moqtap_proxy/framer.rs
1//! Byte-exact object framing for MoQT data streams.
2//!
3//! [`ObjectFramer`] turns the raw bytes of one unidirectional data stream
4//! into individually addressable objects without altering them: every item
5//! it yields is a slice of the bytes that were fed in, and concatenating
6//! those items reproduces the stream exactly.
7//!
8//! Framing is opt-in because it costs latency and memory — an object is
9//! only emitted once it is buffered whole. The proxy pays that cost only
10//! when something is observing; see `session::pipe_data`.
11
12use std::collections::HashMap;
13use std::sync::{Arc, Mutex};
14
15use bytes::{Buf, Bytes, BytesMut};
16
17use moqtap_codec::dispatch::{
18 reemit_subgroup_object, AnyFetchFrame, AnyFetchGroupOrder, AnyFetchHeader, AnyFetchObjectMeta,
19 AnyFetchObjectReader, AnyFetchObjectWriter, AnySubgroupHeader, AnySubgroupObjectMeta,
20 AnySubgroupObjectReader, FetchReemit,
21};
22use moqtap_codec::version::DraftVersion;
23
24use crate::capability::fetch_group_order_is_needed;
25use crate::event::DataStreamHeaderKind;
26use crate::instrument::Recorder;
27use crate::parser::data::is_incomplete_error;
28use crate::types::DataStreamType;
29
30pub use crate::types::{BypassReason, ObjectMeta};
31
32/// Every fetch this session has learned a Group Order for, keyed by the
33/// Request ID that names it on both the control plane and the data stream.
34///
35/// # Why a fetch stream needs something from outside itself
36///
37/// Drafts 18-23 write a fetch Object's Group ID as a difference from the
38/// Object before it, and the fetch's Group Order decides whether the
39/// difference counts up or down — draft-19 Section 11.4.4.1. Nothing on the
40/// data stream states the order, so a framer handed only the stream cannot
41/// reach an absolute Group ID, and the wrong choice decodes every Object
42/// without an error under Group IDs walking the wrong way.
43///
44/// The order is on the FETCH. Draft-19 Section 10.12.3: "The publisher
45/// responding to a FETCH is responsible for delivering all available Objects
46/// in the requested range in the requested order (see Section 10.2.8)." The
47/// session's control pipes read it off each FETCH they carry and file it
48/// here; [`ObjectFramer`] takes it out again when the response stream opens.
49///
50/// # Why an entry is taken rather than read
51///
52/// One FETCH opens one response stream, so an entry has exactly one reader
53/// and is spent by it. Taking bounds the table to the fetches that have been
54/// asked for and not yet answered, on a session that may run for hours, and
55/// it needs no rule about when to forget: the reader is the rule.
56///
57/// What that leaves is a fetch the publisher answered with an error rather
58/// than a stream, whose entry no reader ever comes for. A cap is the
59/// bound on those, and it fails closed — past it nothing is recorded, so the
60/// affected streams are bypassed and say so rather than being read against
61/// somebody else's order.
62#[derive(Debug, Default)]
63pub struct FetchGroupOrders {
64 known: Mutex<HashMap<u64, AnyFetchGroupOrder>>,
65}
66
67impl FetchGroupOrders {
68 /// How many asked-for-but-unanswered fetches one session may hold.
69 ///
70 /// Reached only by a peer that sends FETCHes whose responses never open a
71 /// stream, since every answered one takes its own entry away again.
72 const CAP: usize = 1024;
73
74 /// File the order a FETCH asked for, under the Request ID it asked under.
75 ///
76 /// Call with the order the *peer will act on*, which on a session whose
77 /// control frames a hook may rewrite is the one leaving the proxy rather
78 /// than the one that arrived.
79 pub fn record(&self, request_id: u64, order: AnyFetchGroupOrder) {
80 let mut known = self.known.lock().expect("no task holds the fetch orders across a panic");
81 if known.len() >= Self::CAP && !known.contains_key(&request_id) {
82 return;
83 }
84 known.insert(request_id, order);
85 }
86
87 /// Take the order filed for `request_id`, if one was.
88 #[must_use]
89 pub fn take(&self, request_id: u64) -> Option<AnyFetchGroupOrder> {
90 self.known
91 .lock()
92 .expect("no task holds the fetch orders across a panic")
93 .remove(&request_id)
94 }
95}
96
97/// Padding width that puts every wire length a varint can express within
98/// measuring reach.
99///
100/// Half of `usize::MAX`, so adding the buffered bytes cannot overflow
101/// [`PaddedBuf::remaining`]; a varint tops out at `2^62 - 1`, well inside
102/// it on a 64-bit target.
103const UNBOUNDED_MEASURING_PAD: usize = usize::MAX / 2;
104
105/// Configuration for [`ObjectFramer`].
106///
107/// Construct with [`FramerConfig::default`] and adjust fields, or with
108/// [`FramerConfig::new`] and the builder setters. The struct is
109/// `#[non_exhaustive]` so later releases can add knobs without a break;
110/// that also means a struct literal does not compile from outside this
111/// crate.
112#[derive(Debug, Clone)]
113#[non_exhaustive]
114pub struct FramerConfig {
115 /// Largest object the framer will buffer whole, in bytes. Objects
116 /// larger than this are streamed through as
117 /// [`FramerOut::Passthrough`] and are not individually addressable.
118 /// Default 4 MiB.
119 pub max_buffered_object_bytes: usize,
120}
121
122impl Default for FramerConfig {
123 fn default() -> Self {
124 Self { max_buffered_object_bytes: 4 * 1024 * 1024 }
125 }
126}
127
128impl FramerConfig {
129 /// A config with default values.
130 #[must_use]
131 pub fn new() -> Self {
132 Self::default()
133 }
134
135 /// Set the largest object the framer will buffer whole.
136 #[must_use]
137 pub fn with_max_buffered_object_bytes(mut self, bytes: usize) -> Self {
138 self.max_buffered_object_bytes = bytes;
139 self
140 }
141}
142
143/// One item produced by [`ObjectFramer::poll`].
144#[derive(Debug)]
145#[non_exhaustive]
146pub enum FramerOut {
147 /// The stream's header, with its exact wire bytes (stream-type field
148 /// included).
149 Header {
150 /// The decoded header.
151 header: DataStreamHeaderKind,
152 /// The header's wire bytes, including the stream-type field.
153 raw: Bytes,
154 },
155 /// A complete object, with its exact wire bytes.
156 Object {
157 /// The object's framing, without its payload.
158 meta: ObjectMeta,
159 /// The object's complete wire bytes, framing and payload.
160 raw: Bytes,
161 },
162 /// Bytes the framer is forwarding without interpreting them: an object
163 /// larger than the buffer cap, a fetch stream on a draft whose fetch
164 /// objects are not addressed, or any stream the framer has stopped
165 /// parsing. These bytes are not individually addressable.
166 Passthrough(Bytes),
167 /// The framer has stopped parsing this stream.
168 ///
169 /// Carries **no bytes**, so [`ObjectFramer::poll`]'s concatenation
170 /// invariant is untouched: this item contributes nothing to the
171 /// reconstruction. Emitted exactly once per stream.
