hls_runtime/client/engine.rs
1//! [`HlsClient`] — the sans-IO caller-driven engine.
2
3use alloc::collections::{BTreeMap, BTreeSet, VecDeque};
4use alloc::string::{String, ToString};
5use alloc::vec::Vec;
6
7use broadcast_common::Unpackage;
8use broadcast_hls::{ByteRange, MapTag, MediaPlaylist, MediaSegment, OpenSegment, PreloadHintType};
9use transmux::{Fmp4Demux, TrackSpec, TsDemux};
10
11use super::action::{Action, BlockingReload, ResourceId};
12use super::error::{Error, Result};
13use super::output::Output;
14use super::url;
15
16/// First byte of every MPEG-2 TS packet (ITU-T H.222.0 / ISO/IEC 13818-1
17/// §2.4.3.2 `sync_byte`). Classic MPEG-TS-segment HLS (HLS v3, RFC 8216 —
18/// the dominant legacy/IPTV form) has no `EXT-X-MAP`/init segment at all:
19/// each `.ts` segment is a self-contained PAT/PMT/PES stream, so this byte
20/// is the only available signal to distinguish one from an fMP4/CMAF
21/// segment (which starts with an ISOBMFF box: `ftyp`/`styp`/`moof`) once the
22/// playlist itself has never advertised a Media Initialization Section.
23const TS_SYNC_BYTE: u8 = 0x47;
24
25/// A driveable, sans-IO Low-Latency HLS (RFC 8216bis) playback client.
26///
27/// `HlsClient` never touches a socket or a clock. The caller drives it:
28///
29/// 1. [`HlsClient::new`] seeds the first [`Action::FetchPlaylist`]; drain it
30/// with [`HlsClient::poll`] and perform the GET.
31/// 2. Feed the response back with [`HlsClient::on_playlist`] (playlist) or
32/// [`HlsClient::on_resource`] (init/part/segment bytes) —
33/// [`Action::FetchResource`]'s `id` correlates the two.
34/// 3. Drain [`HlsClient::poll`] again for the next round of actions (a new
35/// reload, newly discoverable parts, a preload-hint prefetch, ...) and
36/// [`HlsClient::next_output`] for newly available [`Output`]s.
37///
38/// # Behaviour
39///
40/// - **Reload scheduling** (issue #717 slice 2): once a playlist advertises
41/// `EXT-X-SERVER-CONTROL`/`EXT-X-PART-INF` **and** the origin's
42/// `CAN-BLOCK-RELOAD` attribute is `YES`
43/// ([`broadcast_hls::LowLatencyConfig::can_block_reload`] is `true` —
44/// *not* merely [`broadcast_hls::MediaPlaylist::low_latency`] being
45/// `Some`, since an origin may carry parts/PART-INF while still
46/// advertising `CAN-BLOCK-RELOAD=NO`), every reload is a Blocking
47/// Playlist Reload (RFC 8216bis §6.2.5.2) naming the next not-yet-seen
48/// Partial Segment's `_HLS_msn`/`_HLS_part`. Otherwise reloads are plain GETs
49/// paced by an [`Action::WaitMs`] hint derived from `#EXT-X-TARGETDURATION`.
50/// `EXT-X-SKIP`/`CAN-SKIP-UNTIL` Playlist Delta Updates (RFC 8216bis §4.4.5.2)
51/// are requested once a full-playlist baseline exists, and merged back into
52/// a full view before further processing — see `merge_delta` internally.
53/// - **Fetch pipeline** (slice 3): the `EXT-X-PRELOAD-HINT`ed part is fetched
54/// ahead of its own appearance as a numbered `EXT-X-PART`; `BYTERANGE`
55/// parts are supported, including the RFC 8216bis §4.4.4.9 "omitted offset
56/// means immediately after the previous sub-range of the same resource"
57/// rule (tracked per resource URL). The Media Initialization Section
58/// (`EXT-X-MAP`) is fetched once and reused for every following resource
59/// until the map changes.
60/// - **Dedup / coalescing**: once *any* of a segment's parts have been
61/// individually fetched, that segment is never re-fetched whole — when it
62/// later closes (`#EXTINF`+URI), the client only fetches whichever of its
63/// parts (if any) are still missing, and marks the segment "delivered" once
64/// every part is accounted for (fetched, or `GAP=YES`). A playlist whose
65/// segments carry **no** parts at all (a non-LL origin) falls back to
66/// fetching the whole segment resource — the two paths never overlap for a
67/// single segment, so a part's samples are never double-counted against its
68/// parent's.
69/// - **Output adapter** (slice 4): exactly one [`Output::Init`] precedes any
70/// [`Output::Samples`]; parts/segments are demuxed via
71/// [`transmux::Fmp4Demux`] (by concatenating the cached init bytes with the
72/// fetched resource — this crate never re-implements ISOBMFF box parsing,
73/// only reuses transmux's), so `Output::Samples` carries real access units,
74/// not opaque container bytes. `#EXT-X-DISCONTINUITY` on a segment surfaces
75/// as [`Output::Discontinuity`] immediately before that segment's first
76/// samples. **Known limitation**: an in-progress ([`OpenSegment`]) segment
77/// carries no discontinuity flag of its own (only a *closed*
78/// [`MediaSegment`] does) — if every part of a segment was already
79/// delivered while it was still open, a discontinuity revealed only once it
80/// closes is signalled late (after those parts' samples, not before). This
81/// is a gap in the current wire model ([`broadcast_hls::OpenSegment`]), not
82/// something this crate can fix locally.
