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::{self, TrailingSlash};
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    trailing_slash: TrailingSlash,
138    immutable_predicate: Option<ImmutablePredicate>,
139    source_folders: Vec<PathBuf>,
140    asset_folders: Vec<PathBuf>,
141    output_dir: PathBuf,
142    css_tool: Option<(CssTool, CssOptions)>,
143    js_tool: Option<(JsTool, JsOptions)>,
144    prune_output: bool,
145}
146
147/// True when two canonical paths are the same path or one contains the other.
148///
149/// Used by the source/output overlap check: `Path::starts_with` compares component-wise, so
150/// `/a/b` neither equals nor contains their sibling `/a/b/c` itself. This is the single
151/// place the overlap rule is defined, so both `with_source_folder` and `with_output_dir`
152/// cannot drift apart.
153fn paths_overlap(a: &Path, b: &Path) -> bool {
154    a.starts_with(b) || b.starts_with(a)
155}
156
157impl Server {
158    /// Create a new server with the given root directory.
159    ///
160    /// Canonicalizes the root once at startup. All subsequent requests use the
161    /// canonical root without re-canonicalizing it, making this suitable for long-lived servers.
162    ///
163    /// # Errors
164    ///
165    /// Returns `Err(StaticError::Io)` if the root cannot be canonicalized (e.g., doesn't exist,
166    /// no read permissions).
167    pub fn new(root: &Path) -> Result<Self, StaticError> {
168        let root_canon = root.canonicalize().map_err(StaticError::Io)?;
169        let output_dir = root_canon.clone();
170        Ok(Server {
171            root_canon,
172            bundle_roots: Vec::new(),
173            max_connections: DEFAULT_MAX_CONNECTIONS,
174            live_reload: false,
175            broadcaster: None,
176            spa_mode: false,
177            spa_root: None,
178            spa_transition: SpaTransition::default(),
179            trailing_slash: TrailingSlash::default(),
180            immutable_predicate: None,
181            source_folders: Vec::new(),
182            asset_folders: Vec::new(),
183            output_dir,
184            css_tool: None,
185            js_tool: None,
186            prune_output: false,
187        })
188    }
189
190    /// Set the maximum number of connections served concurrently (default 1024).
191    ///
192    /// Once this many connections are in flight, `run()`'s accept loop stops accepting
193    /// new ones — without pausing the accept loop, a client that opens a connection and
194    /// sends nothing (see the header-read timeout docs on [`Server::run_on`]) could
195    /// otherwise be used, in enough parallel copies, to exhaust the process's file
196    /// descriptors or memory with no bound at all.
197    pub fn with_max_connections(mut self, max: usize) -> Self {
198        self.max_connections = max;
199        self
200    }
201
202    /// Enable live-reload for this server (disabled by default).
203    ///
204    /// Once enabled, the `run*` methods start a background watcher (mtime polling,
205    /// bounded 500ms interval — see [`crate::start_watching`]) over the server's root
206    /// the first time the server actually starts accepting connections, and:
207    ///
208    /// - serve a live-reload SSE stream at [`crate::LIVE_RELOAD_PATH`], broadcasting a
209    ///   change event (with [`crate::ChangeType`]) whenever a served file is added,
210    ///   modified, or removed;
211    /// - inject a small `<script>` into every served `text/html` response that connects
212    ///   to that stream and reloads the page (or hot-swaps stylesheet `<link>`s, for CSS
213    ///   changes) — no manual client wiring required.
214    ///
215    /// This is meant for local development, not production: leave it disabled (the
216    /// default) for any server serving real traffic. A typical call site gates it behind
217    /// `#[cfg(debug_assertions)]` so a release build never pays for the watcher or the
218    /// injected script.
219    ///
220    /// # Example
221    ///
222    /// ```no_run
223    /// # fn example() -> Result<(), Box<dyn std::error::Error>> {
224    /// use mini_static::Server;
225    /// use std::path::Path;
226    ///
227    /// let server = Server::new(Path::new("./public"))?;
228    /// #[cfg(debug_assertions)]
229    /// let server = server.with_live_reload();
230    /// # Ok(())
231    /// # }
232    /// ```
233    pub fn with_live_reload(mut self) -> Self {
234        self.live_reload = true;
235        self
236    }
237
238    /// Enable spa-mode navigation for this server, swapping `document.body` on
239    /// each navigation (disabled by default).
240    ///
241    /// Once enabled, every served `text/html` response gets a small `<script>`
242    /// injected (see [`Server::with_spa_root`] for what it does) that treats
243    /// `document.body` as the swap target. Calling this after
244    /// [`Server::with_spa_root`] does not clear a previously configured root
245    /// selector — the two methods set independent fields, so
246    /// `.with_spa_root(sel).with_spa_mode()` and
247    /// `.with_spa_mode().with_spa_root(sel)` both end up with spa-mode on and
248    /// root `sel`. Use this one alone when there's no persistent chrome to
249    /// preserve across navigations.
250    ///
251    /// # Example
252    ///
253    /// ```no_run
254    /// # fn example() -> Result<(), Box<dyn std::error::Error>> {
255    /// use mini_static::Server;
256    /// use std::path::Path;
257    ///
258    /// let server = Server::new(Path::new("./public"))?.with_spa_mode();
259    /// # Ok(())
260    /// # }
261    /// ```
262    pub fn with_spa_mode(mut self) -> Self {
263        self.spa_mode = true;
264        self
265    }
266
267    /// Enable spa-mode navigation for this server, swapping only the element
268    /// matched by the CSS `selector` on each navigation (disabled by default;
269    /// also enables spa-mode the same as [`Server::with_spa_mode`]).
270    ///
271    /// Once enabled, every served `text/html` response gets a small `<script>`
272    /// injected that intercepts left-clicks on same-origin `<a href>`
273    /// elements (skipping links with a non-`_self` `target`, a `download`
274    /// attribute, `rel="external"`, a `data-no-spa` attribute, or a same-page
275    /// hash-only href) and, instead of a normal navigation:
276    ///
277    /// - fetches the target URL;
278    /// - on a non-OK or non-`text/html` response (or a fetch error), falls
279    ///   back to a real `location.href` navigation — spa-mode never renders a
280    ///   broken page;
281    /// - otherwise replaces the matched element's `innerHTML` with the
282    ///   corresponding content from the fetched document, updates the page
283    ///   title, and pushes the new URL via `history.pushState`, animating the
284    ///   swap with `document.startViewTransition()` where supported;
285    /// - dispatches a `mini-static:navigate` `CustomEvent` on `window` after
286    ///   every client-side navigation, so page scripts can re-run any
287    ///   per-page initialization that would otherwise only execute once
288    ///   (content swapped in via `innerHTML` never executes its own
289    ///   `<script>` tags);
290    /// - handles browser back/forward by re-fetching and swapping to the new
291    ///   `location.href`.
292    ///
293    /// `selector` is matched against both the current page and the fetched
294    /// page; a link click where the selector matches neither falls back to a
295    /// real navigation, same as a fetch failure. Choose a `selector` that
296    /// wraps only the content that varies between pages, leaving persistent
297    /// chrome (nav/header/footer) outside it so it survives navigation
298    /// untouched.
299    ///
300    /// This is meant to be usable in production, not just local development
301    /// (unlike [`Server::with_live_reload`]): a click on a link mini-static
302    /// doesn't intercept, or on a browser without JS or View Transitions
303    /// support, still works as a normal navigation.
304    ///
305    /// # Example
306    ///
307    /// ```no_run
308    /// # fn example() -> Result<(), Box<dyn std::error::Error>> {
309    /// use mini_static::Server;
310    /// use std::path::Path;
311    ///
312    /// let server = Server::new(Path::new("./public"))?.with_spa_root("#app");
313    /// # Ok(())
314    /// # }
315    /// ```
316    pub fn with_spa_root(mut self, selector: &str) -> Self {
317        self.spa_mode = true;
318        self.spa_root = Some(selector.to_string());
319        self
320    }
321
322    /// Set how spa-mode animates the swap between pages (also enables
323    /// spa-mode the same as [`Server::with_spa_mode`]; default
324    /// [`SpaTransition::Fade`] when spa-mode is enabled without calling this).
325    ///
326    /// [`SpaTransition::Slide`] injects its own `<style>` tag alongside the
327    /// spa-mode `<script>` — no site CSS is required. See [`SpaTransition`]
328    /// and [`SlideOptions`] for what each variant does and how to
329    /// configure the slide's duration, direction, and easing.