172 ///
173 /// **Ordering: immediately after the item the bypass was decided
174 /// during, not in place of it.** Every one of `latch_bypass`'s call
175 /// sites returns a *different* `FramerOut` from the same `poll()` —
176 /// the two header sites fall through to [`Self::Header`], the
177 /// measuring-reach sites return [`Self::Passthrough`], the decode and
178 /// fix-up sites return [`Self::Error`] — and `poll()` returns one
179 /// item. The variant is therefore **deferred**: `latch_bypass` stores
180 /// a `pending_bypass: Option<BypassReason>`, and `poll()` drains it at
181 /// the top of its next call, before anything else.
182 Bypassed {
183 /// Why parsing stopped.
184 reason: BypassReason,
185 /// Whether an elide fix-up was still owed when parsing stopped.
186 ///
187 /// `true` means the remainder of this stream cannot be renumbered
188 /// and the destination must be reset rather than forwarded — the
189 /// bytes still in flight decode to Object IDs one ahead of what
190 /// was actually delivered.
191 fixup_owed: bool,
192 },
193 /// Not enough buffered bytes to produce another item.
194 NeedMore,
195 /// The stream could not be parsed. The framer switches to
196 /// [`Passthrough`](Self::Passthrough) for the remainder of the stream
197 /// and never reports a second error.
198 Error(String),
199}
200
201/// The state an elide leaves behind on a subgroup stream.
202///
203/// Read it with [`ObjectFramer::elide_cursor`]. The framer owns this state
204/// rather than the caller because it forwards objects the caller never
205/// sees an [`ObjectMeta`] for — an object one byte over
206/// [`FramerConfig::max_buffered_object_bytes`] leaves as
207/// [`FramerOut::Passthrough`], and a cursor kept outside the framer would
208/// let its stale delta go out on the wire.
209#[derive(Debug, Clone, Copy, PartialEq, Eq)]
210pub struct ElideCursor {
211 /// Absolute Object ID of the last object actually forwarded, or `None`
212 /// when none has been.
213 pub last_forwarded_id: Option<u64>,
214 /// Whether the next object the framer emits will have its leading ID
215 /// field rewritten. Always `false` on drafts 07-13, where IDs are
216 /// absolute, and on fetch streams.
217 pub fixup_pending: bool,
218}
219
220/// The object the framer emitted most recently, whose forward-or-elide
221/// disposition has not been committed to the cursor yet.
222#[derive(Debug, Clone, Copy)]
223struct PendingObject {
224 object_id: u64,
225 index_in_stream: u64,
226}
227
228/// Identity fields the stream header contributes to every object on a
229/// subgroup stream.
230#[derive(Debug, Clone, Copy, Default)]
231struct SubgroupContext {
232 track_alias: u64,
233 group_id: u64,
234 subgroup_id: Option<u64>,
235 publisher_priority: Option<u8>,
236}
237
238/// What the framer is currently decoding.
239#[derive(Debug)]
240enum Stage {
241 /// Nothing decoded yet; the next bytes are the stream header.
242 AwaitingHeader,
243 /// Subgroup objects, with the per-stream delta state.
244 Subgroup(AnySubgroupObjectReader),
245 /// Fetch objects, with the state on both sides of the framer: what the
246 /// stream said, and what has actually been forwarded from it.
247 ///
248 /// The two are the same value until something is elided, and every frame
249 /// goes through the writer regardless. A frame it never saw would leave
250 /// it a frame behind, and a survivor re-encoded against a stale
251 /// predecessor is a renumbered stream rather than a decode error.
252 Fetch { reader: AnyFetchObjectReader, writer: AnyFetchObjectWriter },
253}
254
255/// Frames a unidirectional MoQT data stream into individually addressable
256/// objects while preserving the exact wire bytes.
257///
258/// Feed it the raw bytes of one stream, in order, from the stream's first
259/// byte. Drain [`poll`](Self::poll) after each [`feed`](Self::feed) until
260/// it returns [`FramerOut::NeedMore`]. The concatenation of every `raw`
261/// and `Passthrough` payload it yields equals the concatenation of every
262/// chunk fed to it, unless [`note_elided`](Self::note_elided) has been
263/// called — see [`Self::poll`].
264#[derive(Debug)]
265pub struct ObjectFramer {
266 stream_type: DataStreamType,
267 draft: DraftVersion,
268 config: FramerConfig,
269 buf: BytesMut,
270 stage: Stage,
271 context: SubgroupContext,
272 /// Bytes still owed to an oversized object that is being streamed
273 /// through uninterpreted. Denominated in **source** bytes — see
274 /// [`Self::poll_oversized`].
275 passthrough_remaining: u64,
276 /// Set once parsing has been abandoned for the rest of the stream.
277 bypassed: bool,
278 /// Why parsing was abandoned, waiting for [`Self::poll`] to drain it
279 /// as a [`FramerOut::Bypassed`].
280 ///
281 /// `fixup_owed` is read off [`Self::fixup_pending`] at drain time
282 /// rather than being captured here, and the two are the same value:
283 /// once `bypassed` is set, `poll` never re-enters `poll_object`, which
284 /// is the only thing that clears the flag, and no `Object` item can be
285 /// emitted between the latch and the drain for `note_elided` to set it.
286 pending_bypass: Option<BypassReason>,
287 index_in_stream: u64,
288 /// Absolute Object ID of the last object actually forwarded.
289 last_forwarded_id: Option<u64>,
290 /// Whether the next object emitted still owes a fix-up: on a subgroup
291 /// stream the leading Object ID varint rewritten against
292 /// [`Self::last_forwarded_id`], on a fetch stream a survivor re-encoded
293 /// against the frame that is now in front of it.
294 ///
295 /// Both are settled by emitting one object, after which the writing
296 /// cursor is level with the reading one again.
297 fixup_pending: bool,
298 /// The fetch writer as it stood before the object emitted most recently
299 /// was re-emitted through it.
300 ///
301 /// Re-emitting advances the writer, and the caller's verdict on that
302 /// object does not land until the next poll, so an elide has to put the
303 /// writer back where it was — otherwise the survivor after it is encoded
304 /// against an object nobody received.
305 fetch_rollback: Option<AnyFetchObjectWriter>,
306 /// The object emitted most recently, awaiting the caller's verdict.
307 /// Committed lazily at the top of [`Self::poll_object`].
308 pending_disposition: Option<PendingObject>,
309 /// Slow-path counters for the session this framer belongs to.
310 counters: Arc<Recorder>,
311 /// The session's answered fetches, on the drafts where a fetch stream
312 /// cannot be read without one. `None` for a framer built outside a
313 /// session, and on every draft that needs no order.
314 fetch_orders: Option<Arc<FetchGroupOrders>>,
315}
316
317impl ObjectFramer {
318 /// A framer whose slow-path counters go to `counters`.
319 ///
320 /// **The only constructor `session.rs` may use.** The counters are
321 /// session-scoped, and a framer that cannot reach its session's
322 /// [`Recorder`] would leave `framers_created`, `framer_header_polls`,
323 /// `framer_object_polls`, `objects_not_addressable` and
324 /// `object_ids_rewritten` at zero for every real session — which is a
325 /// *passing* `Interest::NONE` proof obtained by measuring nothing.
326 ///
327 /// [`Self::new`] is kept, unchanged, for tests and downstream callers
328 /// that construct a framer to parse bytes rather than to forward them;
329 /// it is equivalent to passing a fresh `Recorder` whose counts nobody
330 /// reads. That is why this is an **addition** and not a signature
331 /// change.