83/// - **Classic MPEG-TS-segment HLS** (issue #760): a playlist that never
84/// advertises an `EXT-X-MAP` (HLS v3, the dominant legacy/IPTV form —
85/// self-contained `.ts` segments carrying their own PAT/PMT/PES, no
86/// separate init resource) routes each fetched Part/Segment through
87/// [`transmux::TsDemux`] instead, content-sniffed by the MPEG-TS sync byte
88/// rather than blocked on an init fetch that will never come. The first
89/// successfully demuxed segment's recovered
90/// [`TrackSpec`]s synthesize the one [`Output::Init`] this crate's contract
91/// requires (via [`transmux::build_init_segment`]) so downstream callers
92/// (e.g. `multimux`'s `HlsPull`, which recovers track specs from
93/// `Output::Init`) need no TS-specific handling of their own. The
94/// fMP4/CMAF plus LL (parts/preload-hint) path above is entirely
95/// unchanged; the two never overlap for a single playlist.
96#[derive(Debug)]
97pub struct HlsClient {
98 playlist_url: String,
99
100 pending_actions: VecDeque<Action>,
101 pending_outputs: VecDeque<Output>,
102
103 init_uri: Option<String>,
104 init_bytes: Option<Vec<u8>>,
105 init_emitted: bool,
106 /// Part/Segment resources delivered before the init segment arrived —
107 /// buffered (in arrival order) and replayed once [`Self::init_bytes`] is
108 /// set, so the caller's fetch/response IO can complete in any order
109 /// (a real HTTP client has no reason to serialize on init-first).
110 pending_demux: VecDeque<(ResourceId, Vec<u8>)>,
111
112 requested: BTreeSet<ResourceId>,
113 delivered_parts: BTreeSet<(u64, u64)>,
114 delivered_segments: BTreeSet<u64>,
115 discontinuous_msns: BTreeSet<u64>,
116 discontinuity_emitted: BTreeSet<u64>,
117 byte_range_cursor: BTreeMap<String, u64>,
118
119 outstanding_fetches: u64,
120 saw_endlist: bool,
121 end_emitted: bool,
122 last_full_playlist: Option<MediaPlaylist>,
123}
124
125impl HlsClient {
126 /// Create a new client for the Media Playlist at `playlist_url`, seeding
127 /// the first [`Action::FetchPlaylist`] (a plain, non-blocking GET — the
128 /// client does not yet know whether the origin supports blocking reload).
129 pub fn new(playlist_url: impl Into<String>) -> Self {
130 let playlist_url = playlist_url.into();
131 let mut pending_actions = VecDeque::new();
132 pending_actions.push_back(Action::FetchPlaylist {
133 url: playlist_url.clone(),
134 blocking: None,
135 skip: false,
136 });
137 Self {
138 playlist_url,
139 pending_actions,
140 pending_outputs: VecDeque::new(),
141 init_uri: None,
142 init_bytes: None,
143 init_emitted: false,
144 pending_demux: VecDeque::new(),
145 requested: BTreeSet::new(),
146 delivered_parts: BTreeSet::new(),
147 delivered_segments: BTreeSet::new(),
148 discontinuous_msns: BTreeSet::new(),
149 discontinuity_emitted: BTreeSet::new(),
150 byte_range_cursor: BTreeMap::new(),
151 outstanding_fetches: 0,
152 saw_endlist: false,
153 end_emitted: false,
154 last_full_playlist: None,
155 }
156 }
157
158 /// The Media Playlist URL this client is following.
159 pub fn playlist_url(&self) -> &str {
160 &self.playlist_url
161 }
162
163 /// Drain the next IO [`Action`] the caller must perform, if any.
164 pub fn poll(&mut self) -> Option<Action> {
165 self.pending_actions.pop_front()
166 }
167
168 /// Drain the next [`Output`] event, if any.
169 pub fn next_output(&mut self) -> Option<Output> {
170 self.pending_outputs.pop_front()
171 }
172
173 /// Feed a freshly fetched Media Playlist response.
174 ///
175 /// # Errors
176 /// [`Error::PlaylistNotUtf8`] / [`Error::PlaylistParse`] on malformed
177 /// input.
178 pub fn on_playlist(&mut self, bytes: &[u8]) -> Result<()> {
179 let text = core::str::from_utf8(bytes)?;
180 let playlist = MediaPlaylist::parse(text)?;
181 let playlist = self.merge_delta(playlist);
182
183 for (i, seg) in playlist.segments.iter().enumerate() {
184 let msn = playlist.media_sequence + i as u64;
185 self.process_closed_segment(msn, seg)?;
186 }
187
188 let next_msn = playlist.media_sequence + playlist.segments.len() as u64;
189 if let Some(open) = &playlist.open_segment {
190 self.process_open_segment(next_msn, open)?;
191 }
192
193 // Prefer the *open* segment's map when present: it's the most
194 // recent (`#EXT-X-MAP` carries forward, so the open segment's view
195 // is never older than the last closed segment's) and, crucially, is
196 // the only way to learn the init segment's URI at all when NO
197 // segment has closed yet (issue #717 slice 5 fix — previously this
198 // only ever looked at the last *closed* segment's map, so a client
199 // tuning into a stream mid-segment couldn't fetch the init segment,
200 // and therefore couldn't demux any of that segment's parts, until
201 // it closed — needlessly inflating glass-to-glass latency by up to
202 // a full segment duration on every fresh connection).