330    ///
331    /// # Example
332    ///
333    /// ```no_run
334    /// # fn example() -> Result<(), Box<dyn std::error::Error>> {
335    /// use mini_static::{Server, SlideOptions, SpaTransition};
336    /// use std::path::Path;
337    ///
338    /// let server = Server::new(Path::new("./public"))?
339    ///     .with_spa_root("#app")
340    ///     .with_spa_transition(SpaTransition::Slide(
341    ///         SlideOptions::default().duration_ms(500),
342    ///     ));
343    /// # Ok(())
344    /// # }
345    /// ```
346    /// Sets how a directory request that omits its trailing slash is answered.
347    ///
348    /// Defaults to [`TrailingSlash::Redirect`] — a `301` to the slashed URL, which is
349    /// the canonical address of a directory index and the only form under which the
350    /// page's relative links resolve correctly.
351    ///
352    /// [`TrailingSlash::Serve`] answers the unslashed URL directly instead, saving the
353    /// round-trip. Read that variant's documentation before choosing it: it is only
354    /// correct for a site whose links and asset paths are all root-absolute, and
355    /// nothing here can check that for you.
356    pub fn with_trailing_slash(mut self, trailing_slash: TrailingSlash) -> Self {
357        self.trailing_slash = trailing_slash;
358        self
359    }
360
361    pub fn with_spa_transition(mut self, transition: SpaTransition) -> Self {
362        self.spa_mode = true;
363        self.spa_transition = transition;
364        self
365    }
366
367    /// Serve files matching `predicate` with a long-lived, immutable cache policy
368    /// instead of the default `Cache-Control: no-cache`.
369    ///
370    /// `predicate` is evaluated against each resolved file's path; a match sends
371    /// `Cache-Control: public, max-age=31536000, immutable` on that file's 200 and 304
372    /// responses. This is correct only for fingerprinted assets (e.g.
373    /// `main.a1b2c3.js`) where a content change always produces a new filename —
374    /// caching a mutable filename indefinitely would serve stale content to every
375    /// client that already has it cached.
376    ///
377    /// # Example
378    ///
379    /// ```no_run
380    /// # fn example() -> Result<(), Box<dyn std::error::Error>> {
381    /// use mini_static::Server;
382    /// use std::path::Path;
383    ///
384    /// let server = Server::new(Path::new("./public"))?
385    ///     .with_immutable_assets(|path| {
386    ///         path.file_name()
387    ///             .and_then(|name| name.to_str())
388    ///             .is_some_and(|name| name.contains(".fingerprint."))
389    ///     });
390    /// # Ok(())
391    /// # }
392    /// ```
393    pub fn with_immutable_assets<F>(mut self, predicate: F) -> Self
394    where
395        F: Fn(&Path) -> bool + Send + Sync + 'static,
396    {
397        self.immutable_predicate = Some(Arc::new(predicate));
398        self
399    }
400
401    /// The `Cache-Control` header value for a resolved file path: the immutable policy
402    /// if `with_immutable_assets`'s predicate matches, `no-cache` otherwise.
403    fn cache_control_for(&self, path: &Path) -> &'static str {
404        match &self.immutable_predicate {
405            Some(predicate) if predicate(path) => "public, max-age=31536000, immutable",
406            _ => "no-cache",
407        }
408    }
409
410    /// Register `path` as an additional directory whose changes should trigger a CSS
411    /// bundle rebuild, alongside the registered source folders.
412    ///
413    /// Useful for build pipelines where CSS partials referenced via `@import` live in a
414    /// separate directory tree from the source folders proper: without registering that
415    /// tree here, editing a partial wouldn't be noticed by the watcher and the bundle
416    /// would go stale until something else touched it.
417    ///
418    /// Files under `path` are never directly HTTP-servable: `Server::resolve` and the
419    /// request-handling path never consult bundle roots. This is purely a watch target,
420    /// not a second served root, and — since `@import` resolution is delegated entirely
421    /// to the configured [`CssTool`] (see [`Server::with_css_tool`]) — not an `@import`
422    /// traversal boundary either; the external tool resolves its own imports with no
423    /// root mini-static can enforce.
424    ///
425    /// This method is fallible and canonicalizes the path once at call time, matching
426    /// `Server::new`'s canonicalize-once policy. Call it multiple times to register
427    /// more than one external source tree.
428    ///
429    /// # Errors
430    ///
431    /// Returns `Err(StaticError::Io)` if the path cannot be canonicalized.
432    pub fn with_bundle_root(mut self, path: &Path) -> Result<Self, StaticError> {
433        let canon = path.canonicalize().map_err(StaticError::Io)?;
434        self.bundle_roots.push(canon);
435        Ok(self)
436    }
437
438    /// Designate `dir` as a source folder whose changes drive the build pipelines.
439    ///
440    /// Watched when `with_live_reload()` is enabled; `.css` files under it feed the single
441    /// CSS bundle, `.js`/`.mjs` files are minified per-file into the output dir.
442    ///
443    /// Rejected if `dir` overlaps the output dir or an already-registered source folder: a
444    /// source folder that is also the output would feed every pipeline its own output — the
445    /// feedback loop this layering exists to prevent.
446    ///
447    /// # Errors
448    ///
449    /// Returns `Err(StaticError::Io)` if `dir` cannot be canonicalized, or
450    /// `Err(StaticError::Traversal)` if it overlaps the output dir or another source folder.
451    pub fn with_source_folder(mut self, dir: &Path) -> Result<Self, StaticError> {
452        let canon = dir.canonicalize().map_err(StaticError::Io)?;
453
454        if paths_overlap(&canon, &self.output_dir) {
455            return Err(StaticError::Traversal(format!(
456                "source folder {} overlaps the output dir {}",
457                canon.display(),
458                self.output_dir.display()
459            )));
460        }
461        if self
462            .source_folders
463            .iter()
464            .chain(self.asset_folders.iter())
465            .any(|existing| paths_overlap(&canon, existing))
466        {
467            return Err(StaticError::Traversal(format!(
468                "source folder {} overlaps an already-registered source/asset folder",
469                canon.display()
470            )));
471        }
472
473        self.source_folders.push(canon);
474        Ok(self)
475    }
476
477    /// Designate `dir` as an asset source folder: every file under it (any extension)
478    /// is mirrored byte-identical into the output dir at server startup and on every
479    /// live-reload change — no CSS/JS transformation, just a flat copy preserving each
480    /// file's path relative to `dir`. Use this for hand-authored static files
481    /// (`index.html`, images) that should live outside the served/output dir as
482    /// source, the same source/output separation `with_source_folder`'s CSS/JS
483    /// pipelines already have.
484    ///
485    /// Rejected if `dir` overlaps the output dir or an already-registered
486    /// source/asset folder, for the same reason `with_source_folder` rejects it: a
487    /// folder that is also the output would feed the pipeline its own output.
488    ///
489    /// # Errors
490    ///
491    /// Returns `Err(StaticError::Io)` if `dir` cannot be canonicalized, or
492    /// `Err(StaticError::Traversal)` if it overlaps the output dir or another
493    /// registered source/asset folder.
494    pub fn with_asset_folder(mut self, dir: &Path) -> Result<Self, StaticError> {
495        let canon = dir.canonicalize().map_err(StaticError::Io)?;
496
497        if paths_overlap(&canon, &self.output_dir) {
498            return Err(StaticError::Traversal(format!(
499                "asset folder {} overlaps the output dir {}",
500                canon.display(),
501                self.output_dir.display()
502            )));
503        }
504        if self
505            .source_folders
506            .iter()
507            .chain(self.asset_folders.iter())
508            .any(|existing| paths_overlap(&canon, existing))
509        {
510            return Err(StaticError::Traversal(format!(
511                "asset folder {} overlaps an already-registered source/asset folder",
512                canon.display()
513            )));
514        }
515
516        self.asset_folders.push(canon);
517        Ok(self)
518    }
519
520    /// Designate `dir` as the output directory processed outputs are written to.
521    ///
522    /// Defaults to the served root. The output dir is never a watcher trigger: pipelines
523    /// react to source folders only, so a pipeline's own output can never re-trigger it.
524    /// Call this before `with_css_tool` so a bundle output path reflects the override.