332 pub fn with_recorder(
333 stream_type: DataStreamType,
334 draft: DraftVersion,
335 config: FramerConfig,
336 counters: Arc<Recorder>,
337 ) -> Self {
338 counters.note_framer_created();
339 Self {
340 stream_type,
341 draft,
342 config,
343 buf: BytesMut::with_capacity(4096),
344 stage: Stage::AwaitingHeader,
345 context: SubgroupContext::default(),
346 passthrough_remaining: 0,
347 bypassed: false,
348 pending_bypass: None,
349 index_in_stream: 0,
350 last_forwarded_id: None,
351 fixup_pending: false,
352 fetch_rollback: None,
353 pending_disposition: None,
354 counters,
355 fetch_orders: None,
356 }
357 }
358
359 /// Read this stream's fetch Objects against the order its FETCH asked for.
360 ///
361 /// Only drafts 18-23 need it — see [`FetchGroupOrders`] — and only a
362 /// fetch stream consults it; a subgroup framer given one ignores it. A
363 /// framer built without it on a draft that needs one reports
364 /// [`BypassReason::FetchGroupOrderUnknown`] and forwards the stream
365 /// uninterpreted, which is what every caller outside a session gets and
366 /// what the session itself gets for a stream naming a request it never
367 /// saw asked for.
368 #[must_use]
369 pub fn with_fetch_group_orders(mut self, orders: Arc<FetchGroupOrders>) -> Self {
370 self.fetch_orders = Some(orders);
371 self
372 }
373
374 /// Create a framer for a stream of the given kind on the given draft,
375 /// discarding its counters.
376 ///
377 /// Increments go to a private [`Recorder`] nothing can read, so a
378 /// caller that wants a session's counters to move must use
379 /// [`Self::with_recorder`]. Retained for callers that parse bytes
380 /// rather than forward them, where the counts are not the point.
381 pub fn new(stream_type: DataStreamType, draft: DraftVersion, config: FramerConfig) -> Self {
382 Self::with_recorder(stream_type, draft, config, Arc::new(Recorder::new()))
383 }
384
385 /// The state an elide has left behind on this stream.
386 ///
387 /// Read-only, and read-anytime: the framer applies the fix-up itself,
388 /// so nothing outside has to act on this.
389 #[must_use]
390 pub fn elide_cursor(&self) -> ElideCursor {
391 ElideCursor { last_forwarded_id: self.last_forwarded_id, fixup_pending: self.fixup_pending }
392 }
393
394 /// Record that the object the framer emitted **most recently** was not
395 /// forwarded.
396 ///
397 /// Call exactly once, immediately after deciding to drop an object,
398 /// and before the next [`Self::poll`]. `meta` must be the meta the
399 /// framer handed out for that object; it is `debug_assert`ed against
400 /// the framer's own record, because calling this out of order is the
401 /// one way to corrupt a stream silently.
402 ///
403 /// On drafts 07-13 subgroup streams and on drafts 07-14 fetch streams
404 /// this only suppresses the cursor advance — those Object IDs are
405 /// absolute, so the bytes of every later object already say the truth.
406 /// Elsewhere it also arms a fix-up: the leading Object ID varint on a
407 /// drafts 14-23 subgroup stream, and the whole framing of the next frame
408 /// on a drafts 15-23 fetch stream, where it additionally puts the fetch
409 /// writer back to where the last forwarded frame left it.
410 pub fn note_elided(&mut self, meta: &ObjectMeta) {
411 let pending = self.pending_disposition.take();
412 debug_assert!(
413 matches!(
414 pending,
415 Some(p) if p.object_id == meta.object_id
416 && p.index_in_stream == meta.index_in_stream
417 ),
418 "note_elided must name the object the framer emitted most recently; \
419 got object {} at index {}, framer holds {:?}",
420 meta.object_id,
421 meta.index_in_stream,
422 pending.map(|p| (p.object_id, p.index_in_stream)),
423 );
424 // Taking `pending` is the elide: the cursor simply does not
425 // advance to it. Arming the fix-up on a `None` would renumber an
426 // object whose predecessor was already committed as forwarded.
427 if pending.is_none() {
428 return;
429 }
430 if self.elide_owes_a_fixup() {
431 self.fixup_pending = true;
432 }
433 // The frame was re-emitted through the writer before the caller saw
434 // it, so the writer is one frame ahead of what the destination
435 // received. Nothing else can restore it: the per-draft state is what
436 // the *forwarded* frames left behind, and this frame was not one.
437 if let Some(rollback) = self.fetch_rollback.take() {
438 if let Stage::Fetch { writer, .. } = &mut self.stage {
439 *writer = rollback;
440 }
441 }
442 }
443
444 /// Buffer a chunk of stream bytes.
445 pub fn feed(&mut self, chunk: &[u8]) {
446 self.buf.extend_from_slice(chunk);
447 }
448
449 /// Bytes buffered but not yet emitted. Non-zero only mid-object.
450 pub fn buffered(&self) -> usize {
451 self.buf.len()
452 }
453
454 /// `true` once the framer has stopped parsing this stream and is
455 /// forwarding bytes uninterpreted.
456 pub fn is_bypassed(&self) -> bool {
457 self.bypassed
458 }
459
460 /// Flush any buffered bytes at end of stream.
461 ///
462 /// Called when the source signals FIN. Returns whatever the framer
463 /// still holds — a truncated final object, or bytes buffered behind an
464 /// incomplete framing — so the caller can forward them before
465 /// finishing the destination stream. Forgetting this call turns a
466 /// clean FIN into silent truncation.
467 pub fn finish(&mut self) -> Option<Bytes> {
468 self.passthrough_remaining = 0;
469 if self.buf.is_empty() {
470 None
471 } else {
472 Some(self.buf.split().freeze())
473 }
474 }
475
476 /// Produce the next item, or [`FramerOut::NeedMore`].
477 ///
478 /// # Invariant
479 ///
480 /// Concatenating the `raw` field of every [`Header`](FramerOut::Header)
481 /// and [`Object`](FramerOut::Object) and the payload of every
482 /// [`Passthrough`](FramerOut::Passthrough), in the order produced,
483 /// reproduces the fed bytes exactly — on every draft, including
484 /// streams that fall back to bypass — **unless**
485 /// [`Self::note_elided`] has been called, which deliberately removes
486 /// an object's bytes and may rewrite one leading varint.
487 ///
488 /// That exception is the *only* one, and it is opt-in per stream: a
489 /// framer used as a pure observer never calls `note_elided`, so its
490 /// output stays byte-identical to its input. When `note_elided` has
491 /// been called on a drafts 14-23 subgroup stream, the next object the
492 /// framer emits has its leading Object ID varint re-encoded against
493 /// the last object actually forwarded — one field, in one object, and
494 /// every byte after it copied verbatim. Everything else, on every
495 /// other stream, still concatenates back to the source exactly.
496 ///
497 /// [`FramerOut::Bypassed`] and [`FramerOut::NeedMore`] carry no bytes
498 /// and so contribute nothing to the reconstruction either way.
499 pub fn poll(&mut self) -> FramerOut {
500 // Drained before anything else. `latch_bypass` cannot return this
501 // item itself — each of its call sites returns a different
502 // `FramerOut` from the same `poll()` — so the bypass is reported
503 // on the next call, immediately after the item it was decided
504 // during. Zero bytes, so the invariant above is untouched.