203 let map = playlist
204 .open_segment
205 .as_ref()
206 .and_then(|o| o.map.as_ref())
207 .or_else(|| playlist.segments.last().and_then(|s| s.map.as_ref()));
208 if let Some(map) = map {
209 self.ensure_init_requested(map)?;
210 }
211
212 if let Some(ll) = &playlist.low_latency
213 && let Some(hint_uri) = &ll.preload_hint_part
214 {
215 match ll.preload_hint_type {
216 PreloadHintType::Part => {
217 let part_idx = playlist
218 .open_segment
219 .as_ref()
220 .map(|o| o.parts.len() as u64)
221 .unwrap_or(0);
222 let id = ResourceId::Part {
223 msn: next_msn,
224 part: part_idx,
225 };
226 let url = url::resolve(&self.playlist_url, hint_uri);
227 let byte_range = self.resolve_hint_byte_range(&url, ll)?;
228 self.request_resource(id, url, byte_range);
229 }
230 PreloadHintType::Map => {
231 let map = MapTag {
232 uri: hint_uri.clone(),
233 byte_range: ll.preload_hint_byte_range_length.map(|length| ByteRange {
234 length,
235 offset: ll.preload_hint_byte_range_start,
236 }),
237 extra_attrs: Vec::new(),
238 };
239 self.ensure_init_requested(&map)?;
240 }
241 _ => {
242 // RFC 8216bis §4.4.5.3 defines only PART/MAP today; a
243 // future hint type from a newer transmux is simply not
244 // prefetched rather than treated as an error
245 // (`PreloadHintType` is `#[non_exhaustive]`).
246 }
247 }
248 }
249
250 if playlist.endlist {
251 self.saw_endlist = true;
252 } else {
253 // Issue #717 slice 1 fix: block only when the origin actually
254 // advertises `CAN-BLOCK-RELOAD=YES` — `low_latency.is_some()`
255 // alone is not enough (an origin sending `CAN-BLOCK-RELOAD=NO`
256 // still carries parts/PART-INF, e.g. while ramping up support).
257 let blocking = playlist
258 .low_latency
259 .as_ref()
260 .filter(|ll| ll.can_block_reload)
261 .map(|_| {
262 let part = playlist
263 .open_segment
264 .as_ref()
265 .map(|o| o.parts.len() as u64)
266 .unwrap_or(0);
267 BlockingReload {
268 msn: next_msn,
269 part: Some(part),
270 }
271 });
272 let can_skip = playlist
273 .low_latency
274 .as_ref()
275 .and_then(|ll| ll.can_skip_until)
276 .is_some();
277 let skip = can_skip && self.last_full_playlist.is_some();
278 self.pending_actions.push_back(Action::FetchPlaylist {
279 url: self.playlist_url.clone(),
280 blocking,
281 skip,
282 });
283 if blocking.is_none() {
284 // RFC 8216 §4.3.3.1: a client SHOULD NOT reload more
285 // frequently than once per Target Duration; half that as a
286 // reasonable non-blocking poll cadence.
287 let wait_ms = (u64::from(playlist.target_duration.max(1)) * 1000) / 2;
288 self.pending_actions.push_back(Action::WaitMs(wait_ms));
289 }
290 }
291
292 if playlist.skip.is_none() {
293 self.last_full_playlist = Some(playlist);
294 }
295
296 self.maybe_emit_end_of_stream();
297 Ok(())
298 }
299
300 /// Feed the bytes fetched for a previously requested [`ResourceId`]
301 /// (`init`/part/segment). Part/Segment resources delivered before the
302 /// init segment are buffered internally and demuxed once the init
303 /// arrives — the caller's fetches may complete in any order.
304 ///
305 /// # Errors
306 /// [`Error::UnrequestedResource`] if `id` was never requested (the
307 /// `requested` bookkeeping — or, for `Init`, `init_uri` — has no record
308 /// of it): a caller/driver bug, or a stale/duplicate delivery after the
309 /// client already moved past this id.
310 /// [`Error::Demux`] if `transmux::Fmp4Demux` rejects the concatenation of
311 /// the cached init + `bytes`.
