # fn example() -> Result<(), Box<dyn std::error::Error>> {
253    /// use mini_static::Server;
254    /// use std::path::Path;
255    ///
256    /// let server = Server::new(Path::new("./public"))?.with_spa_mode();
257    /// # Ok(())
258    /// # }
259    /// ```
260    pub fn with_spa_mode(mut self) -> Self {
261        self.spa_mode = true;
262        self
263    }
264
265    /// Enable spa-mode navigation for this server, swapping only the element
266    /// matched by the CSS `selector` on each navigation (disabled by default;
267    /// also enables spa-mode the same as [`Server::with_spa_mode`]).
268    ///
269    /// Once enabled, every served `text/html` response gets a small `<script>`
270    /// injected that intercepts left-clicks on same-origin `<a href>`
271    /// elements (skipping links with a non-`_self` `target`, a `download`
272    /// attribute, `rel="external"`, a `data-no-spa` attribute, or a same-page
273    /// hash-only href) and, instead of a normal navigation:
274    ///
275    /// - fetches the target URL;
276    /// - on a non-OK or non-`text/html` response (or a fetch error), falls
277    ///   back to a real `location.href` navigation — spa-mode never renders a
278    ///   broken page;
279    /// - otherwise replaces the matched element's `innerHTML` with the
280    ///   corresponding content from the fetched document, updates the page
281    ///   title, and pushes the new URL via `history.pushState`, animating the
282    ///   swap with `document.startViewTransition()` where supported;
283    /// - dispatches a `mini-static:navigate` `CustomEvent` on `window` after
284    ///   every client-side navigation, so page scripts can re-run any
285    ///   per-page initialization that would otherwise only execute once
286    ///   (content swapped in via `innerHTML` never executes its own
287    ///   `<script>` tags);
288    /// - handles browser back/forward by re-fetching and swapping to the new
289    ///   `location.href`.
290    ///
291    /// `selector` is matched against both the current page and the fetched
292    /// page; a link click where the selector matches neither falls back to a
293    /// real navigation, same as a fetch failure. Choose a `selector` that
294    /// wraps only the content that varies between pages, leaving persistent
295    /// chrome (nav/header/footer) outside it so it survives navigation
296    /// untouched.
297    ///
298    /// This is meant to be usable in production, not just local development
299    /// (unlike [`Server::with_live_reload`]): a click on a link mini-static
300    /// doesn't intercept, or on a browser without JS or View Transitions
301    /// support, still works as a normal navigation.
302    ///
303    /// # Example
304    ///
305    /// ```no_run
306    /// # fn example() -> Result<(), Box<dyn std::error::Error>> {
307    /// use mini_static::Server;
308    /// use std::path::Path;
309    ///
310    /// let server = Server::new(Path::new("./public"))?.with_spa_root("#app");
311    /// # Ok(())
312    /// # }
313    /// ```
314    pub fn with_spa_root(mut self, selector: &str) -> Self {
315        self.spa_mode = true;
316        self.spa_root = Some(selector.to_string());
317        self
318    }
319
320    /// Set how spa-mode animates the swap between pages (also enables
321    /// spa-mode the same as [`Server::with_spa_mode`]; default
322    /// [`SpaTransition::Fade`] when spa-mode is enabled without calling this).
323    ///
324    /// [`SpaTransition::Slide`] injects its own `<style>` tag alongside the
325    /// spa-mode `<script>` — no site CSS is required. See [`SpaTransition`]
326    /// and [`SlideOptions`] for what each variant does and how to
327    /// configure the slide's duration, direction, and easing.
328    ///
329    /// # Example
330    ///
331    /// ```no_run
332    /// # fn example() -> Result<(), Box<dyn std::error::Error>> {
333    /// use mini_static::{Server, SlideOptions, SpaTransition};
334    /// use std::path::Path;
335    ///
336    /// let server = Server::new(Path::new("./public"))?
337    ///     .with_spa_root("#app")
338    ///     .with_spa_transition(SpaTransition::Slide(
339    ///         SlideOptions::default().duration_ms(500),
340    ///     ));
341    /// # Ok(())
342    /// # }
343    /// ```
344    pub fn with_spa_transition(mut self, transition: SpaTransition) -> Self {
345        self.spa_mode = true;
346        self.spa_transition = transition;
347        self
348    }
349
350    /// Serve files matching `predicate` with a long-lived, immutable cache policy
351    /// instead of the default `Cache-Control: no-cache`.
352    ///
353    /// `predicate` is evaluated against each resolved file's path; a match sends
354    /// `Cache-Control: public, max-age=31536000, immutable` on that file's 200 and 304
355    /// responses. This is correct only for fingerprinted assets (e.g.
356    /// `main.a1b2c3.js`) where a content change always produces a new filename —
357    /// caching a mutable filename indefinitely would serve stale content to every
358    /// client that already has it cached.
359    ///
360    /// # Example
361    ///
362    /// ```no_run
363    /// # fn example() -> Result<(), Box<dyn std::error::Error>> {
364    /// use mini_static::Server;
365    /// use std::path::Path;
366    ///
367    /// let server = Server::new(Path::new("./public"))?
368    ///     .with_immutable_assets(|path| {
369    ///         path.file_name()
370    ///             .and_then(|name| name.to_str())
371    ///             .is_some_and(|name| name.contains(".fingerprint."))
372    ///     });
373    /// # Ok(())
374    /// # }
375    /// ```
376    pub fn with_immutable_assets<F>(mut self, predicate: F) -> Self
377    where
378        F: Fn(&Path) -> bool + Send + Sync + 'static,
379    {
380        self.immutable_predicate = Some(Arc::new(predicate));
381        self
382    }
383
384    /// The `Cache-Control` header value for a resolved file path: the immutable policy
385    /// if `with_immutable_assets`'s predicate matches, `no-cache` otherwise.
386    fn cache_control_for(&self, path: &Path) -> &'static str {
387        match &self.immutable_predicate {
388            Some(predicate) if predicate(path) => "public, max-age=31536000, immutable",
389            _ => "no-cache",
390        }
391    }
392
393    /// Register `path` as an additional directory whose changes should trigger a CSS
394    /// bundle rebuild, alongside the registered source folders.
395    ///
396    /// Useful for build pipelines where CSS partials referenced via `@import` live in a
397    /// separate directory tree from the source folders proper: without registering that
398    /// tree here, editing a partial wouldn't be noticed by the watcher and the bundle
399    /// would go stale until something else touched it.
400    ///
401    /// Files under `path` are never directly HTTP-servable: `Server::resolve` and the
402    /// request-handling path never consult bundle roots. This is purely a watch target,
403    /// not a second served root, and — since `@import` resolution is delegated entirely
404    /// to the configured [`CssTool`] (see [`Server::with_css_tool`]) — not an `@import`
405    /// traversal boundary either; the external tool resolves its own imports with no
406    /// root mini-static can enforce.
407    ///
408    /// This method is fallible and canonicalizes the path once at call time, matching
409    /// `Server::new`'s canonicalize-once policy. Call it multiple times to register
410    /// more than one external source tree.
411    ///
412    /// # Errors
413    ///
414    /// Returns `Err(StaticError::Io)` if the path cannot be canonicalized.
415    pub fn with_bundle_root(mut self, path: &Path) -> Result<Self, StaticError> {
416        let canon = path.canonicalize().map_err(StaticError::Io)?;
417        self.bundle_roots.push(canon);
418        Ok(self)
419    }
420
421    /// Designate `dir` as a source folder whose changes drive the build pipelines.
422    ///
423    /// Watched when `with_live_reload()` is enabled; `.css` files under it feed the single
424    /// CSS bundle, `.js`/`.mjs` files are minified per-file into the output dir.