505 if let Some(reason) = self.pending_bypass.take() {
506 return FramerOut::Bypassed { reason, fixup_owed: self.fixup_pending };
507 }
508 if self.passthrough_remaining > 0 {
509 let owed = self.passthrough_remaining;
510 return self.emit_passthrough(owed);
511 }
512 if self.bypassed {
513 return self.emit_passthrough(u64::MAX);
514 }
515 match self.stage {
516 Stage::AwaitingHeader => self.poll_header(),
517 Stage::Subgroup(_) | Stage::Fetch { .. } => self.poll_object(),
518 }
519 }
520
521 /// Hand out up to `limit` buffered bytes without interpreting them.
522 fn emit_passthrough(&mut self, limit: u64) -> FramerOut {
523 if self.buf.is_empty() {
524 return FramerOut::NeedMore;
525 }
526 let take = usize::try_from(limit).unwrap_or(usize::MAX).min(self.buf.len());
527 self.passthrough_remaining = self.passthrough_remaining.saturating_sub(take as u64);
528 FramerOut::Passthrough(self.buf.split_to(take).freeze())
529 }
530
531 /// Stop parsing this stream for good, record why, and release
532 /// everything buffered.
533 ///
534 /// Draining here is what keeps memory bounded: without it a stream the
535 /// framer cannot parse grows the buffer for as long as the peer keeps
536 /// writing.
537 ///
538 /// `reason` is stored rather than returned, because every call site
539 /// returns a different [`FramerOut`] from the same
540 /// [`poll`](Self::poll); the next `poll` drains it. Guarded on
541 /// `bypassed` so a second latch — which today cannot happen, since
542 /// `poll` short-circuits once the flag is set — could never overwrite
543 /// the first reason or double-count
544 /// `Counters::streams_not_shapeable`, whose whole job is to be one per
545 /// stream.
546 fn latch_bypass(&mut self, reason: BypassReason) {
547 if !self.bypassed {
548 self.bypassed = true;
549 self.pending_bypass = Some(reason);
550 self.counters.note_stream_not_shapeable();
551 }
552 self.passthrough_remaining = 0;
553 }
554
555 fn poll_header(&mut self) -> FramerOut {
556 self.counters.note_framer_header_poll();
557 if self.buf.is_empty() {
558 return FramerOut::NeedMore;
559 }
560
561 let snapshot: &[u8] = &self.buf[..];
562 let mut cursor: &[u8] = snapshot;
563 let decoded = match self.stream_type {
564 DataStreamType::Subgroup => AnySubgroupHeader::decode_stream(self.draft, &mut cursor)
565 .map(DataStreamHeaderKind::Subgroup),
566 DataStreamType::Fetch => AnyFetchHeader::decode_stream(self.draft, &mut cursor)
567 .map(DataStreamHeaderKind::Fetch),
568 };
569
570 match decoded {
571 Ok(header) => {
572 let consumed = snapshot.len() - cursor.remaining();
573 match &header {
574 DataStreamHeaderKind::Subgroup(h) => match AnySubgroupObjectReader::new(h) {
575 Ok(reader) => {
576 self.context = subgroup_context(h);
577 self.stage = Stage::Subgroup(reader);
578 }
579 // A subgroup stream type the object reader
580 // rejects. The header still decoded, so report it,
581 // then forward the rest uninterpreted.
582 Err(_) => self.latch_bypass(BypassReason::UnsupportedSubgroupStreamType),
583 },
584 // Whether this stream's Objects can be addressed is
585 // asked of `fetch_group_order_is_needed` and of the
586 // session's own record of what each FETCH asked for,
587 // rather than inferred from the reader refusing. A codec
588 // able to decode a layout is not on its own enough: on
589 // drafts 18-23 it decodes under a Group Order nothing
590 // on this stream states, and the wrong one decodes as
591 // willingly as the right one.
592 DataStreamHeaderKind::Fetch(h) => match self.fetch_stage(h) {
593 Ok(stage) => self.stage = stage,
594 Err(reason) => self.latch_bypass(reason),
595 },
596 }
597 let raw = self.buf.split_to(consumed).freeze();
598 FramerOut::Header { header, raw }
599 }
600 Err(e) if is_incomplete_error(&e) => {
601 if self.buf.len() >= self.config.max_buffered_object_bytes {
602 self.latch_bypass(BypassReason::ObjectBeyondMeasuringReach);
603 self.emit_passthrough(u64::MAX)
604 } else {
605 FramerOut::NeedMore
606 }
607 }
608 Err(e) => {
609 self.latch_bypass(BypassReason::DecodeError);
610 FramerOut::Error(format!("data stream header decode: {e}"))
611 }
612 }
613 }
614
615 /// The reader and writer a fetch stream is parsed with, or why it is not.
616 ///
617 /// Three drafts need an answer from off the stream and the other eleven
618 /// do not, which is [`fetch_group_order_is_needed`]. Where one is needed
619 /// it comes from the FETCH the session carried, filed under the Request ID
620 /// this header names; a stream naming a request that was never asked for
621 /// is bypassed rather than guessed at, because the guess would decode.
622 ///
623 /// The reader and writer are built with the same order. They are the two
624 /// halves of one stream — the writer re-encodes a survivor after an elide
625 /// against the frames now in front of it — so a pair built against
626 /// different orders would renumber the stream it was meant to preserve.
627 fn fetch_stage(&mut self, h: &AnyFetchHeader) -> Result<Stage, BypassReason> {
628 let order = if fetch_group_order_is_needed(self.draft) {
629 let orders = self.fetch_orders.as_ref().ok_or(BypassReason::FetchGroupOrderUnknown)?;
630 Some(orders.take(h.request_id()).ok_or(BypassReason::FetchGroupOrderUnknown)?)
631 } else {
632 None
633 };
634 // Ascending where the draft needs no answer: the branch above supplies
635 // one for every draft that does, so the fallback is only ever reached
636 // where nothing on the stream is signed and the value cannot be wrong.
637 let order = order.unwrap_or(AnyFetchGroupOrder::Ascending);
638 let built = (AnyFetchObjectReader::new(h, order), AnyFetchObjectWriter::new(h, order));
639 match built {
640 (Ok(reader), Ok(writer)) => Ok(Stage::Fetch { reader, writer }),
641 // A draft this build did not compile. Not an error — the stream
642 // is simply not addressable. The header decode above would
643 // already have failed on such a draft, so this is a second line
644 // rather than the first.
645 _ => Err(BypassReason::NoFetchObjectCodec),
646 }
647 }
648
649 fn poll_object(&mut self) -> FramerOut {
650 self.counters.note_framer_object_poll();
651
652 // Lazy commit, once, here. The pipe loop is sequential — the
653 // caller's verdict on object *N* always lands before `poll`
654 // produces object *N+1* — so an object still pending at the top of
655 // this call was forwarded. `poll_oversized` deliberately does
656 // **not** repeat this: it is only ever reached from inside this
657 // function, and a second commit would consume a disposition this
658 // one already took. If it ever gains an entry path of its own,
659 // make the commit idempotent rather than adding a second site.
660 self.commit_disposition();
661
662 if self.buf.is_empty() {
663 return FramerOut::NeedMore;
664 }
665
666 let snapshot: &[u8] = &self.buf[..];
667 let mut cursor: &[u8] = snapshot;
668
669 // Probe against a clone: a reader mutated before the object is
670 // known to be complete carries a corrupt delta state into every
671 // later object on the stream.