312 pub fn on_resource(&mut self, id: ResourceId, bytes: &[u8]) -> Result<()> {
313 let was_requested = match id {
314 ResourceId::Init => self.init_uri.is_some(),
315 ResourceId::Part { .. } | ResourceId::Segment { .. } => self.requested.contains(&id),
316 };
317 if !was_requested {
318 return Err(Error::UnrequestedResource { id });
319 }
320 self.outstanding_fetches = self.outstanding_fetches.saturating_sub(1);
321 match id {
322 ResourceId::Init => {
323 self.init_bytes = Some(bytes.to_vec());
324 if !self.init_emitted {
325 self.pending_outputs.push_back(Output::Init(bytes.to_vec()));
326 self.init_emitted = true;
327 }
328 let buffered: Vec<_> = self.pending_demux.drain(..).collect();
329 for (bid, bbytes) in buffered {
330 self.finish_media_resource(bid, &bbytes)?;
331 }
332 }
333 ResourceId::Part { .. } | ResourceId::Segment { .. } => {
334 if self.is_ts_segment(bytes) {
335 // Classic MPEG-TS-segment HLS (issue #760): no init
336 // resource will ever arrive for this playlist, so demux
337 // this self-contained TS segment straight away rather
338 // than buffering it forever waiting for one.
339 self.finish_ts_resource(id, bytes)?;
340 } else if self.init_bytes.is_none() {
341 self.pending_demux.push_back((id, bytes.to_vec()));
342 } else {
343 self.finish_media_resource(id, bytes)?;
344 }
345 }
346 }
347 self.maybe_emit_end_of_stream();
348 Ok(())
349 }
350
351 /// Demux + emit + mark-delivered for a Part/Segment resource, once the
352 /// init segment is known to be available.
353 fn finish_media_resource(&mut self, id: ResourceId, bytes: &[u8]) -> Result<()> {
354 match id {
355 ResourceId::Part { msn, part } => {
356 self.emit_discontinuity_if_needed(msn);
357 self.demux_and_emit(id, bytes)?;
358 self.delivered_parts.insert((msn, part));
359 }
360 ResourceId::Segment { msn } => {
361 self.emit_discontinuity_if_needed(msn);
362 self.demux_and_emit(id, bytes)?;
363 self.delivered_segments.insert(msn);
364 }
365 ResourceId::Init => {}
366 }
367 Ok(())
368 }
369
370 /// The classic-TS-HLS counterpart to [`Self::finish_media_resource`]:
371 /// demux + emit + mark-delivered for a self-contained MPEG-TS Part/
372 /// Segment resource — never buffered pending an init fetch, since
373 /// [`Self::is_ts_segment`] only routes here once this playlist is known
374 /// to advertise no `EXT-X-MAP` at all.
375 fn finish_ts_resource(&mut self, id: ResourceId, bytes: &[u8]) -> Result<()> {
376 match id {
377 ResourceId::Part { msn, part } => {
378 self.emit_discontinuity_if_needed(msn);
379 self.demux_and_emit_ts(id, bytes)?;
380 self.delivered_parts.insert((msn, part));
381 }
382 ResourceId::Segment { msn } => {
383 self.emit_discontinuity_if_needed(msn);
384 self.demux_and_emit_ts(id, bytes)?;
385 self.delivered_segments.insert(msn);
386 }
387 ResourceId::Init => {}
388 }
389 Ok(())
390 }
391
392 /// `true` when `bytes` should be routed to [`Self::finish_ts_resource`]
393 /// (classic MPEG-TS-segment HLS, issue #760) rather than the fMP4/CMAF
394 /// path: this playlist has never advertised an `EXT-X-MAP` (no init
395 /// fetch is outstanding or cached — [`Self::init_uri`] is `None`; by the
396 /// time any Part/Segment fetch response reaches [`Self::on_resource`],
397 /// [`Self::on_playlist`] has already fully processed the playlist that
398 /// requested it, including any map it carries, so this check is never
399 /// stale) **and** `bytes` starts with the MPEG-TS sync byte — an
400 /// fMP4/CMAF resource always starts with an ISOBMFF box
401 /// (`ftyp`/`styp`/`moof`), never [`TS_SYNC_BYTE`].
402 fn is_ts_segment(&self, bytes: &[u8]) -> bool {
403 self.init_uri.is_none() && bytes.first() == Some(&TS_SYNC_BYTE)
404 }
405
406 /// Report that a previously requested [`ResourceId`] (or the playlist
407 /// itself, via [`None`]) failed. Clears the id's "requested" bookkeeping
408 /// so the next [`Self::on_playlist`] call naturally re-requests it (no
409 /// automatic retry timer — the caller drives retry cadence).
410 pub fn on_error(&mut self, id: Option<ResourceId>) {
411 if let Some(id) = id {
412 self.outstanding_fetches = self.outstanding_fetches.saturating_sub(1);
413 match id {
414 ResourceId::Init => self.init_uri = None,
415 other => {
416 self.requested.remove(&other);
417 }
418 }
419 }
420 self.maybe_emit_end_of_stream();
421 }
422
423 // -- internals ------------------------------------------------------
424
425 /// Reconstruct a full playlist view from an `EXT-X-SKIP` delta update
426 /// (RFC 8216bis §4.4.5.2), by splicing the skipped prefix back in from
427 /// the last full playlist this client observed. Best-effort: if there is
428 /// no cached baseline, or it doesn't cover the skipped range, the delta
429 /// is returned as-is (never an error — "at least don't break").