425    ///
426    /// Rejected if `dir` overlaps the output dir or an already-registered source folder: a
427    /// source folder that is also the output would feed every pipeline its own output — the
428    /// feedback loop this layering exists to prevent.
429    ///
430    /// # Errors
431    ///
432    /// Returns `Err(StaticError::Io)` if `dir` cannot be canonicalized, or
433    /// `Err(StaticError::Traversal)` if it overlaps the output dir or another source folder.
434    pub fn with_source_folder(mut self, dir: &Path) -> Result<Self, StaticError> {
435        let canon = dir.canonicalize().map_err(StaticError::Io)?;
436
437        if paths_overlap(&canon, &self.output_dir) {
438            return Err(StaticError::Traversal(format!(
439                "source folder {} overlaps the output dir {}",
440                canon.display(),
441                self.output_dir.display()
442            )));
443        }
444        if self
445            .source_folders
446            .iter()
447            .chain(self.asset_folders.iter())
448            .any(|existing| paths_overlap(&canon, existing))
449        {
450            return Err(StaticError::Traversal(format!(
451                "source folder {} overlaps an already-registered source/asset folder",
452                canon.display()
453            )));
454        }
455
456        self.source_folders.push(canon);
457        Ok(self)
458    }
459
460    /// Designate `dir` as an asset source folder: every file under it (any extension)
461    /// is mirrored byte-identical into the output dir at server startup and on every
462    /// live-reload change — no CSS/JS transformation, just a flat copy preserving each
463    /// file's path relative to `dir`. Use this for hand-authored static files
464    /// (`index.html`, images) that should live outside the served/output dir as
465    /// source, the same source/output separation `with_source_folder`'s CSS/JS
466    /// pipelines already have.
467    ///
468    /// Rejected if `dir` overlaps the output dir or an already-registered
469    /// source/asset folder, for the same reason `with_source_folder` rejects it: a
470    /// folder that is also the output would feed the pipeline its own output.
471    ///
472    /// # Errors
473    ///
474    /// Returns `Err(StaticError::Io)` if `dir` cannot be canonicalized, or
475    /// `Err(StaticError::Traversal)` if it overlaps the output dir or another
476    /// registered source/asset folder.
477    pub fn with_asset_folder(mut self, dir: &Path) -> Result<Self, StaticError> {
478        let canon = dir.canonicalize().map_err(StaticError::Io)?;
479
480        if paths_overlap(&canon, &self.output_dir) {
481            return Err(StaticError::Traversal(format!(
482                "asset folder {} overlaps the output dir {}",
483                canon.display(),
484                self.output_dir.display()
485            )));
486        }
487        if self
488            .source_folders
489            .iter()
490            .chain(self.asset_folders.iter())
491            .any(|existing| paths_overlap(&canon, existing))
492        {
493            return Err(StaticError::Traversal(format!(
494                "asset folder {} overlaps an already-registered source/asset folder",
495                canon.display()
496            )));
497        }
498
499        self.asset_folders.push(canon);
500        Ok(self)
501    }
502
503    /// Designate `dir` as the output directory processed outputs are written to.
504    ///
505    /// Defaults to the served root. The output dir is never a watcher trigger: pipelines
506    /// react to source folders only, so a pipeline's own output can never re-trigger it.
507    /// Call this before `with_css_tool` so a bundle output path reflects the override.
508    ///
509    /// # Errors
510    ///
511    /// Returns `Err(StaticError::Io)` if `dir` cannot be canonicalized, or
512    /// `Err(StaticError::Traversal)` if it overlaps a registered source or asset folder.
513    pub fn with_output_dir(mut self, dir: &Path) -> Result<Self, StaticError> {
514        let canon = dir.canonicalize().map_err(StaticError::Io)?;
515
516        if self
517            .source_folders
518            .iter()
519            .chain(self.asset_folders.iter())
520            .any(|existing| paths_overlap(&canon, existing))
521        {
522            return Err(StaticError::Traversal(format!(
523                "output dir {} overlaps a registered source/asset folder",
524                canon.display()
525            )));
526        }
527
528        self.output_dir = canon;
529        Ok(self)
530    }
531
532    /// Configure CSS bundling/minification via an external tool (disabled by default).
533    ///
534    /// `tool` is a preset naming the CLI mini-static invokes (see [`CssTool`]) —
535    /// mini-static does not install or manage the binary, only looks it up on `PATH`;
536    /// [`Server::run_on`] fails fast at startup if it's missing. `options` selects
537    /// `bundle`/`minify` independently (see [`CssOptions`]):
538    ///
539    /// - Neither: every `.css` under the source folders is copied through unchanged,
540    ///   mirrored into the output dir.
541    /// - `minify` only: each file is minified independently and mirrored (no `@import`
542    ///   following).
543    /// - `bundle` only: every `.css` under the source folders is discovered,
544    ///   `@import`-resolved, and concatenated into one output file, unminified.
545    /// - Both: the bundle above, minified.
546    ///
547    /// # Example
548    ///
549    /// ```no_run
550    /// # fn example() -> Result<(), Box<dyn std::error::Error>> {
551    /// use mini_static::{CssOptions, CssTool, Server};
552    /// use std::path::Path;
553    ///
554    /// let server = Server::new(Path::new("./public"))?
555    ///     .with_css_tool(CssTool::LightningCss, CssOptions::new().bundle(true).minify(true));
556    /// # Ok(())
557    /// # }
558    /// ```
559    pub fn with_css_tool(mut self, tool: CssTool, options: CssOptions) -> Self {
560        self.css_tool = Some((tool, options));
561        self
562    }
563
564    /// Configure JS bundling/minification via an external tool (disabled by default).
565    ///
566    /// `tool` is a preset naming the CLI mini-static invokes (see [`JsTool`]) —
567    /// mini-static does not install or manage the binary, only looks it up on `PATH`;
568    /// [`Server::run_on`] fails fast at startup if it's missing. Unlike CSS, JS bundling
569    /// requires an explicit entry point ([`JsOptions::bundle_entry`]) since a JS module
570    /// graph has no well-defined "concatenate everything" meaning; without it, `options`
571    /// runs in per-file mode (every `.js`/`.mjs` under the source folders processed and
572    /// mirrored independently).
573    ///
574    /// # Errors
575    ///
576    /// Returns `Err(StaticError::Io)` if `options` specifies a bundle entry that cannot
577    /// be canonicalized, or `Err(StaticError::Traversal)` if it doesn't lie under a
578    /// registered source folder — checked eagerly here so a bad entry path fails at
579    /// configuration time, not on the first rebuild.
580    ///
581    /// # Example
582    ///
583    /// ```no_run
584    /// # fn example() -> Result<(), Box<dyn std::error::Error>> {
585    /// use mini_static::{JsOptions, JsTool, Server};
586    /// use std::path::Path;
587    ///
588    /// let server = Server::new(Path::new("./public"))?
589    ///     .with_source_folder(Path::new("./js-src"))?
590    ///     .with_js_tool(
591    ///         JsTool::Esbuild,
592    ///         JsOptions::new()
593    ///             .bundle_entry(Path::new("./js-src/main.js"), "bundle.js")
594    ///             .minify(true),
595    ///     )?;
596    /// # Ok(())
597    /// # }
598    /// ```
599    pub fn with_js_tool(mut self, tool: JsTool, options: JsOptions) -> Result<Self, StaticError> {
600        if let Some(entry) = options.entry() {
601            let entry_canon = entry.canonicalize().map_err(StaticError::Io)?;
602            let under_source_folder = self
603                .source_folders
604                .iter()
605                .any(|folder| entry_canon.starts_with(folder));
606            if !under_source_folder {
607                return Err(StaticError::Traversal(format!(
608                    "js bundle entry {} is not under any registered source folder",
609                    entry_canon.display()
610                )));
611            }
612        }
613
614        self.js_tool = Some((tool, options));
615        Ok(self)
616    }
617
618    /// Remove stale CSS bundle output at build time — specifically, delete the bundle file
619    /// when no CSS sources remain, rather than serving an orphan. Applies only to the
620    /// one-shot startup build, never during live-reload.