672 let probed = match &self.stage {
673 Stage::Subgroup(reader) => {
674 let mut probe = reader.clone();
675 probe
676 .read_object_meta(&mut cursor)
677 .map(|m| (Probe::Subgroup(probe), Meta::Sub(m), None))
678 }
679 // `read_object_frame` rather than `read_object_meta`: it reports
680 // the same framing and additionally keeps the shape the frame
681 // arrived in, which is what re-encoding it against a different
682 // predecessor takes.
683 Stage::Fetch { reader, .. } => {
684 let mut probe = reader.clone();
685 probe
686 .read_object_frame(&mut cursor)
687 .map(|frame| (Probe::Fetch(probe), Meta::Fetch(frame.meta), Some(frame)))
688 }
689 Stage::AwaitingHeader => return FramerOut::NeedMore,
690 };
691
692 match probed {
693 Ok((probe, meta, frame)) => {
694 let consumed = snapshot.len() - cursor.remaining();
695 let object_id = meta.object_id();
696 // Before the source bytes are split off, so a failure
697 // leaves them in the buffer for the bypassed poll that
698 // follows to forward verbatim rather than dropping them.
699 let rewritten = match self.apply_elide_fixup(consumed, object_id, frame.as_ref()) {
700 Ok(rewritten) => rewritten,
701 Err(e) => {
702 self.latch_bypass(BypassReason::DecodeError);
703 return FramerOut::Error(e);
704 }
705 };
706 let source = self.buf.split_to(consumed).freeze();
707 let raw = rewritten.unwrap_or(source);
708 self.commit(probe);
709 // An object that arrives in one read completes before
710 // buffering ever reaches the cap, so the incomplete-decode
711 // path below never sees it. Judging it by its own wire
712 // length too is what makes the cap a property of the object
713 // rather than of the caller's read size. Judged on the
714 // *source* length, so an elide fix-up that widens the ID
715 // varint cannot move an object across the boundary.
716 let out = if consumed > self.config.max_buffered_object_bytes {
717 // A `Passthrough` that is a whole object: the meta was
718 // decoded and is about to be discarded, so nothing
719 // outside the framer can address it.
720 self.counters.note_object_not_addressable();
721 FramerOut::Passthrough(raw)
722 } else {
723 FramerOut::Object { meta: self.object_meta(meta), raw }
724 };
725 // Both branches emitted an object, so both arm the cursor:
726 // the oversized one has no `ObjectMeta` for the caller to
727 // name, which is exactly why the framer keeps the record.
728 self.pending_disposition =
729 Some(PendingObject { object_id, index_in_stream: self.index_in_stream });
730 self.index_in_stream += 1;
731 out
732 }
733 Err(e) if is_incomplete_error(&e) => {
734 if self.buf.len() < self.config.max_buffered_object_bytes {
735 FramerOut::NeedMore
736 } else {
737 self.poll_oversized()
738 }
739 }
740 Err(e) => {
741 self.latch_bypass(BypassReason::DecodeError);
742 FramerOut::Error(format!("object decode: {e}"))
743 }
744 }
745 }
746
747 /// The buffer hit its cap without completing an object. Decide whether
748 /// the object's *framing* is understood — in which case its bytes can
749 /// be streamed through and framing resumes afterwards — or whether the
750 /// stream has to be abandoned.
751 ///
752 /// The framing is read against the buffered bytes followed by padding,
753 /// so the decoder can advance past a payload that has not arrived. How
754 /// wide that padding may be is [`Self::measuring_pad`]'s call.
755 fn poll_oversized(&mut self) -> FramerOut {
756 let pad = self.measuring_pad();
757 let mut padded = PaddedBuf::new(&self.buf[..], pad);
758 let probed = match &self.stage {
759 Stage::Subgroup(reader) => {
760 let mut probe = reader.clone();
761 probe.read_object_meta(&mut padded).map(|m| {
762 (Probe::Subgroup(probe), m.object_id, m.wire_len, m.payload_length, None)
763 })
764 }
765 Stage::Fetch { reader, .. } => {
766 let mut probe = reader.clone();
767 probe.read_object_frame(&mut padded).map(|frame| {
768 let meta = frame.meta;
769 (
770 Probe::Fetch(probe),
771 meta.object_id,
772 meta.wire_len,
773 meta.payload_length,
774 Some(frame),
775 )
776 })
777 }
778 Stage::AwaitingHeader => return FramerOut::NeedMore,
779 };
780
781 let Ok((probe, object_id, wire_len, payload_length, frame)) = probed else {
782 self.latch_bypass(BypassReason::ObjectBeyondMeasuringReach);
783 return self.emit_passthrough(u64::MAX);
784 };
785
786 // Only trust the framing when every field ahead of the payload
787 // came from real bytes rather than padding.
788 let framing_len = wire_len - payload_length;
789 let buffered = self.buf.len() as u64;
790 if framing_len > buffered || wire_len <= buffered {
791 self.latch_bypass(BypassReason::ObjectBeyondMeasuringReach);
792 return self.emit_passthrough(u64::MAX);
793 }
794
795 // `framing_len <= buffered` is what makes the fix-up reachable
796 // here: the whole framing, leading Object ID varint included, is
797 // in the chunk about to be emitted. The chunk is a *prefix* of the
798 // object, which `reemit_subgroup_object` accepts — it decodes the
799 // ID field and copies the rest verbatim without validating any
800 // length.
801 let chunk = self.buf.len();
802 let rewritten = match self.apply_elide_fixup(chunk, object_id, frame.as_ref()) {
803 Ok(rewritten) => rewritten,
804 Err(e) => {
805 self.latch_bypass(BypassReason::DecodeError);
806 return FramerOut::Error(e);
807 }
808 };
809
810 self.commit(probe);
811 // The second `Passthrough`-as-an-object path. Counted once here,
812 // not once per emitted chunk: this function is entered once per
813 // oversized object, and the chunks that follow come from `poll`'s
814 // `passthrough_remaining` branch.
815 self.counters.note_object_not_addressable();
816 // Denominated in **source** bytes, and deliberately not adjusted
817 // by the fix-up's `id_bytes_after - id_bytes_before`. Its job is
818 // to say how many more bytes of the *incoming* stream belong to
819 // this object; the emitted stream is a byte or two shorter or
820 // longer for exactly one object, which is what an elide fix-up is.
821 // Correcting it here would make the framer stop consuming this
822 // object early or late and resynchronise mid-way through the next
823 // one — a silent corruption that only fires when the new delta
824 // needs a wider varint.
825 self.passthrough_remaining = wire_len - buffered;
826 self.pending_disposition =
827 Some(PendingObject { object_id, index_in_stream: self.index_in_stream });
828 self.index_in_stream += 1;
829 let source = self.buf.split().freeze();
830 FramerOut::Passthrough(rewritten.unwrap_or(source))
831 }
832
833 /// Commit the disposition of the object emitted most recently.
834 ///
835 /// An object still pending was forwarded; one the caller elided was
836 /// taken out of `pending_disposition` by [`Self::note_elided`], so the
837 /// cursor never advances to it.
838 fn commit_disposition(&mut self) {
839 if let Some(pending) = self.pending_disposition.take() {
840 self.last_forwarded_id = Some(pending.object_id);
841 // Forwarded, so the writer's advance stands and there is nothing
842 // to put back. Cleared rather than left for the next re-emit to
843 // overwrite, so that a rollback can only ever undo the frame it
844 // was taken for.
845 self.fetch_rollback = None;
846 }
847 }
848
849 /// `true` when this stream's Object IDs are written as `id - prev - 1`
850 /// rather than absolutely, so eliding one renumbers every later object.