525    ///
526    /// # Errors
527    ///
528    /// Returns `Err(StaticError::Io)` if `dir` cannot be canonicalized, or
529    /// `Err(StaticError::Traversal)` if it overlaps a registered source or asset folder.
530    pub fn with_output_dir(mut self, dir: &Path) -> Result<Self, StaticError> {
531        let canon = dir.canonicalize().map_err(StaticError::Io)?;
532
533        if self
534            .source_folders
535            .iter()
536            .chain(self.asset_folders.iter())
537            .any(|existing| paths_overlap(&canon, existing))
538        {
539            return Err(StaticError::Traversal(format!(
540                "output dir {} overlaps a registered source/asset folder",
541                canon.display()
542            )));
543        }
544
545        self.output_dir = canon;
546        Ok(self)
547    }
548
549    /// Configure CSS bundling/minification via an external tool (disabled by default).
550    ///
551    /// `tool` is a preset naming the CLI mini-static invokes (see [`CssTool`]) —
552    /// mini-static does not install or manage the binary, only looks it up on `PATH`;
553    /// [`Server::run_on`] fails fast at startup if it's missing. `options` selects
554    /// `bundle`/`minify` independently (see [`CssOptions`]):
555    ///
556    /// - Neither: every `.css` under the source folders is copied through unchanged,
557    ///   mirrored into the output dir.
558    /// - `minify` only: each file is minified independently and mirrored (no `@import`
559    ///   following).
560    /// - `bundle` only: every `.css` under the source folders is discovered,
561    ///   `@import`-resolved, and concatenated into one output file, unminified.
562    /// - Both: the bundle above, minified.
563    ///
564    /// # Example
565    ///
566    /// ```no_run
567    /// # fn example() -> Result<(), Box<dyn std::error::Error>> {
568    /// use mini_static::{CssOptions, CssTool, Server};
569    /// use std::path::Path;
570    ///
571    /// let server = Server::new(Path::new("./public"))?
572    ///     .with_css_tool(CssTool::LightningCss, CssOptions::new().bundle(true).minify(true));
573    /// # Ok(())
574    /// # }
575    /// ```
576    pub fn with_css_tool(mut self, tool: CssTool, options: CssOptions) -> Self {
577        self.css_tool = Some((tool, options));
578        self
579    }
580
581    /// Configure JS bundling/minification via an external tool (disabled by default).
582    ///
583    /// `tool` is a preset naming the CLI mini-static invokes (see [`JsTool`]) —
584    /// mini-static does not install or manage the binary, only looks it up on `PATH`;
585    /// [`Server::run_on`] fails fast at startup if it's missing. Unlike CSS, JS bundling
586    /// requires an explicit entry point ([`JsOptions::bundle_entry`]) since a JS module
587    /// graph has no well-defined "concatenate everything" meaning; without it, `options`
588    /// runs in per-file mode (every `.js`/`.mjs` under the source folders processed and
589    /// mirrored independently).
590    ///
591    /// # Errors
592    ///
593    /// Returns `Err(StaticError::Io)` if `options` specifies a bundle entry that cannot
594    /// be canonicalized, or `Err(StaticError::Traversal)` if it doesn't lie under a
595    /// registered source folder — checked eagerly here so a bad entry path fails at
596    /// configuration time, not on the first rebuild.
597    ///
598    /// # Example
599    ///
600    /// ```no_run
601    /// # fn example() -> Result<(), Box<dyn std::error::Error>> {
602    /// use mini_static::{JsOptions, JsTool, Server};
603    /// use std::path::Path;
604    ///
605    /// let server = Server::new(Path::new("./public"))?
606    ///     .with_source_folder(Path::new("./js-src"))?
607    ///     .with_js_tool(
608    ///         JsTool::Esbuild,
609    ///         JsOptions::new()
610    ///             .bundle_entry(Path::new("./js-src/main.js"), "bundle.js")
611    ///             .minify(true),
612    ///     )?;
613    /// # Ok(())
614    /// # }
615    /// ```
616    pub fn with_js_tool(mut self, tool: JsTool, options: JsOptions) -> Result<Self, StaticError> {
617        if let Some(entry) = options.entry() {
618            let entry_canon = entry.canonicalize().map_err(StaticError::Io)?;
619            let under_source_folder = self
620                .source_folders
621                .iter()
622                .any(|folder| entry_canon.starts_with(folder));
623            if !under_source_folder {
624                return Err(StaticError::Traversal(format!(
625                    "js bundle entry {} is not under any registered source folder",
626                    entry_canon.display()
627                )));
628            }
629        }
630
631        self.js_tool = Some((tool, options));
632        Ok(self)
633    }
634
635    /// Remove stale CSS bundle output at build time — specifically, delete the bundle file
636    /// when no CSS sources remain, rather than serving an orphan. Applies only to the
637    /// one-shot startup build, never during live-reload.
638    pub fn with_prune_output(mut self) -> Self {
639        self.prune_output = true;
640        self
641    }
642
643    /// True when any build pipeline is configured (a CSS/JS tool and/or a source
644    /// folder), i.e. the server should run a startup build.
645    fn has_pipeline(&self) -> bool {
646        self.css_tool.is_some()
647            || self.js_tool.is_some()
648            || !self.source_folders.is_empty()
649            || !self.asset_folders.is_empty()
650    }
651
652    /// Every external tool binary this configuration actually needs at some point
653    /// (bundle and/or minify enabled — a pure passthrough config never spawns its
654    /// configured tool, so it has nothing to fail-fast on), paired with its
655    /// human-readable install hint for a fail-fast startup error.
656    fn required_tool_binaries(&self) -> Vec<(&'static str, &'static str)> {
657        let mut required = Vec::new();
658        if let Some((css_tool, options)) = &self.css_tool {
659            if options.is_bundle() || options.is_minify() {
660                required.push((css_tool.binary_name(), css_tool.install_hint()));
661            }
662        }
663        if let Some((js_tool, options)) = &self.js_tool {
664            if options.is_bundle() || options.is_minify() {
665                required.push((js_tool.binary_name(), js_tool.install_hint()));
666            }
667        }
668        required
669    }
670
671    /// Every directory to watch for source changes: the source folders, the CSS
672    /// `@import` roots, and the asset folders, deduplicated so a directory registered
673    /// under more than one role is watched once.
674    fn watch_targets(&self) -> Vec<PathBuf> {
675        let mut targets = Vec::new();
676        for dir in self
677            .source_folders
678            .iter()
679            .chain(self.bundle_roots.iter())
680            .chain(self.asset_folders.iter())
681        {
682            if !targets.contains(dir) {
683                targets.push(dir.clone());
684            }
685        }
686        targets
687    }
688
689    /// Run every configured build pipeline (CSS/JS tools, asset folders) once and
690    /// return, without starting the HTTP server. A one-shot equivalent of the
691    /// startup build `run*` does automatically — for deploy tooling that wants to
692    /// populate the output dir ahead of time (e.g. a `cargo run --bin build_static`
693    /// step before baking a Docker image), mirroring a one-shot content
694    /// builder's `build()` (e.g. `mini_docs::Builder::build()`).
695    ///
696    /// # Errors
697    ///
698    /// - `Err(StaticError::PipelineSetup)` if a configured tool's binary that's
699    ///   actually needed (bundle or minify enabled) is missing from `PATH` — checked
700    ///   before anything runs, same as [`Server::run_on`].
701    /// - `Err(StaticError::Build)` if a configured pipeline step fails (a tool
702    ///   invocation error, a filesystem error writing output, etc.).
703    ///
704    /// # Example
705    ///
706    /// ```no_run
707    /// # async fn example() -> Result<(), Box<dyn std::error::Error>> {
708    /// use mini_static::Server;
709    /// use std::path::Path;
710    ///
711    /// let server = Server::new(Path::new("./public"))?;
712    /// server.build().await?;
713    /// # Ok(())
714    /// # }
715    /// ```
716    pub async fn build(&self) -> Result<(), StaticError> {
717        for (binary, install_hint) in self.required_tool_binaries() {
718            if !tool::locate_on_path(binary) {
719                return Err(StaticError::PipelineSetup(format!(
720                    "{binary} not found on PATH ({install_hint})"
721                )));
722            }
723        }
724
725        let pipeline = SourcePipeline::new(
726            self.source_folders.clone(),
727            self.bundle_roots.clone(),
728            self.asset_folders.clone(),
729            self.output_dir.clone(),
730            self.css_tool.clone(),
731            self.js_tool.clone(),
732            self.prune_output,
733            Broadcaster::new(),
734        );
735        pipeline
736            .full_build()
737            .await
738            .map_err(|e| StaticError::Build(e.to_string()))
739    }
740
741    /// Resolve a request path under the server's root.