430 fn merge_delta(&self, playlist: MediaPlaylist) -> MediaPlaylist {
431 let Some(skip) = &playlist.skip else {
432 return playlist;
433 };
434 if skip.skipped_segments == 0 {
435 return playlist;
436 }
437 let Some(prev) = &self.last_full_playlist else {
438 return playlist;
439 };
440 if playlist.media_sequence < prev.media_sequence {
441 return playlist;
442 }
443 let prefix_start = (playlist.media_sequence - prev.media_sequence) as usize;
444 // `skip.skipped_segments` (`EXT-X-SKIP`'s `SKIPPED-SEGMENTS`,
445 // RFC 8216bis §4.4.5.2) is untrusted `u64` straight from the remote
446 // origin's playlist text, with no upper bound enforced by
447 // `broadcast_hls::MediaPlaylist::parse`. `usize::try_from` +
448 // `checked_add` guard both the u64->usize narrowing and the
449 // addition itself, so an adversarial/corrupt value falls through to
450 // the same "can't merge, return the delta as-is" fallback as every
451 // other guard in this function rather than panicking (debug) or
452 // wrapping to a bogus, silently-wrong slice bound (release).
453 let prefix_end = usize::try_from(skip.skipped_segments)
454 .ok()
455 .and_then(|skipped| prefix_start.checked_add(skipped));
456 let Some(prefix) = prefix_end.and_then(|end| prev.segments.get(prefix_start..end)) else {
457 return playlist;
458 };
459 let mut merged = playlist;
460 let mut segments = prefix.to_vec();
461 segments.extend(merged.segments);
462 merged.segments = segments;
463 merged
464 }
465
466 fn process_closed_segment(&mut self, msn: u64, seg: &MediaSegment) -> Result<()> {
467 if seg.discontinuous {
468 self.discontinuous_msns.insert(msn);
469 }
470 if self.delivered_segments.contains(&msn) {
471 return Ok(());
472 }
473 if seg.parts.is_empty() {
474 // Either a genuinely non-LL segment (never had parts), OR an LL
475 // segment whose parts were already fetched individually while it
476 // was still open and whose *closed* rendering simply omits them
477 // — RFC 8216bis does not require a closed segment to keep
478 // listing `#EXT-X-PART` lines, and real origins commonly don't
479 // (e.g. `multimux`'s: `MediaSegment.parts` is always empty for a
480 // closed segment; only the still-open segment carries parts).
481 // Detect the latter via `delivered_parts`: if any part for this
482 // `msn` was ever delivered, every one of its non-`GAP` parts was
483 // already requested while it was open (`process_open_segment`
484 // requests every known part each time it's polled, so by the
485 // time the segment closes none can have been missed) — fetching
486 // the whole segment *as well* would demux and emit its samples a
487 // second time. Caught by `hls-runtime/tests/glass_to_glass.rs`
488 // (issue #717 slice 5): every sample was double-delivered for
489 // the first two segments of a real, live-paced run.
490 let already_have_parts = self
491 .delivered_parts
492 .range((msn, 0)..(msn + 1, 0))
493 .next()
494 .is_some();
495 if already_have_parts {
496 self.delivered_segments.insert(msn);
497 return Ok(());
498 }
499 let id = ResourceId::Segment { msn };
500 if !self.requested.contains(&id) {
501 let url = url::resolve(&self.playlist_url, &seg.uri);
502 let byte_range = self.resolve_byte_range(&url, &seg.byte_range)?;
503 self.request_resource(id, url, byte_range);
504 }
505 return Ok(());
506 }
507
508 let mut fully_accounted = true;
509 for (i, part) in seg.parts.iter().enumerate() {
510 let i = i as u64;
511 if part.gap || self.delivered_parts.contains(&(msn, i)) {
512 continue;
513 }
514 fully_accounted = false;
515 let id = ResourceId::Part { msn, part: i };
516 if !self.requested.contains(&id) {
517 let url = url::resolve(&self.playlist_url, &part.uri);
518 let byte_range = self.resolve_byte_range(&url, &part.byte_range)?;
519 self.request_resource(id, url, byte_range);
520 }
521 }
522 if fully_accounted {
523 self.delivered_segments.insert(msn);
524 }
525 Ok(())
526 }
527
528 fn process_open_segment(&mut self, msn: u64, open: &OpenSegment) -> Result<()> {
529 for (i, part) in open.parts.iter().enumerate() {
530 let i = i as u64;
531 if part.gap || self.delivered_parts.contains(&(msn, i)) {
532 continue;
533 }
534 let id = ResourceId::Part { msn, part: i };
535 if !self.requested.contains(&id) {
536 let url = url::resolve(&self.playlist_url, &part.uri);
537 let byte_range = self.resolve_byte_range(&url, &part.byte_range)?;
538 self.request_resource(id, url, byte_range);
539 }
540 }
541 Ok(())
542 }
543
544 fn ensure_init_requested(&mut self, map: &MapTag) -> Result<()> {
545 let url = url::resolve(&self.playlist_url, &map.uri);
546 if self.init_uri.as_deref() == Some(url.as_str()) {
547 return Ok(());
548 }
549 self.init_uri = Some(url.clone());
550 self.init_bytes = None;
551 self.init_emitted = false;
552 let byte_range = self.resolve_byte_range(&url, &map.byte_range)?;
553 self.pending_actions.push_back(Action::FetchResource {
554 id: ResourceId::Init,
555 url,
556 byte_range,
557 });
558 self.outstanding_fetches += 1;
559 Ok(())
560 }
561
562 fn request_resource(&mut self, id: ResourceId, url: String, byte_range: Option<(u64, u64)>) {
563 self.requested.insert(id);
564 self.outstanding_fetches += 1;
565 self.pending_actions.push_back(Action::FetchResource {
566 id,
567 url,
568 byte_range,
569 });
570 }
571
572 /// Resolve a `PartSpec`/`MediaSegment`/`MapTag` `BYTERANGE` into an
573 /// absolute `(offset, length)`, honouring the "omitted offset continues
574 /// the previous sub-range of the same resource" rule (tracked per
575 /// resolved URL).