621    pub fn with_prune_output(mut self) -> Self {
622        self.prune_output = true;
623        self
624    }
625
626    /// True when any build pipeline is configured (a CSS/JS tool and/or a source
627    /// folder), i.e. the server should run a startup build.
628    fn has_pipeline(&self) -> bool {
629        self.css_tool.is_some()
630            || self.js_tool.is_some()
631            || !self.source_folders.is_empty()
632            || !self.asset_folders.is_empty()
633    }
634
635    /// Every external tool binary this configuration actually needs at some point
636    /// (bundle and/or minify enabled — a pure passthrough config never spawns its
637    /// configured tool, so it has nothing to fail-fast on), paired with its
638    /// human-readable install hint for a fail-fast startup error.
639    fn required_tool_binaries(&self) -> Vec<(&'static str, &'static str)> {
640        let mut required = Vec::new();
641        if let Some((css_tool, options)) = &self.css_tool {
642            if options.is_bundle() || options.is_minify() {
643                required.push((css_tool.binary_name(), css_tool.install_hint()));
644            }
645        }
646        if let Some((js_tool, options)) = &self.js_tool {
647            if options.is_bundle() || options.is_minify() {
648                required.push((js_tool.binary_name(), js_tool.install_hint()));
649            }
650        }
651        required
652    }
653
654    /// Every directory to watch for source changes: the source folders, the CSS
655    /// `@import` roots, and the asset folders, deduplicated so a directory registered
656    /// under more than one role is watched once.
657    fn watch_targets(&self) -> Vec<PathBuf> {
658        let mut targets = Vec::new();
659        for dir in self
660            .source_folders
661            .iter()
662            .chain(self.bundle_roots.iter())
663            .chain(self.asset_folders.iter())
664        {
665            if !targets.contains(dir) {
666                targets.push(dir.clone());
667            }
668        }
669        targets
670    }
671
672    /// Run every configured build pipeline (CSS/JS tools, asset folders) once and
673    /// return, without starting the HTTP server. A one-shot equivalent of the
674    /// startup build `run*` does automatically — for deploy tooling that wants to
675    /// populate the output dir ahead of time (e.g. a `cargo run --bin build_static`
676    /// step before baking a Docker image), mirroring a one-shot content
677    /// builder's `build()` (e.g. `mini_docs::Builder::build()`).
678    ///
679    /// # Errors
680    ///
681    /// - `Err(StaticError::PipelineSetup)` if a configured tool's binary that's
682    ///   actually needed (bundle or minify enabled) is missing from `PATH` — checked
683    ///   before anything runs, same as [`Server::run_on`].
684    /// - `Err(StaticError::Build)` if a configured pipeline step fails (a tool
685    ///   invocation error, a filesystem error writing output, etc.).
686    ///
687    /// # Example
688    ///
689    /// ```no_run
690    /// # async fn example() -> Result<(), Box<dyn std::error::Error>> {
691    /// use mini_static::Server;
692    /// use std::path::Path;
693    ///
694    /// let server = Server::new(Path::new("./public"))?;
695    /// server.build().await?;
696    /// # Ok(())
697    /// # }
698    /// ```
699    pub async fn build(&self) -> Result<(), StaticError> {
700        for (binary, install_hint) in self.required_tool_binaries() {
701            if !tool::locate_on_path(binary) {
702                return Err(StaticError::PipelineSetup(format!(
703                    "{binary} not found on PATH ({install_hint})"
704                )));
705            }
706        }
707
708        let pipeline = SourcePipeline::new(
709            self.source_folders.clone(),
710            self.bundle_roots.clone(),
711            self.asset_folders.clone(),
712            self.output_dir.clone(),
713            self.css_tool.clone(),
714            self.js_tool.clone(),
715            self.prune_output,
716            Broadcaster::new(),
717        );
718        pipeline
719            .full_build()
720            .await
721            .map_err(|e| StaticError::Build(e.to_string()))
722    }
723
724    /// Resolve a request path under the server's root.
725    ///
726    /// This is a lower-level API for resolving paths without generating HTTP responses.
727    /// For most use cases, prefer [`Server::handle_request`] or the `run*` methods.
728    ///
729    /// # Returns
730    ///
731    /// - `Ok(PathBuf)` if the path resolves to a file within root.
732    /// - `Err(StaticError)` if the path is invalid, missing, or attempts traversal.
733    pub fn resolve(&self, request_path: &str) -> Result<PathBuf, StaticError> {
734        resolve::resolve_with_canonical_root(&self.root_canon, request_path)
735    }
736
737    /// Run the server on a specific address with a configurable header-read timeout.
738    ///
739    /// Spawns the server in a background Tokio task and returns immediately with the
740    /// assigned port number and a [`ServerHandle`]. Call `handle.shutdown().await` to
741    /// stop accepting new connections and wait for in-flight connections to finish.
742    /// Dropping the handle instead leaves the server running for the life of the process.
743    ///
744    /// # Header-Read Timeout
745    ///
746    /// Connections that don't send complete HTTP headers within `header_timeout` are closed.
747    /// This prevents slowloris attacks and resource exhaustion from incomplete requests. The
748    /// timeout applies only to the header-read phase — once a complete header block has been
749    /// read, the connection is handed off with no further time bound, so long-lived response
750    /// bodies (e.g. the live-reload SSE stream from [`Server::with_live_reload`]) are not cut
751    /// off mid-stream.
752    ///
753    /// # Precompressed Sidecars
754    ///
755    /// If a request's `Accept-Encoding` allows `br` or `gzip` (preferring `br`) and a
756    /// sibling `<path>.br`/`<path>.gz` exists next to the resolved file, its bytes are
757    /// served instead with a matching `Content-Encoding`. Every file response carries
758    /// `Vary: Accept-Encoding` so intermediate caches don't serve the wrong variant to a
759    /// differently-capable client.
760    ///
761    /// # Arguments
762    ///
763    /// * `addr` - Socket address to bind to (e.g., `127.0.0.1:0` for loopback ephemeral,
764    ///   or `0.0.0.0:8080` to bind all interfaces on a fixed port).
765    /// * `header_timeout` - Maximum time to wait for complete HTTP headers on each connection.
766    ///
767    /// # Returns
768    ///
769    /// - `Ok((u16, ServerHandle))` with the assigned port number and a handle for graceful shutdown.
770    /// - `Err(StaticError::Io)` if binding to the socket fails.
771    /// - `Err(StaticError::PipelineSetup)` if a configured [`CssTool`]/[`JsTool`]'s binary
772    ///   is not found on `PATH`. Checked before the listener binds: a deployment whose
773    ///   configured pipeline can never run should fail visibly at boot, not be discovered
774    ///   later as a missing/stale asset.
775    pub async fn run_on(
776        &self,
777        addr: SocketAddr,
778        header_timeout: Duration,
779    ) -> Result<(u16, ServerHandle), StaticError> {
780        for (binary, install_hint) in self.required_tool_binaries() {
781            if !tool::locate_on_path(binary) {
782                return Err(StaticError::PipelineSetup(format!(
783                    "{binary} not found on PATH ({install_hint})"
784                )));
785            }
786        }
787
788        let listener = TcpListener::bind(addr).await.map_err(StaticError::Io)?;
789        let port = listener.local_addr().map_err(StaticError::Io)?.port();
790
791        let mut server = self.clone();
792        if server.live_reload {
793            let broadcaster = Broadcaster::new();
794
795            // The build pipelines react to SOURCE folders only; the output dir is never
796            // watched. Watching the output would feed each pipeline its own writes back
797            // into its trigger — the feedback loop this layering exists to prevent.