742    ///
743    /// This is a lower-level API for resolving paths without generating HTTP responses.
744    /// For most use cases, prefer [`Server::handle_request`] or the `run*` methods.
745    ///
746    /// # Returns
747    ///
748    /// - `Ok(PathBuf)` if the path resolves to a file within root.
749    /// - `Err(StaticError)` if the path is invalid, missing, or attempts traversal.
750    pub fn resolve(&self, request_path: &str) -> Result<PathBuf, StaticError> {
751        resolve::resolve_with_canonical_root(&self.root_canon, request_path)
752    }
753
754    /// Run the server on a specific address with a configurable header-read timeout.
755    ///
756    /// Spawns the server in a background Tokio task and returns immediately with the
757    /// assigned port number and a [`ServerHandle`]. Call `handle.shutdown().await` to
758    /// stop accepting new connections and wait for in-flight connections to finish.
759    /// Dropping the handle instead leaves the server running for the life of the process.
760    ///
761    /// # Header-Read Timeout
762    ///
763    /// Connections that don't send complete HTTP headers within `header_timeout` are closed.
764    /// This prevents slowloris attacks and resource exhaustion from incomplete requests. The
765    /// timeout applies only to the header-read phase — once a complete header block has been
766    /// read, the connection is handed off with no further time bound, so long-lived response
767    /// bodies (e.g. the live-reload SSE stream from [`Server::with_live_reload`]) are not cut
768    /// off mid-stream.
769    ///
770    /// # Precompressed Sidecars
771    ///
772    /// If a request's `Accept-Encoding` allows `br` or `gzip` (preferring `br`) and a
773    /// sibling `<path>.br`/`<path>.gz` exists next to the resolved file, its bytes are
774    /// served instead with a matching `Content-Encoding`. Every file response carries
775    /// `Vary: Accept-Encoding` so intermediate caches don't serve the wrong variant to a
776    /// differently-capable client.
777    ///
778    /// # Arguments
779    ///
780    /// * `addr` - Socket address to bind to (e.g., `127.0.0.1:0` for loopback ephemeral,
781    ///   or `0.0.0.0:8080` to bind all interfaces on a fixed port).
782    /// * `header_timeout` - Maximum time to wait for complete HTTP headers on each connection.
783    ///
784    /// # Returns
785    ///
786    /// - `Ok((u16, ServerHandle))` with the assigned port number and a handle for graceful shutdown.
787    /// - `Err(StaticError::Io)` if binding to the socket fails.
788    /// - `Err(StaticError::PipelineSetup)` if a configured [`CssTool`]/[`JsTool`]'s binary
789    ///   is not found on `PATH`. Checked before the listener binds: a deployment whose
790    ///   configured pipeline can never run should fail visibly at boot, not be discovered
791    ///   later as a missing/stale asset.
792    pub async fn run_on(
793        &self,
794        addr: SocketAddr,
795        header_timeout: Duration,
796    ) -> Result<(u16, ServerHandle), StaticError> {
797        for (binary, install_hint) in self.required_tool_binaries() {
798            if !tool::locate_on_path(binary) {
799                return Err(StaticError::PipelineSetup(format!(
800                    "{binary} not found on PATH ({install_hint})"
801                )));
802            }
803        }
804
805        let listener = TcpListener::bind(addr).await.map_err(StaticError::Io)?;
806        let port = listener.local_addr().map_err(StaticError::Io)?.port();
807
808        let mut server = self.clone();
809        if server.live_reload {
810            let broadcaster = Broadcaster::new();
811
812            // The build pipelines react to SOURCE folders only; the output dir is never
813            // watched. Watching the output would feed each pipeline its own writes back
814            // into its trigger — the feedback loop this layering exists to prevent.
815            if server.has_pipeline() {
816                let pipeline = Arc::new(SourcePipeline::new(
817                    server.source_folders.clone(),
818                    server.bundle_roots.clone(),
819                    server.asset_folders.clone(),
820                    server.output_dir.clone(),
821                    server.css_tool.clone(),
822                    server.js_tool.clone(),
823                    server.prune_output,
824                    broadcaster.clone(),
825                ));
826                let mut rx = broadcaster.subscribe();
827                tokio::spawn(async move {
828                    // One-shot startup build (and optional prune) first, so the earliest
829                    // request already sees fresh output rather than yesterday's.
830                    if let Err(e) = pipeline.full_build().await {
831                        eprintln!("source pipeline build error: {e}");
832                    }
833                    while let Some(event) = rx.recv().await {
834                        if let Err(e) = pipeline
835                            .process_change(&event.path, &event.change_type)
836                            .await
837                        {
838                            eprintln!("source pipeline error: {e}");
839                        }
840                    }
841                });
842            }
843
844            for dir in server.watch_targets() {
845                start_watching(Arc::new(dir), broadcaster.clone());
846            }
847
848            server.broadcaster = Some(broadcaster);
849        } else if server.has_pipeline() {
850            // No live-reload: still run the one-shot build so a release boot reflects the
851            // current sources. The broadcaster is a throwaway — there is no browser to
852            // notify, so broadcasting into it is a no-op.
853            let pipeline = Arc::new(SourcePipeline::new(
854                server.source_folders.clone(),
855                server.bundle_roots.clone(),
856                server.asset_folders.clone(),
857                server.output_dir.clone(),
858                server.css_tool.clone(),
859                server.js_tool.clone(),
860                server.prune_output,
861                Broadcaster::new(),
862            ));
863            tokio::spawn(async move {
864                if let Err(e) = pipeline.full_build().await {
865                    eprintln!("source pipeline build error: {e}");
866                }
867            });
868        }
869        let semaphore = Arc::new(Semaphore::new(server.max_connections));
870        let (shutdown_tx, shutdown_rx) = tokio::sync::oneshot::channel();
871
872        let accept_task = tokio::spawn(async move {
873            let mut backoff = ACCEPT_BACKOFF_INITIAL;
874            let mut join_set: tokio::task::JoinSet<()> = tokio::task::JoinSet::new();
875            let mut shutdown_pin = std::pin::pin!(shutdown_rx);
876            let mut shutting_down = false;
877
878            loop {
879                if !shutting_down {
880                    // The accept-and-permit step and the shutdown signal race in a single
881                    // `select!` so shutdown can preempt a pending accept or a permit wait
882                    // cleanly, at any point — not just between loop iterations.
883                    tokio::select! {
884                        accepted = accept_and_permit(&listener, &mut backoff, &semaphore) => {
885                            match accepted {
886                                Some((stream, permit)) => {
887                                    let server = server.clone();
888                                    join_set.spawn(async move {
889                                        let _permit = permit;
890                                        serve_connection(stream, server, header_timeout).await;
891                                    });
892                                }
893                                None => shutting_down = true,
894                            }
895                        }
896                        _ = shutdown_pin.as_mut() => {
897                            shutting_down = true;
898                        }
899                    }
900                    continue;
901                }
902
903                // Stop accepting; drain already-spawned connections before returning.
904                match join_set.join_next().await {
905                    Some(_) => continue,
906                    None => break,
907                }
908            }
909        });
910
911        Ok((
912            port,
913            ServerHandle {
914                shutdown_tx: Some(shutdown_tx),
915                accept_task,
916            },
917        ))
918    }
919
920    /// Run the server on loopback (127.0.0.1), binding an ephemeral port.
921    ///
922    /// Thin wrapper around [`Server::run_on`] — see it for the header-read timeout and
923    /// sidecar semantics, and for what the returned [`ServerHandle`] does.
924    pub async fn run(&self, header_timeout: Duration) -> Result<(u16, ServerHandle), StaticError> {
925        self.run_on(([127, 0, 0, 1], 0).into(), header_timeout)
926            .await
927    }
928
929    /// Run the server on all interfaces (0.0.0.0) at `port` (0 for an ephemeral port).
930    ///
931    /// Useful for containerized deployments and reverse-proxy setups. Thin wrapper
932    /// around [`Server::run_on`] — see it for the header-read timeout and sidecar
933    /// semantics, and for what the returned [`ServerHandle`] does.