576 ///
577 /// # Errors
578 /// [`Error::ByteRangeOverflow`] if `offset + length` (both taken
579 /// straight from the untrusted remote playlist) overflows `u64` — see
580 /// that variant's doc for why this is rejected rather than saturated.
581 fn resolve_byte_range(
582 &mut self,
583 url: &str,
584 br: &Option<ByteRange>,
585 ) -> Result<Option<(u64, u64)>> {
586 let Some(br) = br.as_ref() else {
587 return Ok(None);
588 };
589 let offset = br
590 .offset
591 .unwrap_or_else(|| *self.byte_range_cursor.get(url).unwrap_or(&0));
592 let next_cursor =
593 offset
594 .checked_add(br.length)
595 .ok_or_else(|| Error::ByteRangeOverflow {
596 url: url.to_string(),
597 offset,
598 length: br.length,
599 })?;
600 self.byte_range_cursor.insert(url.to_string(), next_cursor);
601 Ok(Some((offset, br.length)))
602 }
603
604 /// Same overflow contract as [`Self::resolve_byte_range`] — see
605 /// [`Error::ByteRangeOverflow`].
606 fn resolve_hint_byte_range(
607 &mut self,
608 url: &str,
609 ll: &broadcast_hls::LowLatencyConfig,
610 ) -> Result<Option<(u64, u64)>> {
611 let Some(length) = ll.preload_hint_byte_range_length else {
612 return Ok(None);
613 };
614 let br = ByteRange {
615 length,
616 offset: ll.preload_hint_byte_range_start,
617 };
618 self.resolve_byte_range(url, &Some(br))
619 }
620
621 fn demux_and_emit(&mut self, id: ResourceId, bytes: &[u8]) -> Result<()> {
622 let init = self
623 .init_bytes
624 .as_ref()
625 .ok_or(Error::InitNotYetAvailable { id })?;
626 let mut combined = Vec::with_capacity(init.len() + bytes.len());
627 combined.extend_from_slice(init);
628 combined.extend_from_slice(bytes);
629 let mut demux = Fmp4Demux::new();
630 let media = demux
631 .unpackage(combined.as_slice())
632 .map_err(|source| Error::Demux { id, source })?;
633 for track in media.tracks {
634 if !track.samples.is_empty() {
635 self.pending_outputs.push_back(Output::Samples {
636 track_id: track.spec.track_id,
637 samples: track.samples,
638 });
639 }
640 }
641 Ok(())
642 }
643
644 /// The classic-TS-HLS counterpart to [`Self::demux_and_emit`]: demux a
645 /// self-contained MPEG-TS Part/Segment resource via [`TsDemux`] directly
646 /// (no init bytes to concatenate — each `.ts` segment carries its own
647 /// PAT/PMT/PES). On the very first such resource this client demuxes,
648 /// also synthesizes the one [`Output::Init`] the crate's output contract
649 /// requires ("exactly one `Init` precedes any `Samples`") from the
650 /// recovered [`TrackSpec`]s via [`transmux::build_init_segment`] — a real
651 /// `ftyp`+fragmented-`moov`, byte-for-byte demuxable by
652 /// `transmux::Fmp4Demux` like any other init segment, so callers built
653 /// against the fMP4 path (e.g. `multimux`'s `HlsPull`, which recovers
654 /// track specs from `Output::Init`) need no TS-specific handling.
655 fn demux_and_emit_ts(&mut self, id: ResourceId, bytes: &[u8]) -> Result<()> {
656 let mut demux = TsDemux::new();
657 let media = demux
658 .demux(bytes)
659 .map_err(|source| Error::Demux { id, source })?;
660 if !self.init_emitted {
661 let specs: Vec<TrackSpec> = media.tracks.iter().map(|t| t.spec.clone()).collect();
662 let init_bytes = transmux::build_init_segment(&specs, media.movie_timescale)
663 .map_err(|source| Error::Demux { id, source })?;
664 self.pending_outputs.push_back(Output::Init(init_bytes));
665 self.init_emitted = true;
666 }
667 for track in media.tracks {
668 if !track.samples.is_empty() {
669 self.pending_outputs.push_back(Output::Samples {
670 track_id: track.spec.track_id,
671 samples: track.samples,
672 });
673 }
674 }
675 Ok(())
676 }
677
678 fn emit_discontinuity_if_needed(&mut self, msn: u64) {
679 if self.discontinuous_msns.contains(&msn) && !self.discontinuity_emitted.contains(&msn) {
680 self.pending_outputs.push_back(Output::Discontinuity);
681 self.discontinuity_emitted.insert(msn);
682 }
683 }
684
685 fn maybe_emit_end_of_stream(&mut self) {
686 if self.saw_endlist && !self.end_emitted && self.outstanding_fetches == 0 {
687 self.pending_outputs.push_back(Output::EndOfStream);
688 self.end_emitted = true;
689 }
690 }
691}
692
693#[cfg(test)]
694mod tests {
695 use super::*;
696
697 // Regression: `on_resource` documents (see `Error::UnrequestedResource`)
698 // that it rejects a `ResourceId` the client never requested, but
699 // previously never actually checked — any bytes for any id (a
700 // caller/driver bug, or a stale/duplicate delivery) were silently
701 // accepted. Must FAIL if that check is ever removed.