798            if server.has_pipeline() {
799                let pipeline = Arc::new(SourcePipeline::new(
800                    server.source_folders.clone(),
801                    server.bundle_roots.clone(),
802                    server.asset_folders.clone(),
803                    server.output_dir.clone(),
804                    server.css_tool.clone(),
805                    server.js_tool.clone(),
806                    server.prune_output,
807                    broadcaster.clone(),
808                ));
809                let mut rx = broadcaster.subscribe();
810                tokio::spawn(async move {
811                    // One-shot startup build (and optional prune) first, so the earliest
812                    // request already sees fresh output rather than yesterday's.
813                    if let Err(e) = pipeline.full_build().await {
814                        eprintln!("source pipeline build error: {e}");
815                    }
816                    while let Some(event) = rx.recv().await {
817                        if let Err(e) = pipeline
818                            .process_change(&event.path, &event.change_type)
819                            .await
820                        {
821                            eprintln!("source pipeline error: {e}");
822                        }
823                    }
824                });
825            }
826
827            for dir in server.watch_targets() {
828                start_watching(Arc::new(dir), broadcaster.clone());
829            }
830
831            server.broadcaster = Some(broadcaster);
832        } else if server.has_pipeline() {
833            // No live-reload: still run the one-shot build so a release boot reflects the
834            // current sources. The broadcaster is a throwaway — there is no browser to
835            // notify, so broadcasting into it is a no-op.
836            let pipeline = Arc::new(SourcePipeline::new(
837                server.source_folders.clone(),
838                server.bundle_roots.clone(),
839                server.asset_folders.clone(),
840                server.output_dir.clone(),
841                server.css_tool.clone(),
842                server.js_tool.clone(),
843                server.prune_output,
844                Broadcaster::new(),
845            ));
846            tokio::spawn(async move {
847                if let Err(e) = pipeline.full_build().await {
848                    eprintln!("source pipeline build error: {e}");
849                }
850            });
851        }
852        let semaphore = Arc::new(Semaphore::new(server.max_connections));
853        let (shutdown_tx, shutdown_rx) = tokio::sync::oneshot::channel();
854
855        let accept_task = tokio::spawn(async move {
856            let mut backoff = ACCEPT_BACKOFF_INITIAL;
857            let mut join_set: tokio::task::JoinSet<()> = tokio::task::JoinSet::new();
858            let mut shutdown_pin = std::pin::pin!(shutdown_rx);
859            let mut shutting_down = false;
860
861            loop {
862                if !shutting_down {
863                    // The accept-and-permit step and the shutdown signal race in a single
864                    // `select!` so shutdown can preempt a pending accept or a permit wait
865                    // cleanly, at any point — not just between loop iterations.
866                    tokio::select! {
867                        accepted = accept_and_permit(&listener, &mut backoff, &semaphore) => {
868                            match accepted {
869                                Some((stream, permit)) => {
870                                    let server = server.clone();
871                                    join_set.spawn(async move {
872                                        let _permit = permit;
873                                        serve_connection(stream, server, header_timeout).await;
874                                    });
875                                }
876                                None => shutting_down = true,
877                            }
878                        }
879                        _ = shutdown_pin.as_mut() => {
880                            shutting_down = true;
881                        }
882                    }
883                    continue;
884                }
885
886                // Stop accepting; drain already-spawned connections before returning.
887                match join_set.join_next().await {
888                    Some(_) => continue,
889                    None => break,
890                }
891            }
892        });
893
894        Ok((
895            port,
896            ServerHandle {
897                shutdown_tx: Some(shutdown_tx),
898                accept_task,
899            },
900        ))
901    }
902
903    /// Run the server on loopback (127.0.0.1), binding an ephemeral port.
904    ///
905    /// Thin wrapper around [`Server::run_on`] — see it for the header-read timeout and
906    /// sidecar semantics, and for what the returned [`ServerHandle`] does.
907    pub async fn run(&self, header_timeout: Duration) -> Result<(u16, ServerHandle), StaticError> {
908        self.run_on(([127, 0, 0, 1], 0).into(), header_timeout)
909            .await
910    }
911
912    /// Run the server on all interfaces (0.0.0.0) at `port` (0 for an ephemeral port).
913    ///
914    /// Useful for containerized deployments and reverse-proxy setups. Thin wrapper
915    /// around [`Server::run_on`] — see it for the header-read timeout and sidecar
916    /// semantics, and for what the returned [`ServerHandle`] does.
917    pub async fn run_all(
918        &self,
919        port: u16,
920        header_timeout: Duration,
921    ) -> Result<(u16, ServerHandle), StaticError> {
922        self.run_on(([0, 0, 0, 0], port).into(), header_timeout)
923            .await
924    }
925
926    /// Run the server on loopback with the default 30-second header-read timeout.
927    ///
928    /// The recommended entry point for tests and lightweight services that don't need a
929    /// custom timeout. Thin wrapper around [`Server::run`].
930    ///
931    /// # Example
932    ///
933    /// ```no_run
934    /// # async fn example() -> Result<(), Box<dyn std::error::Error>> {
935    /// use mini_static::Server;
936    /// use std::path::Path;
937    ///
938    /// let server = Server::new(Path::new("./public"))?;
939    /// let (port, handle) = server.run_ephemeral().await?;
940    /// println!("Server ready on http://127.0.0.1:{}", port);
941    /// handle.shutdown().await;
942    /// # Ok(())
943    /// # }
944    /// ```
945    pub async fn run_ephemeral(&self) -> Result<(u16, ServerHandle), StaticError> {
946        self.run(DEFAULT_HEADER_TIMEOUT).await
947    }
948
949    /// Produce the HTTP response for a request, streaming file bodies to the client.
950    ///
951    /// This is the crate's single request-handling path: the `run*` accept loop calls it,
952    /// and so should any async server embedding `mini-static` as a fallback route (e.g.
953    /// `mini-unified`). It never blocks the calling task — path resolution runs on Tokio's
954    /// blocking-thread pool via `spawn_blocking`, and the file is read via async I/O.
955    ///
956    /// File responses are backed by `FileBody`, which hands hyper one 64 KB chunk at a
957    /// time as `poll_frame` is driven: memory use stays bounded to one chunk per in-flight
958    /// response regardless of file size.
959    ///
960    /// `headers` are the request's headers; `If-None-Match` (304 on a matching ETag) and
961    /// `Accept-Encoding` (precompressed sidecar selection, see [`Server::run_on`]) are the
962    /// ones read today. Only `GET` and `HEAD` are allowed; anything else gets a 405 with an
963    /// `Allow` header. Missing files and traversal attempts both get an identical 404, so a
964    /// response never discloses whether a path exists outside the root.
965    pub async fn handle_request(
966        &self,
967        method: &Method,
968        request_path: &str,
969        headers: &HeaderMap,
970    ) -> Response<ResponseBody> {
971        if method != Method::GET && method != Method::HEAD {
972            return text(
973                response(StatusCode::METHOD_NOT_ALLOWED).header("Allow", "GET, HEAD"),
974                "method not allowed\n",
975            );
976        }
977
978        // Live-reload SSE stream — only reachable when `with_live_reload()` was called
979        // and the server was started via a `run*` method (those are the only paths that
980        // populate `broadcaster`).