851 ///
852 /// Subgroup streams on drafts 14-23, and no fetch stream on any draft:
853 /// this is the predicate for the *varint rewrite*, and a fetch frame is
854 /// paid for by [`Self::reemit_fetch_frame`] instead. Drafts 07-13 write
855 /// subgroup Object IDs absolutely and need neither.
856 ///
857 /// The draft half is an exhaustive match rather than a `matches!`: a draft
858 /// left off the list answers "written absolutely", the framer then forwards
859 /// the object after an elide without rewriting its leading varint, and
860 /// every later Object ID on the stream is off by one with nothing to say
861 /// so. The stream-kind half stays a `matches!` — `DataStreamType` is not a
862 /// draft list and its two variants are both named here.
863 fn delta_encodes_object_ids(&self) -> bool {
864 matches!(self.stream_type, DataStreamType::Subgroup)
865 && match self.draft {
866 DraftVersion::Draft07
867 | DraftVersion::Draft08
868 | DraftVersion::Draft09
869 | DraftVersion::Draft10
870 | DraftVersion::Draft11
871 | DraftVersion::Draft12
872 | DraftVersion::Draft13 => false,
873 DraftVersion::Draft14
874 | DraftVersion::Draft15
875 | DraftVersion::Draft16
876 | DraftVersion::Draft17
877 | DraftVersion::Draft18
878 | DraftVersion::Draft19
879 | DraftVersion::Draft20
880 | DraftVersion::Draft21
881 | DraftVersion::Draft22
882 | DraftVersion::Draft23 => true,
883 }
884 }
885
886 /// `true` when eliding an object from this stream leaves the next one
887 /// encoded against something that is no longer on the wire, so a fix-up
888 /// is owed before another object may be forwarded.
889 ///
890 /// The two stream kinds owe it for different reasons and pay it in
891 /// different ways, and the debt itself is the same: until one more object
892 /// has been emitted, the bytes the destination would receive decode to
893 /// Locations nobody sent. `fixup_owed` on a `Bypassed` is what a session
894 /// resets its destination over, and it reads this.
895 ///
896 /// Fetch streams on drafts 15-23, where a Serialization Flags field lets
897 /// a frame take any of its Group ID, Subgroup ID, Object ID and Priority
898 /// from the frame before it — draft-17 Section 10.4.4.1, Table 7: "Object
899 /// ID is the prior Object's ID plus one". Not drafts 07-14, whose fetch
900 /// objects state all four outright.
901 ///
902 /// Exhaustive rather than a `matches!` for the same reason as
903 /// [`Self::delta_encodes_object_ids`], and the consequence is larger here:
904 /// `false` means no fix-up is owed, so the session never resets the
905 /// destination and the receiver keeps a stream whose frames decode to
906 /// Locations nobody sent.
907 fn elide_owes_a_fixup(&self) -> bool {
908 match self.stream_type {
909 DataStreamType::Subgroup => self.delta_encodes_object_ids(),
910 DataStreamType::Fetch => match self.draft {
911 DraftVersion::Draft07
912 | DraftVersion::Draft08
913 | DraftVersion::Draft09
914 | DraftVersion::Draft10
915 | DraftVersion::Draft11
916 | DraftVersion::Draft12
917 | DraftVersion::Draft13
918 | DraftVersion::Draft14 => false,
919 DraftVersion::Draft15
920 | DraftVersion::Draft16
921 | DraftVersion::Draft17
922 | DraftVersion::Draft18
923 | DraftVersion::Draft19
924 | DraftVersion::Draft20
925 | DraftVersion::Draft21
926 | DraftVersion::Draft22
927 | DraftVersion::Draft23 => true,
928 },
929 }
930 }
931
932 /// Rewrite the leading Object ID varint of `self.buf[..len]` when an
933 /// elide has left the wire's delta chain one object ahead of what was
934 /// actually forwarded.
935 ///
936 /// `len` may be a prefix of the object rather than the whole of it —
937 /// the oversized path calls this with the first chunk. Returns
938 /// `Ok(None)` when no fix-up was owed, in which case the caller emits
939 /// the source bytes untouched; the borrow of `self.buf` ends before
940 /// this returns, so the caller is free to split it afterwards.
941 ///
942 /// The error is unreachable in practice — Object IDs are strictly
943 /// increasing within a stream and the caller has already established
944 /// that the ID field is whole in the buffer — but it is reported
945 /// rather than swallowed, because the alternative is emitting bytes
946 /// known to decode to the wrong ID.
947 fn apply_elide_fixup(
948 &mut self,
949 len: usize,
950 object_id: u64,
951 frame: Option<&AnyFetchFrame>,
952 ) -> Result<Option<Bytes>, String> {
953 // `frame` is `Some` exactly on a fetch stream, where the payment is a
954 // re-encode of the whole framing rather than a rewrite of one varint,
955 // and where it is the writer rather than a flag that decides whether
956 // anything is owed.
957 if let Some(frame) = frame {
958 return self.reemit_fetch_frame(len, frame);
959 }
960 if !self.fixup_pending || !self.delta_encodes_object_ids() {
961 return Ok(None);
962 }
963 let mut out = BytesMut::with_capacity(len + 8);
964 let outcome = reemit_subgroup_object(
965 self.draft,
966 self.last_forwarded_id,
967 object_id,
968 &self.buf[..len],
969 &mut out,
970 );
971 if let Err(e) = outcome {
972 // Leaves `fixup_pending` set, so the cursor keeps reporting
973 // that this stream still owes a fix-up.
974 return Err(format!("elide fix-up: {e}"));
975 }
976 self.fixup_pending = false;
977 self.counters.note_object_id_rewritten();
978 Ok(Some(out.freeze()))
979 }
980
981 /// Re-emit one fetch frame through this stream's writer.
982 ///
983 /// Called for **every** fetch frame the framer emits, not only after an
984 /// elide. The writer is what a survivor is re-encoded against, and it
985 /// only moves when it is shown a frame, so skipping the frames nothing
986 /// was owed for would leave it as many frames behind as were skipped.
987 ///
988 /// Answers `Ok(None)` when the frame's own bytes still say what the frame
989 /// says, which is every frame on a stream nothing has been removed from.
990 /// The writer is cloned first so that [`Self::note_elided`] can put it
991 /// back: this runs before the caller's verdict on the frame, and a frame
992 /// the caller drops must leave no trace on the writing side.
993 ///
994 /// `len` may be a prefix of the frame — the oversized path calls this
995 /// with the first chunk — and the codec asks only that the whole framing
996 /// be inside it, which is what the caller has already established.
997 fn reemit_fetch_frame(
998 &mut self,
999 len: usize,
1000 frame: &AnyFetchFrame,
1001 ) -> Result<Option<Bytes>, String> {
1002 let mut out = BytesMut::with_capacity(len + 16);
1003 let (outcome, rollback) = {
1004 let Stage::Fetch { writer, .. } = &mut self.stage else {
1005 return Ok(None);
1006 };
1007 let rollback = writer.clone();
1008 (writer.reemit_object(frame, &self.buf[..len], &mut out), rollback)
1009 };
1010 self.fetch_rollback = Some(rollback);
1011
1012 match outcome {
1013 Ok(FetchReemit::Unchanged) => {
1014 // The debt such as it was is settled either way. A removal
1015 // whose survivor happened to state every field it needed
1016 // costs nothing, and leaving the flag set would have the
1017 // stream report an unpaid fix-up at every later bypass.