702 #[test]
703 fn on_resource_rejects_a_never_requested_id() {
704 let mut client = HlsClient::new("http://example.com/playlist.m3u8");
705 let id = ResourceId::Segment { msn: 0 };
706
707 let err = client
708 .on_resource(id, b"some bytes")
709 .expect_err("an id the client never requested must be rejected");
710 assert!(
711 matches!(err, Error::UnrequestedResource { id: got } if got == id),
712 "wrong error variant: {err:?}"
713 );
714
715 // Init is checked too (tracked via `init_uri` rather than
716 // `requested`, since it's never inserted into that set).
717 let err = client
718 .on_resource(ResourceId::Init, b"init bytes")
719 .expect_err("an unrequested Init must be rejected");
720 assert!(
721 matches!(
722 err,
723 Error::UnrequestedResource {
724 id: ResourceId::Init
725 }
726 ),
727 "wrong error variant: {err:?}"
728 );
729 }
730
731 // The flip side of the regression above: a `ResourceId` the client
732 // actually asked for (via its own internal `request_resource`
733 // bookkeeping, mirroring what a real `poll()`-driven fetch populates)
734 // must still be accepted, not spuriously rejected.
735 #[test]
736 fn on_resource_accepts_a_previously_requested_id() {
737 let mut client = HlsClient::new("http://example.com/playlist.m3u8");
738 let id = ResourceId::Segment { msn: 0 };
739 client.request_resource(id, "http://example.com/seg0.m4s".to_string(), None);
740
741 // No init segment cached yet, so this is buffered rather than
742 // demuxed — the point here is only that it is *not* rejected as
743 // unrequested.
744 let result = client.on_resource(id, b"some bytes");
745 assert!(
746 result.is_ok(),
747 "a requested id must be accepted: {result:?}"
748 );
749 assert!(
750 client.pending_demux.iter().any(|(bid, _)| *bid == id),
751 "expected the resource to be buffered pending the init segment"
752 );
753 }
754
755 // Issue #760: classic MPEG-TS-segment HLS routing. `is_ts_segment` must
756 // say yes to a genuine TS resource (sync byte, no map ever seen)...
757 #[test]
758 fn is_ts_segment_true_when_no_map_seen_and_sync_byte_present() {
759 let client = HlsClient::new("http://example.com/playlist.m3u8");
760 assert!(client.is_ts_segment(&[TS_SYNC_BYTE, 0x40, 0x11, 0x00]));
761 }
762
763 // ...but say no to an ISOBMFF (fMP4/CMAF) resource even when no map has
764 // been seen yet — the content itself is never TS, so it must fall
765 // through to the ordinary init-buffering path rather than being
766 // misrouted into `TsDemux` (which would reject it as malformed TS).
767 #[test]
768 fn is_ts_segment_false_for_an_isobmff_resource_with_no_map_seen() {
769 let client = HlsClient::new("http://example.com/playlist.m3u8");
770 let ftyp_box = b"\x00\x00\x00\x18ftypiso5\x00\x00\x02\x00iso5iso6mp41";
771 assert!(!client.is_ts_segment(ftyp_box));
772 }
773
774 // The playlist signal takes precedence over content-sniffing: once this
775 // playlist is known to advertise an `EXT-X-MAP` (an init fetch has been
776 // requested/cached), even a resource whose first byte happens to be
777 // `0x47` must NOT be misrouted through `TsDemux` -- it is that
778 // playlist's own fMP4/CMAF init + part/segment concatenation the
779 // fetched bytes belong with.
780 #[test]
781 fn is_ts_segment_false_once_a_map_has_been_requested() {
782 let mut client = HlsClient::new("http://example.com/playlist.m3u8");
783 client
784 .ensure_init_requested(&MapTag {
785 uri: "init.mp4".to_string(),
786 byte_range: None,
787 extra_attrs: Vec::new(),
788 })
789 .expect("ensure_init_requested succeeds");
790 assert!(!client.is_ts_segment(&[TS_SYNC_BYTE, 0x40, 0x11, 0x00]));
791 }
792
793 // Biting test for the u64-overflow defect: a remote origin's
794 // `#EXT-X-BYTERANGE` (or preload-hint byte range) is untrusted text —
795 // `broadcast_hls::MediaPlaylist::parse` places no upper bound on either
796 // the offset or the length (confirmed: bare `str::parse::<u64>()`). A
797 // playlist advertising `BYTERANGE:18446744073709551615@1` must be
798 // rejected with `Error::ByteRangeOverflow`, not panic (debug) or
799 // silently wrap `offset + length` into a bogus cursor (release).