981        if *method == Method::GET && request_path == reload::LIVE_RELOAD_PATH {
982            if let Some(broadcaster) = &self.broadcaster {
983                return finish(
984                    response(StatusCode::OK)
985                        .header("Content-Type", "text/event-stream")
986                        .header("Cache-Control", "no-cache")
987                        .header("Connection", "keep-alive")
988                        .body(ResponseBody::Sse(SseBody::new(broadcaster.subscribe()))),
989                );
990            }
991        }
992
993        // `resolve()` does blocking filesystem syscalls (`canonicalize()`, up to two per
994        // request). Running those directly in this `async fn` would block whichever
995        // Tokio worker thread happens to be driving it, stalling every other task
996        // scheduled on that thread for the duration of the syscalls. `spawn_blocking`
997        // moves the work onto Tokio's dedicated blocking thread pool instead.
998        let server = self.clone();
999        let owned_request_path = request_path.to_string();
1000        let resolved =
1001            tokio::task::spawn_blocking(move || server.resolve(&owned_request_path)).await;
1002        let path = match resolved {
1003            Err(_) => return internal_error_response(),
1004            Ok(Err(e)) => {
1005                return text(
1006                    response(StatusCode::NOT_FOUND),
1007                    format!("{}\n", e.user_message()),
1008                )
1009            }
1010            Ok(Ok(path)) => path,
1011        };
1012
1013        // A directory served via its `index.html` needs a trailing slash to establish the
1014        // correct base for the page's relative links. Compare against the *decoded*
1015        // request path so a percent-encoded explicit request for index.html (e.g.
1016        // `/docs/index.htm%6c`) is recognized as such instead of producing a redirect to a
1017        // still-encoded, broken Location.
1018        let decoded_request_path = resolve::decode_request_path(request_path);
1019        if path.file_name().is_some_and(|name| name == "index.html")
1020            && !decoded_request_path.ends_with('/')
1021            && !decoded_request_path.ends_with("index.html")
1022        {
1023            // `location` is built from the (attacker-controlled) request path; `finish()`
1024            // degrades to 400 instead of panicking if it ever contains bytes invalid in a
1025            // header value.
1026            let location = format!("{}/", request_path.trim_end_matches('/'));
1027            return text(
1028                response(StatusCode::MOVED_PERMANENTLY).header("Location", location),
1029                "moved\n",
1030            );
1031        }
1032
1033        let Ok(file) = File::open(&path).await else {
1034            return internal_error_response();
1035        };
1036        let Ok(metadata) = file.metadata().await else {
1037            return internal_error_response();
1038        };
1039
1040        let content_type = mime_type_for_path(&path);
1041        // Live-reload and spa-mode HTML injection both need the original, uncompressed
1042        // bytes to splice their script into — never substitute a precompressed sidecar on
1043        // this path. `broadcaster` is only `Some` when live-reload is enabled (see
1044        // `Server::with_live_reload`); `spa_mode` is independent of it (see
1045        // `Server::with_spa_mode`/`with_spa_root`) — either alone is enough to trigger
1046        // injection.
1047        let html_injection =
1048            (self.broadcaster.is_some() || self.spa_mode) && content_type.starts_with("text/html");
1049
1050        let range_header = header_str(headers, "range");
1051        let if_range_header = header_str(headers, "if-range");
1052
1053        let accept_encoding = header_str(headers, "accept-encoding");
1054        // Skip precompressed sidecars when Range is requested (serve original file instead).
1055        let sidecar = if html_injection || range_header.is_some() {
1056            None
1057        } else {
1058            select_precompressed_sidecar(&path, accept_encoding).await
1059        };
1060        let (mut file, metadata, content_encoding) = match sidecar {
1061            Some((sidecar_file, sidecar_metadata, encoding)) => {
1062                (sidecar_file, sidecar_metadata, Some(encoding))
1063            }
1064            None => (file, metadata, None),
1065        };
1066
1067        // HTML injection is skipped for a served precompressed sidecar (already final
1068        // bytes from a build step) — see `html_injection`'s definition above.
1069        let etag = generate_etag(&metadata);
1070        let cache_control = self.cache_control_for(&path);
1071
1072        if header_str(headers, "if-none-match").is_some_and(|value| is_etag_match(value, &etag)) {
1073            return finish(
1074                Response::builder()
1075                    .status(StatusCode::NOT_MODIFIED)
1076                    .header("Cache-Control", cache_control)
1077                    .header("Vary", "Accept-Encoding")
1078                    .header("ETag", etag)
1079                    .header("Accept-Ranges", "bytes")
1080                    .body(ResponseBody::Buffered(Full::new(Bytes::new()))),
1081            );
1082        }
1083
1084        // `Some` when the served representation differs from the file's raw bytes and had
1085        // to be built in memory; `None` means stream the open file as-is. Computed before
1086        // the HEAD check below because RFC 9110 requires a HEAD response's headers —
1087        // `Content-Length` included — to match what a GET would send, even though the body
1088        // itself is dropped.
1089        let transformed: Option<Bytes> = if html_injection {
1090            let mut html = Vec::with_capacity(metadata.len() as usize);
1091            if file.read_to_end(&mut html).await.is_err() {
1092                return internal_error_response();
1093            }
1094            if self.broadcaster.is_some() {
1095                reload::inject_reload_script(&mut html);
1096            }
1097            if self.spa_mode {
1098                spa::inject_spa_script(&mut html, self.spa_root.as_deref(), &self.spa_transition);
1099            }
1100            Some(Bytes::from(html))
1101        } else {
1102            None
1103        };
1104
1105        let file_size = transformed
1106            .as_ref()
1107            .map_or(metadata.len(), |bytes| bytes.len() as u64);
1108
1109        // Handle Range requests.
1110        let range_outcome = range_header.map(|h| parse_range_header(h, file_size));
1111        let range_check = if let Some(outcome) = &range_outcome {
1112            match outcome {
1113                RangeOutcome::Satisfiable(start, end) => {
1114                    // If-Range validation: stale If-Range ignores Range, serves full 200.
1115                    if let Some(if_range) = if_range_header {
1116                        if !if_range_valid(if_range, &etag) {
1117                            RangeCheck::IgnoreRange
1118                        } else {
1119                            RangeCheck::Satisfiable(*start, *end)
1120                        }
1121                    } else {
1122                        RangeCheck::Satisfiable(*start, *end)
1123                    }
1124                }
1125                RangeOutcome::MultiRangeIgnored => RangeCheck::IgnoreRange,
1126                RangeOutcome::Unsatisfiable => RangeCheck::Unsatisfiable,
1127                RangeOutcome::NoRange => RangeCheck::IgnoreRange,
1128            }
1129        } else {
1130            RangeCheck::IgnoreRange
1131        };
1132
1133        match &range_check {
1134            RangeCheck::Unsatisfiable => {
1135                return finish(
1136                    Response::builder()
1137                        .status(StatusCode::RANGE_NOT_SATISFIABLE)
1138                        .header("Content-Range", format!("bytes */{}", file_size))
1139                        .header("Accept-Ranges", "bytes")
1140                        .body(ResponseBody::Buffered(Full::new(Bytes::new()))),
1141                );
1142            }
1143            RangeCheck::Satisfiable(start, end) => {
1144                let range_len = end - start + 1;
1145
1146                // Seek to start position; if sidecar, we already skipped it above.
1147                if transformed.is_none()
1148                    && file.seek(std::io::SeekFrom::Start(*start)).await.is_err()
1149                {
1150                    return internal_error_response();
1151                }
1152
1153                // HEAD must not return a body (RFC 9110).