1018 self.fixup_pending = false;
1019 Ok(None)
1020 }
1021 Ok(FetchReemit::Reframed { .. }) => {
1022 self.fixup_pending = false;
1023 self.counters.note_object_id_rewritten();
1024 Ok(Some(out.freeze()))
1025 }
1026 // Leaves `fixup_pending` set, as the subgroup path does, so the
1027 // stream keeps reporting that it still owes one.
1028 Err(e) => Err(format!("elide fix-up: {e}")),
1029 }
1030 }
1031
1032 /// How much padding [`Self::poll_oversized`] may put behind the
1033 /// buffered bytes when measuring an object that has not arrived whole.
1034 ///
1035 /// The pad costs no memory by itself — it is never materialised — but
1036 /// every copying read in the codec checks `Buf::remaining()` before it
1037 /// allocates, so the pad is exactly the bound on the allocation a
1038 /// hostile length field can provoke. That makes the choice per layout
1039 /// rather than global:
1040 ///
1041 /// * Subgroup objects on drafts 14-23, and fetch objects on draft-14,
1042 /// are measured without a single copy — their `read_object_meta`
1043 /// advances past the extension block and the payload rather than
1044 /// reading them. Nothing can be talked into allocating, so the pad is
1045 /// effectively unbounded and an object of *any* declared size stays
1046 /// measurable: it streams through and framing resumes after it.
1047 /// * Every other object layout decodes its extension block by copying
1048 /// it out of the buffer. Those keep a pad of one cap, which bounds
1049 /// the copy at the price of reach: an object whose wire length
1050 /// exceeds `buffered + cap` cannot be measured and the stream falls
1051 /// back to passthrough for its remainder.
1052 ///
1053 /// Widening the second case needs the codec to reject an extension
1054 /// length larger than the bytes actually present, which is not this
1055 /// crate's to change.
1056 ///
1057 /// **"Every other" includes the newest fetch layouts, and that is not an
1058 /// omission.** Drafts 15 through 23 all read a fetch frame through
1059 /// `FetchObjectReader::read_object_header`, which materialises the frame's
1060 /// properties — `data_dispatch.rs` `fo15`..`fo23` each carry the block out
1061 /// of the buffer before `read_object_frame` skips the payload — so a
1062 /// declared extension length is still an allocation those drafts can be
1063 /// talked into. Only draft-14's `FetchObject::decode_meta` advances past
1064 /// the block instead. Both arms are exhaustive matches rather than
1065 /// `matches!` so that this stays a decision: a new draft answering `false`
1066 /// by omission would be *safe* here, which is precisely why nothing would
1067 /// ever notice that no one had looked at its fetch layout.
1068 fn measuring_pad(&self) -> usize {
1069 let copy_free = match self.stage {
1070 Stage::Subgroup(_) => match self.draft {
1071 DraftVersion::Draft07
1072 | DraftVersion::Draft08
1073 | DraftVersion::Draft09
1074 | DraftVersion::Draft10
1075 | DraftVersion::Draft11
1076 | DraftVersion::Draft12
1077 | DraftVersion::Draft13 => false,
1078 DraftVersion::Draft14
1079 | DraftVersion::Draft15
1080 | DraftVersion::Draft16
1081 | DraftVersion::Draft17
1082 | DraftVersion::Draft18
1083 | DraftVersion::Draft19
1084 | DraftVersion::Draft20
1085 | DraftVersion::Draft21
1086 | DraftVersion::Draft22
1087 | DraftVersion::Draft23 => true,
1088 },
1089 Stage::Fetch { .. } => match self.draft {
1090 DraftVersion::Draft14 => true,
1091 DraftVersion::Draft07
1092 | DraftVersion::Draft08
1093 | DraftVersion::Draft09
1094 | DraftVersion::Draft10
1095 | DraftVersion::Draft11
1096 | DraftVersion::Draft12
1097 | DraftVersion::Draft13
1098 | DraftVersion::Draft15
1099 | DraftVersion::Draft16
1100 | DraftVersion::Draft17
1101 | DraftVersion::Draft18
1102 | DraftVersion::Draft19
1103 | DraftVersion::Draft20
1104 | DraftVersion::Draft21
1105 | DraftVersion::Draft22
1106 | DraftVersion::Draft23 => false,
1107 },
1108 Stage::AwaitingHeader => false,
1109 };
1110 if copy_free {
1111 UNBOUNDED_MEASURING_PAD
1112 } else {
1113 self.config.max_buffered_object_bytes
1114 }
1115 }
1116
1117 /// Adopt a probe reader whose decode succeeded.
1118 fn commit(&mut self, probe: Probe) {
1119 match (&mut self.stage, probe) {
1120 (Stage::Subgroup(reader), Probe::Subgroup(probe)) => *reader = probe,
1121 (Stage::Fetch { reader, .. }, Probe::Fetch(probe)) => *reader = probe,
1122 // `probe` was cloned from `stage` a few lines earlier, so the
1123 // pairing always matches.
1124 _ => {}
1125 }
1126 }
1127
1128 fn object_meta(&self, meta: Meta) -> ObjectMeta {
1129 match meta {
1130 Meta::Sub(m) => ObjectMeta {
1131 draft: self.draft,
1132 stream_kind: DataStreamType::Subgroup,
1133 track_alias: Some(self.context.track_alias),
1134 group_id: self.context.group_id,
1135 subgroup_id: self.context.subgroup_id,
1136 object_id: m.object_id,
1137 publisher_priority: self.context.publisher_priority,
1138 index_in_stream: self.index_in_stream,
1139 payload_len: m.payload_length,
1140 status: m.status,
1141 end_of_range: None,
1142 },
1143 Meta::Fetch(m) => ObjectMeta {
1144 draft: self.draft,
1145 stream_kind: DataStreamType::Fetch,
1146 track_alias: None,
1147 group_id: m.group_id,
1148 // Not `Some(m.subgroup_id)`: from draft-15 a fetch frame
1149 // may carry no Subgroup ID at all — an object forwarded
1150 // over a datagram has no subgroup, and an End of Range
1151 // indicator names a Location rather than an object — and
1152 // the codec leaves a placeholder in the field when it says
1153 // so. Forwarding that placeholder would key a matcher on a
1154 // subgroup the publisher never named.
1155 subgroup_id: m.has_subgroup_id.then_some(m.subgroup_id),
1156 object_id: m.object_id,
1157 publisher_priority: Some(m.publisher_priority),
1158 index_in_stream: self.index_in_stream,
1159 payload_len: m.payload_length,
1160 status: m.status,
1161 end_of_range: m.end_of_range,
1162 },
1163 }
1164 }
1165}
1166
1167/// A reader clone whose decode succeeded and is ready to be adopted.
1168enum Probe {
1169 Subgroup(AnySubgroupObjectReader),
1170 Fetch(AnyFetchObjectReader),
1171}
1172
1173/// The codec-level framing of one object, either stream kind.
1174#[derive(Clone, Copy)]
1175enum Meta {
1176 Sub(AnySubgroupObjectMeta),
1177 Fetch(AnyFetchObjectMeta),
1178}
1179
1180impl Meta {
1181 /// The object's absolute Object ID, delta encoding already resolved.
1182 fn object_id(self) -> u64 {
1183 match self {
1184 Meta::Sub(m) => m.object_id,
1185 Meta::Fetch(m) => m.object_id,
1186 }
1187 }
1188}
1189
1190/// The identity fields a subgroup header contributes to its objects.