800 //
801 // MUTATION VERIFIED: reverting `resolve_byte_range`'s
802 // `offset.checked_add(br.length).ok_or_else(...)` back to the original
803 // `offset + br.length` makes this test fail — the debug build panics
804 // with "attempt to add with overflow" before `expect_err` ever runs
805 // (confirmed by running it), rather than the release build's silent
806 // wraparound the issue actually reports. Recompiled and re-ran to
807 // observe that exact panic, then restored the checked_add.
808 #[test]
809 fn resolve_byte_range_rejects_an_offset_plus_length_that_overflows_u64() {
810 let mut client = HlsClient::new("http://example.com/playlist.m3u8");
811 let br = Some(ByteRange {
812 length: u64::MAX,
813 offset: Some(1),
814 });
815 let err = client
816 .resolve_byte_range("http://example.com/seg0.m4s", &br)
817 .expect_err("offset 1 + length u64::MAX must overflow and be rejected");
818 assert!(
819 matches!(
820 err,
821 Error::ByteRangeOverflow {
822 offset: 1,
823 length: u64::MAX,
824 ..
825 }
826 ),
827 "wrong error variant/fields: {err:?}"
828 );
829 }
830
831 // Same defect, the other trigger named in the issue: repeated
832 // omitted-offset ranges on the *same* resource URL accumulate via
833 // `byte_range_cursor` (RFC 8216bis §4.4.4.9's "omitted offset continues
834 // the previous sub-range" rule) — a long-lived pull can walk that
835 // cursor arbitrarily close to `u64::MAX` before a single request's
836 // `offset + length` itself overflows.
837 #[test]
838 fn resolve_byte_range_rejects_a_cursor_accumulation_that_overflows_u64() {
839 let mut client = HlsClient::new("http://example.com/playlist.m3u8");
840 let url = "http://example.com/seg0.m4s";
841 // Seed the cursor near the top of the u64 range with an explicit
842 // offset, then the next omitted-offset range pushes it over.
843 let seed = Some(ByteRange {
844 length: u64::MAX - 10,
845 offset: Some(5),
846 });
847 let (offset, length) = client
848 .resolve_byte_range(url, &seed)
849 .expect("seed range does not itself overflow")
850 .expect("Some for a Some(ByteRange)");
851 assert_eq!((offset, length), (5, u64::MAX - 10));
852
853 let next = Some(ByteRange {
854 length: 100,
855 offset: None, // continues the cursor left at 5 + (u64::MAX - 10)
856 });
857 let err = client
858 .resolve_byte_range(url, &next)
859 .expect_err("cursor + next length must overflow and be rejected");
860 assert!(
861 matches!(err, Error::ByteRangeOverflow { length: 100, .. }),
862 "wrong error variant/fields: {err:?}"
863 );
864 }
865
866 // Biting test for the `merge_delta` defect: an `#EXT-X-SKIP` delta's
867 // `SKIPPED-SEGMENTS` is untrusted `u64` text with no upper bound
868 // enforced by `broadcast_hls::MediaPlaylist::parse`. A malicious/corrupt
869 // origin claiming a `SKIPPED-SEGMENTS` value that overflows
870 // `prefix_start + skipped_segments` (here: `prefix_start == 10` from a
871 // 10-segment `media_sequence` advance, plus `skipped_segments ==
872 // u64::MAX - 5`) must fall through to the existing "can't merge, return
873 // the delta as-is" fallback, not panic. `prefix_start` is deliberately
874 // nonzero: `0 + u64::MAX` does not overflow, so a `prefix_start == 0`
875 // case would pass even with the guard removed, for the wrong reason
876 // (the subsequent `.get()` bounds check catching it) rather than the
877 // one this test targets (the addition itself).
878 //
879 // MUTATION VERIFIED: reverting the guard back to the original
880 // `prefix_start + skip.skipped_segments as usize` makes this test fail:
881 // the debug build panics with "attempt to add with overflow" inside
882 // `merge_delta` before the `let merged = ...` assertions ever run
883 // (confirmed by running it), rather than returning the delta unmerged.
884 // Recompiled and re-ran to observe that exact panic, then restored the
885 // checked/try_from guard.
886 #[test]
887 fn merge_delta_does_not_panic_on_an_overflowing_skipped_segments() {
888 let mut client = HlsClient::new("http://example.com/playlist.m3u8");
889 let prev = MediaPlaylist {
890 media_sequence: 0,
891 segments: vec![MediaSegment::default(); 2],
892 ..Default::default()
893 };
894 // `last_full_playlist` is set directly (private field, same module)
895 // rather than driven through a full `on_playlist` round-trip — the
896 // point under test is purely `merge_delta`'s own arithmetic guard.
897 client.last_full_playlist = Some(prev);
898
899 let delta = MediaPlaylist {
900 media_sequence: 10, // prefix_start == 10 - 0 == 10
901 skip: Some(broadcast_hls::SkipInfo {
902 skipped_segments: u64::MAX - 5,
903 ..Default::default()
904 }),
905 segments: Vec::new(),
906 ..Default::default()
907 };
908
909 let merged = client.merge_delta(delta.clone());
910 assert_eq!(
911 merged, delta,
912 "an unmergeable skip count must fall back to the delta as-is, not panic"
913 );
914 }
915}