934    pub async fn run_all(
935        &self,
936        port: u16,
937        header_timeout: Duration,
938    ) -> Result<(u16, ServerHandle), StaticError> {
939        self.run_on(([0, 0, 0, 0], port).into(), header_timeout)
940            .await
941    }
942
943    /// Run the server on loopback with the default 30-second header-read timeout.
944    ///
945    /// The recommended entry point for tests and lightweight services that don't need a
946    /// custom timeout. Thin wrapper around [`Server::run`].
947    ///
948    /// # Example
949    ///
950    /// ```no_run
951    /// # async fn example() -> Result<(), Box<dyn std::error::Error>> {
952    /// use mini_static::Server;
953    /// use std::path::Path;
954    ///
955    /// let server = Server::new(Path::new("./public"))?;
956    /// let (port, handle) = server.run_ephemeral().await?;
957    /// println!("Server ready on http://127.0.0.1:{}", port);
958    /// handle.shutdown().await;
959    /// # Ok(())
960    /// # }
961    /// ```
962    pub async fn run_ephemeral(&self) -> Result<(u16, ServerHandle), StaticError> {
963        self.run(DEFAULT_HEADER_TIMEOUT).await
964    }
965
966    /// Produce the HTTP response for a request, streaming file bodies to the client.
967    ///
968    /// This is the crate's single request-handling path: the `run*` accept loop calls it,
969    /// and so should any async server embedding `mini-static` as a fallback route (e.g.
970    /// `mini-unified`). It never blocks the calling task — path resolution runs on Tokio's
971    /// blocking-thread pool via `spawn_blocking`, and the file is read via async I/O.
972    ///
973    /// File responses are backed by `FileBody`, which hands hyper one 64 KB chunk at a
974    /// time as `poll_frame` is driven: memory use stays bounded to one chunk per in-flight
975    /// response regardless of file size.
976    ///
977    /// `headers` are the request's headers; `If-None-Match` (304 on a matching ETag) and
978    /// `Accept-Encoding` (precompressed sidecar selection, see [`Server::run_on`]) are the
979    /// ones read today. Only `GET` and `HEAD` are allowed; anything else gets a 405 with an
980    /// `Allow` header. Missing files and traversal attempts both get an identical 404, so a
981    /// response never discloses whether a path exists outside the root.
982    pub async fn handle_request(
983        &self,
984        method: &Method,
985        request_path: &str,
986        headers: &HeaderMap,
987    ) -> Response<ResponseBody> {
988        if method != Method::GET && method != Method::HEAD {
989            return text(
990                response(StatusCode::METHOD_NOT_ALLOWED).header("Allow", "GET, HEAD"),
991                "method not allowed\n",
992            );
993        }
994
995        // Live-reload SSE stream — only reachable when `with_live_reload()` was called
996        // and the server was started via a `run*` method (those are the only paths that
997        // populate `broadcaster`).
998        if *method == Method::GET && request_path == reload::LIVE_RELOAD_PATH {
999            if let Some(broadcaster) = &self.broadcaster {
1000                return finish(
1001                    response(StatusCode::OK)
1002                        .header("Content-Type", "text/event-stream")
1003                        .header("Cache-Control", "no-cache")
1004                        .header("Connection", "keep-alive")
1005                        .body(ResponseBody::Sse(SseBody::new(broadcaster.subscribe()))),
1006                );
1007            }
1008        }
1009
1010        // `resolve()` does blocking filesystem syscalls (`canonicalize()`, up to two per
1011        // request). Running those directly in this `async fn` would block whichever
1012        // Tokio worker thread happens to be driving it, stalling every other task
1013        // scheduled on that thread for the duration of the syscalls. `spawn_blocking`
1014        // moves the work onto Tokio's dedicated blocking thread pool instead.
1015        let server = self.clone();
1016        let owned_request_path = request_path.to_string();
1017        let resolved =
1018            tokio::task::spawn_blocking(move || server.resolve(&owned_request_path)).await;
1019        let path = match resolved {
1020            Err(_) => return internal_error_response(),
1021            Ok(Err(e)) => {
1022                return text(
1023                    response(StatusCode::NOT_FOUND),
1024                    format!("{}\n", e.user_message()),
1025                )
1026            }
1027            Ok(Ok(path)) => path,
1028        };
1029
1030        // A directory served via its `index.html` needs a trailing slash to establish the
1031        // correct base for the page's relative links — unless the caller has declared
1032        // (via `with_trailing_slash`) that their markup has no relative paths to break.
1033        // Compare against the *decoded* request path so a percent-encoded explicit
1034        // request for index.html (e.g. `/docs/index.htm%6c`) is recognized as such
1035        // instead of producing a redirect to a still-encoded, broken Location.
1036        let decoded_request_path = resolve::decode_request_path(request_path);
1037        if self.trailing_slash == TrailingSlash::Redirect
1038            && path.file_name().is_some_and(|name| name == "index.html")
1039            && !decoded_request_path.ends_with('/')
1040            && !decoded_request_path.ends_with("index.html")
1041        {
1042            // `location` is built from the (attacker-controlled) request path; `finish()`
1043            // degrades to 400 instead of panicking if it ever contains bytes invalid in a
1044            // header value.
1045            let location = format!("{}/", request_path.trim_end_matches('/'));
1046            return text(
1047                response(StatusCode::MOVED_PERMANENTLY).header("Location", location),
1048                "moved\n",
1049            );
1050        }
1051
1052        let Ok(file) = File::open(&path).await else {
1053            return internal_error_response();
1054        };
1055        let Ok(metadata) = file.metadata().await else {
1056            return internal_error_response();
1057        };
1058
1059        let content_type = mime_type_for_path(&path);
1060        // Live-reload and spa-mode HTML injection both need the original, uncompressed
1061        // bytes to splice their script into — never substitute a precompressed sidecar on
1062        // this path. `broadcaster` is only `Some` when live-reload is enabled (see
1063        // `Server::with_live_reload`); `spa_mode` is independent of it (see
1064        // `Server::with_spa_mode`/`with_spa_root`) — either alone is enough to trigger
1065        // injection.
1066        let html_injection =
1067            (self.broadcaster.is_some() || self.spa_mode) && content_type.starts_with("text/html");
1068
1069        let range_header = header_str(headers, "range");
1070        let if_range_header = header_str(headers, "if-range");
1071
1072        let accept_encoding = header_str(headers, "accept-encoding");
1073        // Skip precompressed sidecars when Range is requested (serve original file instead).
1074        let sidecar = if html_injection || range_header.is_some() {
1075            None
1076        } else {
1077            select_precompressed_sidecar(&path, accept_encoding).await
1078        };
1079        let (mut file, metadata, content_encoding) = match sidecar {
1080            Some((sidecar_file, sidecar_metadata, encoding)) => {
1081                (sidecar_file, sidecar_metadata, Some(encoding))
1082            }
1083            None => (file, metadata, None),
1084        };
1085
1086        // HTML injection is skipped for a served precompressed sidecar (already final
1087        // bytes from a build step) — see `html_injection`'s definition above.
1088        let etag = generate_etag(&metadata);
1089        let cache_control = self.cache_control_for(&path);
1090
1091        if header_str(headers, "if-none-match").is_some_and(|value| is_etag_match(value, &etag)) {
1092            return finish(
1093                Response::builder()
1094                    .status(StatusCode::NOT_MODIFIED)
1095                    .header("Cache-Control", cache_control)
1096                    .header("Vary", "Accept-Encoding")
1097                    .header("ETag", etag)
1098                    .header("Accept-Ranges", "bytes")
1099                    .body(ResponseBody::Buffered(Full::new(Bytes::new()))),
1100            );
1101        }
1102
1103        // `Some` when the served representation differs from the file's raw bytes and had
1104        // to be built in memory; `None` means stream the open file as-is. Computed before
1105        // the HEAD check below because RFC 9110 requires a HEAD response's headers —
1106        // `Content-Length` included — to match what a GET would send, even though the body
1107        // itself is dropped.
1108        let transformed: Option<Bytes> = if html_injection {
1109            let mut html = Vec::with_capacity(metadata.len() as usize);
1110            if file.read_to_end(&mut html).await.is_err() {
1111                return internal_error_response();
1112            }
1113            if self.broadcaster.is_some() {
1114                reload::inject_reload_script(&mut html);
1115            }
1116            if self.spa_mode {
1117                spa::inject_spa_script(&mut html, self.spa_root.as_deref(), &self.spa_transition);
1118            }
1119            Some(Bytes::from(html))
1120        } else {
1121            None
1122        };
1123
1124        let file_size = transformed
1125            .as_ref()
1126            .map_or(metadata.len(), |bytes| bytes.len() as u64);
1127
1128        // Handle Range requests.