1154                let body = if *method == Method::HEAD {
1155                    ResponseBody::Buffered(Full::new(Bytes::new()))
1156                } else {
1157                    match transformed {
1158                        Some(ref bytes) => ResponseBody::Buffered(Full::new(
1159                            bytes.slice(*start as usize..(*end as usize + 1)),
1160                        )),
1161                        None => ResponseBody::Streamed(FileBody::new_ranged(file, range_len)),
1162                    }
1163                };
1164
1165                let mut builder = Response::builder()
1166                    .status(StatusCode::PARTIAL_CONTENT)
1167                    .header("Content-Type", content_type)
1168                    .header("Content-Length", range_len.to_string())
1169                    .header(
1170                        "Content-Range",
1171                        format!("bytes {}-{}/{}", start, end, file_size),
1172                    )
1173                    .header("Cache-Control", cache_control)
1174                    .header("Vary", "Accept-Encoding")
1175                    .header("ETag", etag)
1176                    .header("Accept-Ranges", "bytes");
1177                if let Some(encoding) = content_encoding {
1178                    builder = builder.header("Content-Encoding", encoding);
1179                }
1180                return finish(builder.body(body));
1181            }
1182            RangeCheck::IgnoreRange => {}
1183        }
1184
1185        // HEAD must not return a body (RFC 9110).
1186        let body = if *method == Method::HEAD {
1187            ResponseBody::Buffered(Full::new(Bytes::new()))
1188        } else {
1189            match transformed {
1190                Some(bytes) => ResponseBody::Buffered(Full::new(bytes)),
1191                None => ResponseBody::Streamed(FileBody::new(file)),
1192            }
1193        };
1194
1195        let mut builder = response(StatusCode::OK)
1196            .header("Content-Type", content_type)
1197            .header("Content-Length", file_size.to_string())
1198            .header("Cache-Control", cache_control)
1199            .header("Vary", "Accept-Encoding")
1200            .header("ETag", etag)
1201            .header("Accept-Ranges", "bytes");
1202        if let Some(encoding) = content_encoding {
1203            builder = builder.header("Content-Encoding", encoding);
1204        }
1205        finish(builder.body(body))
1206    }
1207}
1208
1209/// Ceiling on how many bytes `read_header_prefix` buffers before giving up. Without this,
1210/// a client that trickles bytes forever without ever sending the terminating blank line
1211/// could grow the buffer without limit — the header-read timeout alone doesn't bound
1212/// memory, only wall-clock time, and a sufficiently patient sender could still send
1213/// unbounded data before the deadline fires.
1214const MAX_HEADER_BYTES: usize = 64 * 1024;
1215
1216/// Why `read_header_prefix` gave up before seeing a complete header block. Every variant
1217/// is a legitimate reason to drop the connection — none is treated specially by the
1218/// caller today, but the distinction is worth preserving for anyone debugging this later.
1219#[derive(Debug)]
1220enum HeaderReadError {
1221    /// The client closed the connection (or shut down its write half) before sending a
1222    /// complete header block.
1223    ConnectionClosed,
1224    /// More than `MAX_HEADER_BYTES` were buffered without seeing `\r\n\r\n`.
1225    TooLarge,
1226    /// The underlying socket read failed. Kept rather than discarded so a future `log`
1227    /// feature has the real I/O error to report instead of an opaque unit variant.
1228    #[allow(dead_code)]
1229    Io(std::io::Error),
1230}
1231
1232/// Reads from `stream` until a complete HTTP header block (`\r\n\r\n`) has been buffered,
1233/// returning every byte read so far — which may include bytes past the header block
1234/// (request body, or a second pipelined request) if the client sent them in the same
1235/// read. Callers pair this with `tokio::time::timeout` to bound how long the header phase
1236/// itself may take; this function has no timeout of its own, only the size ceiling in
1237/// `MAX_HEADER_BYTES`.
1238async fn read_header_prefix(stream: &mut TcpStream) -> Result<Vec<u8>, HeaderReadError> {
1239    let mut buf = Vec::new();
1240    let mut chunk = [0u8; 4096];
1241
1242    loop {
1243        let n = stream.read(&mut chunk).await.map_err(HeaderReadError::Io)?;
1244        if n == 0 {
1245            return Err(HeaderReadError::ConnectionClosed);
1246        }
1247        buf.extend_from_slice(&chunk[..n]);
1248
1249        if buf.len() > MAX_HEADER_BYTES {
1250            return Err(HeaderReadError::TooLarge);
1251        }
1252        // Only the tail can hold a terminator this read completed: the `n` new bytes plus
1253        // the 3 before them. Rescanning the whole buffer every time would make the header
1254        // read quadratic in the bytes received.
1255        let scan_from = buf.len().saturating_sub(n + 3);
1256        if buf[scan_from..].windows(4).any(|w| w == b"\r\n\r\n") {
1257            return Ok(buf);
1258        }
1259    }
1260}
1261
1262/// Wraps an accepted `TcpStream` whose header block has already been drained into
1263/// `prefix` (by `read_header_prefix`, ahead of the connection being handed to hyper).
1264/// Reads replay `prefix` first, then fall through to the live socket — so hyper sees
1265/// exactly the byte stream it would have seen without the pre-read, just sourced from two
1266/// buffers back-to-back instead of one continuous one. Writes pass straight through.
1267struct PrefixedIo {
1268    prefix: Bytes,
1269    prefix_pos: usize,
1270    inner: TcpStream,
1271}
1272
1273impl PrefixedIo {
1274    fn new(prefix: Vec<u8>, inner: TcpStream) -> Self {
1275        PrefixedIo {
1276            prefix: Bytes::from(prefix),
1277            prefix_pos: 0,
1278            inner,
1279        }
1280    }
1281}
1282
1283impl AsyncRead for PrefixedIo {
1284    fn poll_read(
1285        self: Pin<&mut Self>,
1286        cx: &mut Context<'_>,
1287        buf: &mut ReadBuf<'_>,
1288    ) -> Poll<std::io::Result<()>> {
1289        let this = self.get_mut();
1290        if this.prefix_pos < this.prefix.len() {
1291            let remaining = &this.prefix[this.prefix_pos..];
1292            let n = remaining.len().min(buf.remaining());
1293            buf.put_slice(&remaining[..n]);
1294            this.prefix_pos += n;
1295            return Poll::Ready(Ok(()));
1296        }
1297        Pin::new(&mut this.inner).poll_read(cx, buf)
1298    }
1299}
1300
1301impl AsyncWrite for PrefixedIo {
1302    fn poll_write(
1303        self: Pin<&mut Self>,
1304        cx: &mut Context<'_>,
1305        buf: &[u8],
1306    ) -> Poll<std::io::Result<usize>> {
1307        Pin::new(&mut self.get_mut().inner).poll_write(cx, buf)
1308    }
1309
1310    fn poll_flush(self: Pin<&mut Self>, cx: &mut Context<'_>) -> Poll<std::io::Result<()>> {
1311        Pin::new(&mut self.get_mut().inner).poll_flush(cx)
1312    }
1313
1314    fn poll_shutdown(self: Pin<&mut Self>, cx: &mut Context<'_>) -> Poll<std::io::Result<()>> {
1315        Pin::new(&mut self.get_mut().inner).poll_shutdown(cx)
1316    }
1317}
1318
1319/// Wires an accepted connection up to the hyper HTTP/1 service.