1129        let range_outcome = range_header.map(|h| parse_range_header(h, file_size));
1130        let range_check = if let Some(outcome) = &range_outcome {
1131            match outcome {
1132                RangeOutcome::Satisfiable(start, end) => {
1133                    // If-Range validation: stale If-Range ignores Range, serves full 200.
1134                    if let Some(if_range) = if_range_header {
1135                        if !if_range_valid(if_range, &etag) {
1136                            RangeCheck::IgnoreRange
1137                        } else {
1138                            RangeCheck::Satisfiable(*start, *end)
1139                        }
1140                    } else {
1141                        RangeCheck::Satisfiable(*start, *end)
1142                    }
1143                }
1144                RangeOutcome::MultiRangeIgnored => RangeCheck::IgnoreRange,
1145                RangeOutcome::Unsatisfiable => RangeCheck::Unsatisfiable,
1146                RangeOutcome::NoRange => RangeCheck::IgnoreRange,
1147            }
1148        } else {
1149            RangeCheck::IgnoreRange
1150        };
1151
1152        match &range_check {
1153            RangeCheck::Unsatisfiable => {
1154                return finish(
1155                    Response::builder()
1156                        .status(StatusCode::RANGE_NOT_SATISFIABLE)
1157                        .header("Content-Range", format!("bytes */{}", file_size))
1158                        .header("Accept-Ranges", "bytes")
1159                        .body(ResponseBody::Buffered(Full::new(Bytes::new()))),
1160                );
1161            }
1162            RangeCheck::Satisfiable(start, end) => {
1163                let range_len = end - start + 1;
1164
1165                // Seek to start position; if sidecar, we already skipped it above.
1166                if transformed.is_none()
1167                    && file.seek(std::io::SeekFrom::Start(*start)).await.is_err()
1168                {
1169                    return internal_error_response();
1170                }
1171
1172                // HEAD must not return a body (RFC 9110).
1173                let body = if *method == Method::HEAD {
1174                    ResponseBody::Buffered(Full::new(Bytes::new()))
1175                } else {
1176                    match transformed {
1177                        Some(ref bytes) => ResponseBody::Buffered(Full::new(
1178                            bytes.slice(*start as usize..(*end as usize + 1)),
1179                        )),
1180                        None => ResponseBody::Streamed(FileBody::new_ranged(file, range_len)),
1181                    }
1182                };
1183
1184                let mut builder = Response::builder()
1185                    .status(StatusCode::PARTIAL_CONTENT)
1186                    .header("Content-Type", content_type)
1187                    .header("Content-Length", range_len.to_string())
1188                    .header(
1189                        "Content-Range",
1190                        format!("bytes {}-{}/{}", start, end, file_size),
1191                    )
1192                    .header("Cache-Control", cache_control)
1193                    .header("Vary", "Accept-Encoding")
1194                    .header("ETag", etag)
1195                    .header("Accept-Ranges", "bytes");
1196                if let Some(encoding) = content_encoding {
1197                    builder = builder.header("Content-Encoding", encoding);
1198                }
1199                return finish(builder.body(body));
1200            }
1201            RangeCheck::IgnoreRange => {}
1202        }
1203
1204        // HEAD must not return a body (RFC 9110).
1205        let body = if *method == Method::HEAD {
1206            ResponseBody::Buffered(Full::new(Bytes::new()))
1207        } else {
1208            match transformed {
1209                Some(bytes) => ResponseBody::Buffered(Full::new(bytes)),
1210                None => ResponseBody::Streamed(FileBody::new(file)),
1211            }
1212        };
1213
1214        let mut builder = response(StatusCode::OK)
1215            .header("Content-Type", content_type)
1216            .header("Content-Length", file_size.to_string())
1217            .header("Cache-Control", cache_control)
1218            .header("Vary", "Accept-Encoding")
1219            .header("ETag", etag)
1220            .header("Accept-Ranges", "bytes");
1221        if let Some(encoding) = content_encoding {
1222            builder = builder.header("Content-Encoding", encoding);
1223        }
1224        finish(builder.body(body))
1225    }
1226}
1227
1228/// Ceiling on how many bytes `read_header_prefix` buffers before giving up. Without this,
1229/// a client that trickles bytes forever without ever sending the terminating blank line
1230/// could grow the buffer without limit — the header-read timeout alone doesn't bound
1231/// memory, only wall-clock time, and a sufficiently patient sender could still send
1232/// unbounded data before the deadline fires.
1233const MAX_HEADER_BYTES: usize = 64 * 1024;
1234
1235/// Why `read_header_prefix` gave up before seeing a complete header block. Every variant
1236/// is a legitimate reason to drop the connection — none is treated specially by the
1237/// caller today, but the distinction is worth preserving for anyone debugging this later.
1238#[derive(Debug)]
1239enum HeaderReadError {
1240    /// The client closed the connection (or shut down its write half) before sending a
1241    /// complete header block.
1242    ConnectionClosed,
1243    /// More than `MAX_HEADER_BYTES` were buffered without seeing `\r\n\r\n`.
1244    TooLarge,
1245    /// The underlying socket read failed. Kept rather than discarded so a future `log`
1246    /// feature has the real I/O error to report instead of an opaque unit variant.
1247    #[allow(dead_code)]
1248    Io(std::io::Error),
1249}
1250
1251/// Reads from `stream` until a complete HTTP header block (`\r\n\r\n`) has been buffered,
1252/// returning every byte read so far — which may include bytes past the header block
1253/// (request body, or a second pipelined request) if the client sent them in the same
1254/// read. Callers pair this with `tokio::time::timeout` to bound how long the header phase
1255/// itself may take; this function has no timeout of its own, only the size ceiling in
1256/// `MAX_HEADER_BYTES`.
1257async fn read_header_prefix(stream: &mut TcpStream) -> Result<Vec<u8>, HeaderReadError> {
1258    let mut buf = Vec::new();
1259    let mut chunk = [0u8; 4096];
1260
1261    loop {
1262        let n = stream.read(&mut chunk).await.map_err(HeaderReadError::Io)?;
1263        if n == 0 {
1264            return Err(HeaderReadError::ConnectionClosed);
1265        }
1266        buf.extend_from_slice(&chunk[..n]);
1267
1268        if buf.len() > MAX_HEADER_BYTES {
1269            return Err(HeaderReadError::TooLarge);
1270        }
1271        // Only the tail can hold a terminator this read completed: the `n` new bytes plus
1272        // the 3 before them. Rescanning the whole buffer every time would make the header
1273        // read quadratic in the bytes received.
1274        let scan_from = buf.len().saturating_sub(n + 3);
1275        if buf[scan_from..].windows(4).any(|w| w == b"\r\n\r\n") {
1276            return Ok(buf);
1277        }
1278    }
1279}
1280
1281/// Wraps an accepted `TcpStream` whose header block has already been drained into
1282/// `prefix` (by `read_header_prefix`, ahead of the connection being handed to hyper).
1283/// Reads replay `prefix` first, then fall through to the live socket — so hyper sees
1284/// exactly the byte stream it would have seen without the pre-read, just sourced from two
1285/// buffers back-to-back instead of one continuous one. Writes pass straight through.