1320///
1321/// `header_timeout` bounds only the header-read phase (`read_header_prefix`, run before
1322/// hyper ever sees the connection). Once a complete header block has been read, the
1323/// connection is handed to hyper with no further time bound — deliberately, since a
1324/// response body may legitimately outlive `header_timeout` by design (the live-reload SSE
1325/// stream is the motivating case: it stays open until a watched file changes, which may
1326/// be minutes or hours after the request). Wrapping the whole connection lifetime in
1327/// `header_timeout` — the prior implementation — silently truncated exactly that stream
1328/// once `header_timeout` elapsed, aborting the response mid-write after headers had
1329/// already been sent (the client observes this as a chunked-encoding error, not a clean
1330/// close). The connection-count ceiling (`Server::with_max_connections`) is what bounds
1331/// resource use from connections held open indefinitely, not this timeout.
1332async fn serve_connection(mut stream: TcpStream, server: Server, header_timeout: Duration) {
1333    let prefix = match timeout(header_timeout, read_header_prefix(&mut stream)).await {
1334        Ok(Ok(prefix)) => prefix,
1335        Ok(Err(_)) | Err(_) => return,
1336    };
1337
1338    let io = TokioIo::new(PrefixedIo::new(prefix, stream));
1339    let svc = service_fn(move |req: Request<Incoming>| {
1340        let server = server.clone();
1341        async move {
1342            let resp = server
1343                .handle_request(req.method(), req.uri().path(), req.headers())
1344                .await;
1345            Ok::<_, Infallible>(resp)
1346        }
1347    });
1348    let _ = AutoBuilder::new(TokioExecutor::new())
1349        .serve_connection(io, svc)
1350        .await;
1351}
1352
1353/// Default header-read timeout used by [`Server::run_ephemeral`].
1354const DEFAULT_HEADER_TIMEOUT: Duration = Duration::from_secs(30);
1355
1356/// Default grace period `ServerHandle::shutdown()` waits for in-flight connections to
1357/// finish on their own before aborting whatever is left. A connection with no
1358/// self-imposed end — an open live-reload SSE stream, or any keep-alive connection whose
1359/// peer simply never closes it — would otherwise let `shutdown()` hang forever waiting
1360/// for it to finish naturally. Every wait in this crate has a stated upper bound;
1361/// shutdown is no exception.
1362const DEFAULT_SHUTDOWN_DRAIN_TIMEOUT: Duration = Duration::from_secs(5);
1363
1364/// A handle to a server started by one of the `Server::run*` methods.
1365///
1366/// Dropping this handle without calling `shutdown()` leaves the server running in the
1367/// background for the life of the process. Call `shutdown()` to stop accepting new
1368/// connections and wait for already-accepted connections to finish before returning.
1369pub struct ServerHandle {
1370    shutdown_tx: Option<tokio::sync::oneshot::Sender<()>>,
1371    accept_task: tokio::task::JoinHandle<()>,
1372}
1373
1374impl ServerHandle {
1375    /// Stop accepting new connections and wait up to `DEFAULT_SHUTDOWN_DRAIN_TIMEOUT`
1376    /// (5s) for in-flight connections to finish on their own. Equivalent to
1377    /// `shutdown_with_timeout(DEFAULT_SHUTDOWN_DRAIN_TIMEOUT)` — see that method for what
1378    /// happens to connections still open once the grace period elapses.
1379    pub async fn shutdown(self) {
1380        self.shutdown_with_timeout(DEFAULT_SHUTDOWN_DRAIN_TIMEOUT)
1381            .await;
1382    }
1383
1384    /// Stop accepting new connections and wait up to `drain_timeout` for in-flight
1385    /// connections to finish on their own.
1386    ///
1387    /// Connections still open once `drain_timeout` elapses are aborted rather than
1388    /// waited on further: dropping the accept task drops its `JoinSet`, which aborts
1389    /// every task still tracked in it (see `tokio::task::JoinSet`'s own drop behavior) —
1390    /// which in turn drops each connection's socket, closing it. This is what bounds
1391    /// shutdown when a connection has no natural end of its own (the live-reload SSE
1392    /// stream is the motivating case: it stays open until a watched file changes, which
1393    /// may never happen before the process needs to exit).
1394    pub async fn shutdown_with_timeout(mut self, drain_timeout: Duration) {
1395        if let Some(tx) = self.shutdown_tx.take() {
1396            let _ = tx.send(());
1397        }
1398        if timeout(drain_timeout, &mut self.accept_task).await.is_err() {
1399            self.accept_task.abort();
1400        }
1401    }
1402}
1403
1404/// Read a request header as a `&str`, or `None` if it's absent or not valid ASCII.
1405fn header_str<'h>(headers: &'h HeaderMap, name: &str) -> Option<&'h str> {
1406    headers.get(name).and_then(|value| value.to_str().ok())
1407}
1408
1409/// Start a response carrying the baseline security header every response in this crate
1410/// sends. The 304 path is the one exception and builds its own — a 304 repeats only the
1411/// caching validators, not the full header set.
1412fn response(status: StatusCode) -> Builder {
1413    Response::builder()
1414        .status(status)
1415        .header("X-Content-Type-Options", "nosniff")
1416}
1417
1418/// Finish `builder` with a plain-text body. `&'static str` bodies borrow rather than
1419/// allocate; `String` bodies (the 404 message) are moved in.
1420fn text(builder: Builder, body: impl Into<Bytes>) -> Response<ResponseBody> {
1421    finish(builder.body(ResponseBody::Buffered(Full::new(body.into()))))
1422}
1423
1424/// Finishes building a response, degrading to a generic 400 instead of panicking if any
1425/// header value turns out to be invalid for use as an HTTP header value.
1426///
1427/// Every header value that reaches `Response::builder()` in this module is either a
1428/// static string or formatted from internal, already-validated data (a byte count, an
1429/// mtime, a fixed method list) — none of it can actually fail today. But `.unwrap()`
1430/// on that assumption is exactly the kind of thing that turns "can't happen" into a
1431/// production panic the day someone adds a header built from new input without
1432/// re-deriving that guarantee. Routing every response through this one fallible path
1433/// means that mistake fails safe instead of panicking.
1434fn finish(built: Result<Response<ResponseBody>, hyper::http::Error>) -> Response<ResponseBody> {
1435    built.unwrap_or_else(|_| bad_request_response())
1436}
1437
1438// `internal_error_response()` and `bad_request_response()` are the fallback responses
1439// `finish()` itself degrades to — every header and body here is a fixed string with no
1440// external input, so `.body(...)` cannot fail. They can't be routed through `finish()`
1441// without it degrading to itself on failure.
1442fn internal_error_response() -> Response<ResponseBody> {
1443    response(StatusCode::INTERNAL_SERVER_ERROR)
1444        .body(ResponseBody::Buffered(Full::new(Bytes::from_static(
1445            b"internal server error\n",
1446        ))))
1447        .unwrap()
1448}
1449
1450fn bad_request_response() -> Response<ResponseBody> {
1451    response(StatusCode::BAD_REQUEST)
1452        .body(ResponseBody::Buffered(Full::new(Bytes::from_static(
1453            b"bad request\n",
1454        ))))
1455        .unwrap()
1456}
1457
1458/// `Content-Encoding` name and sidecar file extension for each supported precompressed
1459/// variant, in preference order — brotli wins when a client accepts both and both
1460/// sidecars exist.
1461const SIDECAR_ENCODINGS: [(&str, &str); 2] = [("br", ".br"), ("gzip", ".gz")];
1462
1463/// Whether `accept_encoding` allows `encoding`.
1464///
1465/// Matches by substring rather than parsing `q`-value weights or the `identity`/`*`
1466/// directives — a lighter-weight negotiation than a general HTTP client would need,
1467/// sufficient for deciding between two static sidecar files.
1468fn accepts_encoding(accept_encoding: Option<&str>, encoding: &str) -> bool {
1469    accept_encoding.is_some_and(|header| header.contains(encoding))
1470}
1471
1472/// Look for a precompressed sidecar (`<path>.br` / `<path>.gz`) matching the client's
1473/// `Accept-Encoding`, and return its open file, metadata, and encoding name if found.