1286struct PrefixedIo {
1287    prefix: Bytes,
1288    prefix_pos: usize,
1289    inner: TcpStream,
1290}
1291
1292impl PrefixedIo {
1293    fn new(prefix: Vec<u8>, inner: TcpStream) -> Self {
1294        PrefixedIo {
1295            prefix: Bytes::from(prefix),
1296            prefix_pos: 0,
1297            inner,
1298        }
1299    }
1300}
1301
1302impl AsyncRead for PrefixedIo {
1303    fn poll_read(
1304        self: Pin<&mut Self>,
1305        cx: &mut Context<'_>,
1306        buf: &mut ReadBuf<'_>,
1307    ) -> Poll<std::io::Result<()>> {
1308        let this = self.get_mut();
1309        if this.prefix_pos < this.prefix.len() {
1310            let remaining = &this.prefix[this.prefix_pos..];
1311            let n = remaining.len().min(buf.remaining());
1312            buf.put_slice(&remaining[..n]);
1313            this.prefix_pos += n;
1314            return Poll::Ready(Ok(()));
1315        }
1316        Pin::new(&mut this.inner).poll_read(cx, buf)
1317    }
1318}
1319
1320impl AsyncWrite for PrefixedIo {
1321    fn poll_write(
1322        self: Pin<&mut Self>,
1323        cx: &mut Context<'_>,
1324        buf: &[u8],
1325    ) -> Poll<std::io::Result<usize>> {
1326        Pin::new(&mut self.get_mut().inner).poll_write(cx, buf)
1327    }
1328
1329    fn poll_flush(self: Pin<&mut Self>, cx: &mut Context<'_>) -> Poll<std::io::Result<()>> {
1330        Pin::new(&mut self.get_mut().inner).poll_flush(cx)
1331    }
1332
1333    fn poll_shutdown(self: Pin<&mut Self>, cx: &mut Context<'_>) -> Poll<std::io::Result<()>> {
1334        Pin::new(&mut self.get_mut().inner).poll_shutdown(cx)
1335    }
1336}
1337
1338/// Wires an accepted connection up to the hyper HTTP/1 service.
1339///
1340/// `header_timeout` bounds only the header-read phase (`read_header_prefix`, run before
1341/// hyper ever sees the connection). Once a complete header block has been read, the
1342/// connection is handed to hyper with no further time bound — deliberately, since a
1343/// response body may legitimately outlive `header_timeout` by design (the live-reload SSE
1344/// stream is the motivating case: it stays open until a watched file changes, which may
1345/// be minutes or hours after the request). Wrapping the whole connection lifetime in
1346/// `header_timeout` — the prior implementation — silently truncated exactly that stream
1347/// once `header_timeout` elapsed, aborting the response mid-write after headers had
1348/// already been sent (the client observes this as a chunked-encoding error, not a clean
1349/// close). The connection-count ceiling (`Server::with_max_connections`) is what bounds
1350/// resource use from connections held open indefinitely, not this timeout.
1351async fn serve_connection(mut stream: TcpStream, server: Server, header_timeout: Duration) {
1352    let prefix = match timeout(header_timeout, read_header_prefix(&mut stream)).await {
1353        Ok(Ok(prefix)) => prefix,
1354        Ok(Err(_)) | Err(_) => return,
1355    };
1356
1357    let io = TokioIo::new(PrefixedIo::new(prefix, stream));
1358    let svc = service_fn(move |req: Request<Incoming>| {
1359        let server = server.clone();
1360        async move {
1361            let resp = server
1362                .handle_request(req.method(), req.uri().path(), req.headers())
1363                .await;
1364            Ok::<_, Infallible>(resp)
1365        }
1366    });
1367    let _ = AutoBuilder::new(TokioExecutor::new())
1368        .serve_connection(io, svc)
1369        .await;
1370}
1371
1372/// Default header-read timeout used by [`Server::run_ephemeral`].
1373const DEFAULT_HEADER_TIMEOUT: Duration = Duration::from_secs(30);
1374
1375/// Default grace period `ServerHandle::shutdown()` waits for in-flight connections to
1376/// finish on their own before aborting whatever is left. A connection with no
1377/// self-imposed end — an open live-reload SSE stream, or any keep-alive connection whose
1378/// peer simply never closes it — would otherwise let `shutdown()` hang forever waiting
1379/// for it to finish naturally. Every wait in this crate has a stated upper bound;
1380/// shutdown is no exception.
1381const DEFAULT_SHUTDOWN_DRAIN_TIMEOUT: Duration = Duration::from_secs(5);
1382
1383/// A handle to a server started by one of the `Server::run*` methods.
1384///
1385/// Dropping this handle without calling `shutdown()` leaves the server running in the
1386/// background for the life of the process. Call `shutdown()` to stop accepting new
1387/// connections and wait for already-accepted connections to finish before returning.
1388pub struct ServerHandle {
1389    shutdown_tx: Option<tokio::sync::oneshot::Sender<()>>,
1390    accept_task: tokio::task::JoinHandle<()>,
1391}
1392
1393impl ServerHandle {
1394    /// Stop accepting new connections and wait up to `DEFAULT_SHUTDOWN_DRAIN_TIMEOUT`
1395    /// (5s) for in-flight connections to finish on their own. Equivalent to
1396    /// `shutdown_with_timeout(DEFAULT_SHUTDOWN_DRAIN_TIMEOUT)` — see that method for what
1397    /// happens to connections still open once the grace period elapses.
1398    pub async fn shutdown(self) {
1399        self.shutdown_with_timeout(DEFAULT_SHUTDOWN_DRAIN_TIMEOUT)
1400            .await;
1401    }
1402
1403    /// Stop accepting new connections and wait up to `drain_timeout` for in-flight
1404    /// connections to finish on their own.
1405    ///
1406    /// Connections still open once `drain_timeout` elapses are aborted rather than
1407    /// waited on further: dropping the accept task drops its `JoinSet`, which aborts
1408    /// every task still tracked in it (see `tokio::task::JoinSet`'s own drop behavior) —
1409    /// which in turn drops each connection's socket, closing it. This is what bounds
1410    /// shutdown when a connection has no natural end of its own (the live-reload SSE
1411    /// stream is the motivating case: it stays open until a watched file changes, which
1412    /// may never happen before the process needs to exit).
1413    pub async fn shutdown_with_timeout(mut self, drain_timeout: Duration) {
1414        if let Some(tx) = self.shutdown_tx.take() {
1415            let _ = tx.send(());
1416        }
1417        if timeout(drain_timeout, &mut self.accept_task).await.is_err() {
1418            self.accept_task.abort();
1419        }
1420    }
1421}
1422
1423/// Read a request header as a `&str`, or `None` if it's absent or not valid ASCII.
1424fn header_str<'h>(headers: &'h HeaderMap, name: &str) -> Option<&'h str> {
1425    headers.get(name).and_then(|value| value.to_str().ok())
1426}
1427
1428/// Start a response carrying the baseline security header every response in this crate
1429/// sends. The 304 path is the one exception and builds its own — a 304 repeats only the
1430/// caching validators, not the full header set.
1431fn response(status: StatusCode) -> Builder {
1432    Response::builder()
1433        .status(status)
1434        .header("X-Content-Type-Options", "nosniff")
1435}
1436
1437/// Finish `builder` with a plain-text body. `&'static str` bodies borrow rather than
1438/// allocate; `String` bodies (the 404 message) are moved in.
1439fn text(builder: Builder, body: impl Into<Bytes>) -> Response<ResponseBody> {
1440    finish(builder.body(ResponseBody::Buffered(Full::new(body.into()))))
1441}
1442
1443/// Finishes building a response, degrading to a generic 400 instead of panicking if any
1444/// header value turns out to be invalid for use as an HTTP header value.
1445///
1446/// Every header value that reaches `Response::builder()` in this module is either a
1447/// static string or formatted from internal, already-validated data (a byte count, an
1448/// mtime, a fixed method list) — none of it can actually fail today. But `.unwrap()`
1449/// on that assumption is exactly the kind of thing that turns "can't happen" into a
1450/// production panic the day someone adds a header built from new input without
1451/// re-deriving that guarantee. Routing every response through this one fallible path
1452/// means that mistake fails safe instead of panicking.
1453fn finish(built: Result<Response<ResponseBody>, hyper::http::Error>) -> Response<ResponseBody> {
1454    built.unwrap_or_else(|_| bad_request_response())
1455}
1456
1457// `internal_error_response()` and `bad_request_response()` are the fallback responses
1458// `finish()` itself degrades to — every header and body here is a fixed string with no
1459// external input, so `.body(...)` cannot fail. They can't be routed through `finish()`
1460// without it degrading to itself on failure.
1461fn internal_error_response() -> Response<ResponseBody> {
1462    response(StatusCode::INTERNAL_SERVER_ERROR)
1463        .body(ResponseBody::Buffered(Full::new(Bytes::from_static(
1464            b"internal server error\n",
1465        ))))
1466        .unwrap()
1467}
1468
1469fn bad_request_response() -> Response<ResponseBody> {
1470    response(StatusCode::BAD_REQUEST)
1471        .body(ResponseBody::Buffered(Full::new(Bytes::from_static(
1472            b"bad request\n",
1473        ))))
1474        .unwrap()
1475}
1476
1477/// `Content-Encoding` name and sidecar file extension for each supported precompressed
1478/// variant, in preference order — brotli wins when a client accepts both and both
1479/// sidecars exist.
1480const SIDECAR_ENCODINGS: [(&str, &str); 2] = [("br", ".br"), ("gzip", ".gz")];
1481
1482/// Whether `accept_encoding` allows `encoding`.
1483///
1484/// Matches by substring rather than parsing `q`-value weights or the `identity`/`*`
1485/// directives — a lighter-weight negotiation than a general HTTP client would need,
1486/// sufficient for deciding between two static sidecar files.
1487fn accepts_encoding(accept_encoding: Option<&str>, encoding: &str) -> bool {
1488    accept_encoding.is_some_and(|header| header.contains(encoding))
1489}
1490
1491/// Look for a precompressed sidecar (`<path>.br` / `<path>.gz`) matching the client's
1492/// `Accept-Encoding`, and return its open file, metadata, and encoding name if found.
1493///
1494/// `path` must already be the fully resolved, canonicalized path `resolve()` produced.
1495/// The sidecar path is built by appending an extension to it — never by re-resolving a
1496/// modified request path — so this lookup can't become a second traversal surface: any
1497/// path this function reads is provably a sibling of a path `resolve()` already cleared.
1498async fn select_precompressed_sidecar(
1499    path: &Path,
1500    accept_encoding: Option<&str>,
1501) -> Option<(File, fs::Metadata, &'static str)> {
1502    for (encoding, ext) in SIDECAR_ENCODINGS {
1503        if !accepts_encoding(accept_encoding, encoding) {
1504            continue;
1505        }
1506        let mut sidecar = path.as_os_str().to_os_string();
1507        sidecar.push(ext);
1508        let sidecar_path = PathBuf::from(sidecar);
1509
1510        // Tripwire for the traversal boundary: a sidecar path built by appending a suffix
1511        // must stay in the same directory as `path` (which `resolve()` already proved is
1512        // inside root). `ext` is always one of the two static literals in
1513        // `SIDECAR_ENCODINGS`, never derived from request input, so this can only fire if
1514        // a future change starts deriving `sidecar` some other way.
1515        debug_assert_eq!(
1516            sidecar_path.parent(),
1517            path.parent(),
1518            "sidecar path must stay in the same directory as the already-resolved path"
1519        );
1520
1521        if let Ok(sidecar_file) = File::open(&sidecar_path).await {
1522            if let Ok(sidecar_metadata) = sidecar_file.metadata().await {
1523                return Some((sidecar_file, sidecar_metadata, encoding));
1524            }
1525        }
1526    }
1527    None
1528}
1529
1530/// Generate an ETag for a file based on modification time and size.
1531///
1532/// Format: `"<size>-<mtime_secs>"`
1533fn generate_etag(metadata: &fs::Metadata) -> String {
1534    let mtime = metadata
1535        .modified()
1536        .ok()
1537        .and_then(|t| t.duration_since(SystemTime::UNIX_EPOCH).ok())
1538        .map(|d| d.as_secs())
1539        .unwrap_or(0);
1540    format!("\"{}-{}\"", metadata.len(), mtime)
1541}
1542
1543/// Determine MIME type from file path extension.
1544fn mime_type_for_path(path: &Path) -> &'static str {
1545    let ext = path
1546        .extension()
1547        .and_then(|ext| ext.to_str())
1548        .unwrap_or_default()
1549        .to_lowercase();
1550
1551    match ext.as_str() {
1552        "html" | "htm" => "text/html; charset=utf-8",
1553        "css" => "text/css; charset=utf-8",
1554        "js" => "application/javascript; charset=utf-8",
1555        "json" => "application/json; charset=utf-8",
1556        "svg" => "image/svg+xml",
1557        "png" => "image/png",
1558        "jpg" | "jpeg" => "image/jpeg",
1559        "gif" => "image/gif",
1560        "webp" => "image/webp",
1561        "ico" => "image/x-icon",
1562        "woff" => "font/woff",
1563        "woff2" => "font/woff2",
1564        "ttf" => "font/ttf",
1565        "md" | "markdown" => "text/markdown; charset=utf-8",
1566        "txt" => "text/plain; charset=utf-8",
1567        "xml" => "application/xml",
1568        "pdf" => "application/pdf",
1569        "zip" => "application/zip",
1570        _ => "application/octet-stream",
1571    }
1572}
1573
1574/// Check if the If-None-Match header matches the current ETag.
1575/// Handles both exact match and wildcard (*) comparison per RFC 9110.
1576fn is_etag_match(if_none_match: &str, etag: &str) -> bool {
1577    if if_none_match == "*" {
1578        return true;
1579    }
1580    if_none_match.split(',').any(|tag| tag.trim() == etag)
1581}
1582
1583#[derive(Debug)]
1584enum RangeOutcome {
1585    NoRange,
1586    Satisfiable(u64, u64),
1587    Unsatisfiable,
1588    MultiRangeIgnored,
1589}
1590
1591enum RangeCheck {
1592    IgnoreRange,
1593    Satisfiable(u64, u64),
1594    Unsatisfiable,
1595}
1596
1597fn parse_range_header(header: &str, file_size: u64) -> RangeOutcome {
1598    let header = header.trim();
1599    if !header.starts_with("bytes=") {
1600        return RangeOutcome::NoRange;
1601    }
1602
1603    let range_spec = &header[6..];
1604
1605    if range_spec.contains(',') {
1606        return RangeOutcome::MultiRangeIgnored;
1607    }
1608
1609    if let Some(suffix_pos) = range_spec.find('-') {
1610        if suffix_pos == 0 {
1611            let suffix_len_str = &range_spec[1..];
1612            if let Ok(suffix_len) = suffix_len_str.parse::<u64>() {
1613                if suffix_len == 0 {
1614                    return RangeOutcome::Unsatisfiable;
1615                }
1616                if suffix_len >= file_size {
1617                    return RangeOutcome::Satisfiable(0, file_size - 1);
1618                }
1619                return RangeOutcome::Satisfiable(file_size - suffix_len, file_size - 1);
1620            }
1621            return RangeOutcome::Unsatisfiable;
1622        }
1623
1624        let start_str = &range_spec[..suffix_pos];
1625        let end_str = &range_spec[suffix_pos + 1..];
1626
1627        if let Ok(start) = start_str.parse::<u64>() {
1628            if start >= file_size {
1629                return RangeOutcome::Unsatisfiable;
1630            }
1631
1632            if end_str.is_empty() {
1633                return RangeOutcome::Satisfiable(start, file_size - 1);
1634            }
1635
1636            if let Ok(end) = end_str.parse::<u64>() {
1637                if end < start {
1638                    return RangeOutcome::Unsatisfiable;
1639                }
1640                let clamped_end = (end + 1).min(file_size) - 1;
1641                if start > clamped_end {
1642                    return RangeOutcome::Unsatisfiable;
1643                }
1644                return RangeOutcome::Satisfiable(start, clamped_end);
1645            }
1646        }
1647    }
1648
1649    RangeOutcome::Unsatisfiable
1650}
1651
1652fn if_range_valid(if_range_header: &str, current_etag: &str) -> bool {
1653    if_range_header.trim() == current_etag
1654}
1655
1656#[cfg(test)]
1657#[path = "../tests/unit/server/precompressed_sidecar.rs"]
1658mod precompressed_sidecar_tests;
1659
1660#[cfg(test)]
1661#[path = "../tests/unit/server/file_body.rs"]
1662mod file_body_tests;
1663
1664#[cfg(test)]
1665#[path = "../tests/unit/server/accept.rs"]
1666mod accept_tests;
1667
1668#[cfg(test)]
1669#[path = "../tests/unit/server/finish.rs"]
1670mod finish_tests;
1671
1672#[cfg(test)]
1673#[path = "../tests/unit/server/header_prefix.rs"]
1674mod header_prefix_tests;
1675
1676#[cfg(test)]
1677#[path = "../tests/unit/server/css_bundle.rs"]
1678mod css_bundle_tests;
1679
1680#[cfg(test)]
1681#[path = "../tests/unit/server/asset_folder.rs"]
1682mod asset_folder_tests;
1683
1684#[cfg(test)]
1685#[path = "../tests/unit/server/build_once.rs"]
1686mod build_once_tests;
1687
1688#[cfg(test)]
1689#[path = "../tests/unit/server/range_header.rs"]
1690mod range_header_tests;