1474///
1475/// `path` must already be the fully resolved, canonicalized path `resolve()` produced.
1476/// The sidecar path is built by appending an extension to it — never by re-resolving a
1477/// modified request path — so this lookup can't become a second traversal surface: any
1478/// path this function reads is provably a sibling of a path `resolve()` already cleared.
1479async fn select_precompressed_sidecar(
1480    path: &Path,
1481    accept_encoding: Option<&str>,
1482) -> Option<(File, fs::Metadata, &'static str)> {
1483    for (encoding, ext) in SIDECAR_ENCODINGS {
1484        if !accepts_encoding(accept_encoding, encoding) {
1485            continue;
1486        }
1487        let mut sidecar = path.as_os_str().to_os_string();
1488        sidecar.push(ext);
1489        let sidecar_path = PathBuf::from(sidecar);
1490
1491        // Tripwire for the traversal boundary: a sidecar path built by appending a suffix
1492        // must stay in the same directory as `path` (which `resolve()` already proved is
1493        // inside root). `ext` is always one of the two static literals in
1494        // `SIDECAR_ENCODINGS`, never derived from request input, so this can only fire if
1495        // a future change starts deriving `sidecar` some other way.
1496        debug_assert_eq!(
1497            sidecar_path.parent(),
1498            path.parent(),
1499            "sidecar path must stay in the same directory as the already-resolved path"
1500        );
1501
1502        if let Ok(sidecar_file) = File::open(&sidecar_path).await {
1503            if let Ok(sidecar_metadata) = sidecar_file.metadata().await {
1504                return Some((sidecar_file, sidecar_metadata, encoding));
1505            }
1506        }
1507    }
1508    None
1509}
1510
1511/// Generate an ETag for a file based on modification time and size.
1512///
1513/// Format: `"<size>-<mtime_secs>"`
1514fn generate_etag(metadata: &fs::Metadata) -> String {
1515    let mtime = metadata
1516        .modified()
1517        .ok()
1518        .and_then(|t| t.duration_since(SystemTime::UNIX_EPOCH).ok())
1519        .map(|d| d.as_secs())
1520        .unwrap_or(0);
1521    format!("\"{}-{}\"", metadata.len(), mtime)
1522}
1523
1524/// Determine MIME type from file path extension.
1525fn mime_type_for_path(path: &Path) -> &'static str {
1526    let ext = path
1527        .extension()
1528        .and_then(|ext| ext.to_str())
1529        .unwrap_or_default()
1530        .to_lowercase();
1531
1532    match ext.as_str() {
1533        "html" | "htm" => "text/html; charset=utf-8",
1534        "css" => "text/css; charset=utf-8",
1535        "js" => "application/javascript; charset=utf-8",
1536        "json" => "application/json; charset=utf-8",
1537        "svg" => "image/svg+xml",
1538        "png" => "image/png",
1539        "jpg" | "jpeg" => "image/jpeg",
1540        "gif" => "image/gif",
1541        "webp" => "image/webp",
1542        "ico" => "image/x-icon",
1543        "woff" => "font/woff",
1544        "woff2" => "font/woff2",
1545        "ttf" => "font/ttf",
1546        "md" | "markdown" => "text/markdown; charset=utf-8",
1547        "txt" => "text/plain; charset=utf-8",
1548        "xml" => "application/xml",
1549        "pdf" => "application/pdf",
1550        "zip" => "application/zip",
1551        _ => "application/octet-stream",
1552    }
1553}
1554
1555/// Check if the If-None-Match header matches the current ETag.
1556/// Handles both exact match and wildcard (*) comparison per RFC 9110.
1557fn is_etag_match(if_none_match: &str, etag: &str) -> bool {
1558    if if_none_match == "*" {
1559        return true;
1560    }
1561    if_none_match.split(',').any(|tag| tag.trim() == etag)
1562}
1563
1564#[derive(Debug)]
1565enum RangeOutcome {
1566    NoRange,
1567    Satisfiable(u64, u64),
1568    Unsatisfiable,
1569    MultiRangeIgnored,
1570}
1571
1572enum RangeCheck {
1573    IgnoreRange,
1574    Satisfiable(u64, u64),
1575    Unsatisfiable,
1576}
1577
1578fn parse_range_header(header: &str, file_size: u64) -> RangeOutcome {
1579    let header = header.trim();
1580    if !header.starts_with("bytes=") {
1581        return RangeOutcome::NoRange;
1582    }
1583
1584    let range_spec = &header[6..];
1585
1586    if range_spec.contains(',') {
1587        return RangeOutcome::MultiRangeIgnored;
1588    }
1589
1590    if let Some(suffix_pos) = range_spec.find('-') {
1591        if suffix_pos == 0 {
1592            let suffix_len_str = &range_spec[1..];
1593            if let Ok(suffix_len) = suffix_len_str.parse::<u64>() {
1594                if suffix_len == 0 {
1595                    return RangeOutcome::Unsatisfiable;
1596                }
1597                if suffix_len >= file_size {
1598                    return RangeOutcome::Satisfiable(0, file_size - 1);
1599                }
1600                return RangeOutcome::Satisfiable(file_size - suffix_len, file_size - 1);
1601            }
1602            return RangeOutcome::Unsatisfiable;
1603        }
1604
1605        let start_str = &range_spec[..suffix_pos];
1606        let end_str = &range_spec[suffix_pos + 1..];
1607
1608        if let Ok(start) = start_str.parse::<u64>() {
1609            if start >= file_size {
1610                return RangeOutcome::Unsatisfiable;
1611            }
1612
1613            if end_str.is_empty() {
1614                return RangeOutcome::Satisfiable(start, file_size - 1);
1615            }
1616
1617            if let Ok(end) = end_str.parse::<u64>() {
1618                if end < start {
1619                    return RangeOutcome::Unsatisfiable;
1620                }
1621                let clamped_end = (end + 1).min(file_size) - 1;
1622                if start > clamped_end {
1623                    return RangeOutcome::Unsatisfiable;
1624                }
1625                return RangeOutcome::Satisfiable(start, clamped_end);
1626            }
1627        }
1628    }
1629
1630    RangeOutcome::Unsatisfiable
1631}
1632
1633fn if_range_valid(if_range_header: &str, current_etag: &str) -> bool {
1634    if_range_header.trim() == current_etag
1635}
1636
1637#[cfg(test)]
1638#[path = "../tests/unit/server/precompressed_sidecar.rs"]
1639mod precompressed_sidecar_tests;
1640
1641#[cfg(test)]
1642#[path = "../tests/unit/server/file_body.rs"]
1643mod file_body_tests;
1644
1645#[cfg(test)]
1646#[path = "../tests/unit/server/accept.rs"]
1647mod accept_tests;
1648
1649#[cfg(test)]
1650#[path = "../tests/unit/server/finish.rs"]
1651mod finish_tests;
1652
1653#[cfg(test)]
1654#[path = "../tests/unit/server/header_prefix.rs"]
1655mod header_prefix_tests;
1656
1657#[cfg(test)]
1658#[path = "../tests/unit/server/css_bundle.rs"]
1659mod css_bundle_tests;
1660
1661#[cfg(test)]
1662#[path = "../tests/unit/server/asset_folder.rs"]
1663mod asset_folder_tests;
1664
1665#[cfg(test)]
1666#[path = "../tests/unit/server/build_once.rs"]
1667mod build_once_tests;
1668
1669#[cfg(test)]
1670#[path = "../tests/unit/server/range_header.rs"]
1671mod range_header_tests;