1191///
1192/// Every field comes from an [`AnySubgroupHeader`] accessor, so the
1193/// per-draft knowledge behind them — which stream types leave the subgroup
1194/// ID to the first object, which drafts fold a mode field into the
1195/// header-type octet, which drafts may omit the publisher priority — stays
1196/// in the codec beside the decoders that define it. This crate keeps no
1197/// copy of it, which is the point: a local copy (a mask constant and a
1198/// per-draft match) drifts from the decoders it copies.
1199///
1200/// Where a draft encodes the subgroup ID implicitly as the first object's
1201/// ID, the codec stores zero and [`AnySubgroupHeader::subgroup_id`]
1202/// reports `None` rather than that zero.
1203fn subgroup_context(header: &AnySubgroupHeader) -> SubgroupContext {
1204 SubgroupContext {
1205 track_alias: header.track_alias(),
1206 group_id: header.group_id(),
1207 subgroup_id: header.subgroup_id(),
1208 publisher_priority: header.publisher_priority(),
1209 }
1210}
1211
1212/// A slice followed by a run of zero bytes.
1213///
1214/// Lets a decoder walk an object's framing and advance past a payload that
1215/// has not arrived yet, so the framer can learn an object's total wire
1216/// length without buffering it. Only fields decoded from the real prefix
1217/// are trustworthy; the caller checks that before using the result, and
1218/// picks the pad width from what the decode path may copy — see
1219/// `ObjectFramer::measuring_pad`.
1220struct PaddedBuf<'a> {
1221 real: &'a [u8],
1222 pad: usize,
1223}
1224
1225/// Backing bytes for [`PaddedBuf`]'s padded region.
1226const ZEROS: [u8; 1024] = [0u8; 1024];
1227
1228impl<'a> PaddedBuf<'a> {
1229 fn new(real: &'a [u8], pad: usize) -> Self {
1230 Self { real, pad }
1231 }
1232}
1233
1234impl Buf for PaddedBuf<'_> {
1235 fn remaining(&self) -> usize {
1236 self.real.len() + self.pad
1237 }
1238
1239 fn chunk(&self) -> &[u8] {
1240 if self.real.is_empty() {
1241 &ZEROS[..self.pad.min(ZEROS.len())]
1242 } else {
1243 self.real
1244 }
1245 }
1246
1247 fn advance(&mut self, cnt: usize) {
1248 let from_real = cnt.min(self.real.len());
1249 self.real = &self.real[from_real..];
1250 self.pad = self.pad.saturating_sub(cnt - from_real);
1251 }
1252}
1253
1254// Every test below builds its input with a codec writer and reads it back
1255// with a codec reader, so every one of them needs at least one draft
1256// compiled in. With none, `AnySubgroupHeader` and friends are uninhabited
1257// enums and the helpers stop type-checking — `-D warnings` reports the
1258// `decode_stream` call as an unreachable definition. Gating the module is
1259// the honest shape: with no draft there is no framing to test, so the
1260// module vanishes rather than being kept alive by `#[allow]`.
1261#[cfg(test)]
1262#[cfg(any(
1263 feature = "draft07",
1264 feature = "draft08",
1265 feature = "draft09",
1266 feature = "draft10",
1267 feature = "draft11",
1268 feature = "draft12",
1269 feature = "draft13",
1270 feature = "draft14",
1271 feature = "draft15",
1272 feature = "draft16",
1273 feature = "draft17",
1274 feature = "draft18",
1275 feature = "draft19",
1276 feature = "draft20",
1277 feature = "draft21",
1278 feature = "draft22",
1279 feature = "draft23"
1280))]
1281mod tests {
1282 //! The elide cursor, on the wire.
1283 //! Everything here builds its input with the codec's *writer* and checks
1284 //! the framer's output by *decoding* it again — so the claim under test
1285 //! (*after eliding object N the stream still decodes to the IDs that
1286 //! survived*) is settled by the codec's readers rather than by the rewrite
1287 //! logic being tested. Byte identity in the absence of an elide stays the
1288 //! property `tests/framer_tests.rs` and
1289 //! `tests/object_framing_acceptance.rs` own; the fence here only pins that
1290 //! the new code does not fire when nothing asked it to.
1291
1292 use super::*;
1293
1294 use moqtap_codec::dispatch::{AnySubgroupObject, AnySubgroupObjectWriter};
1295
1296 /// The drafts this build actually compiled.
1297 ///
1298 /// Each element carries its own `#[cfg]`, so the sweep is the enabled
1299 /// set rather than a hardcoded list: a `--features draft14` build
1300 /// sweeps one draft and does not try to decode the headers whose
1301 /// decoders were not compiled.
1302 const DRAFTS: &[DraftVersion] = &[
1303 #[cfg(feature = "draft07")]
1304 DraftVersion::Draft07,
1305 #[cfg(feature = "draft08")]
1306 DraftVersion::Draft08,
1307 #[cfg(feature = "draft09")]
1308 DraftVersion::Draft09,
1309 #[cfg(feature = "draft10")]
1310 DraftVersion::Draft10,
1311 #[cfg(feature = "draft11")]
1312 DraftVersion::Draft11,
1313 #[cfg(feature = "draft12")]
1314 DraftVersion::Draft12,
1315 #[cfg(feature = "draft13")]
1316 DraftVersion::Draft13,
1317 #[cfg(feature = "draft14")]
1318 DraftVersion::Draft14,
1319 #[cfg(feature = "draft15")]
1320 DraftVersion::Draft15,
1321 #[cfg(feature = "draft16")]
1322 DraftVersion::Draft16,
1323 #[cfg(feature = "draft17")]
1324 DraftVersion::Draft17,
1325 #[cfg(feature = "draft18")]
1326 DraftVersion::Draft18,
1327 #[cfg(feature = "draft19")]
1328 DraftVersion::Draft19,
1329 #[cfg(feature = "draft20")]
1330 DraftVersion::Draft20,
1331 #[cfg(feature = "draft21")]
1332 DraftVersion::Draft21,
1333 #[cfg(feature = "draft22")]
1334 DraftVersion::Draft22,
1335 #[cfg(feature = "draft23")]
1336 DraftVersion::Draft23,
1337 ];
1338
1339 /// The drafts that write an Object ID as `id - prev - 1`, and so are
1340 /// the only ones where eliding an object corrupts its successors —
1341 /// intersected with the ones this build compiled.
1342 const DELTA_DRAFTS: &[DraftVersion] = &[
1343 #[cfg(feature = "draft14")]
1344 DraftVersion::Draft14,
1345 #[cfg(feature = "draft15")]
1346 DraftVersion::Draft15,
1347 #[cfg(feature = "draft16")]
1348 DraftVersion::Draft16,
1349 #[cfg(feature = "draft17")]
1350 DraftVersion::Draft17,
1351 #[cfg(feature = "draft18")]
1352 DraftVersion::Draft18,
1353 #[cfg(feature = "draft19")]
1354 DraftVersion::Draft19,
1355 #[cfg(feature = "draft20")]
1356 DraftVersion::Draft20,
1357 #[cfg(feature = "draft21")]
1358 DraftVersion::Draft21,
1359 #[cfg(feature = "draft22")]
1360 DraftVersion::Draft22,
1361 #[cfg(feature = "draft23")]
1362 DraftVersion::Draft23,
1363 ];
1364
1365 /// The stream-type field opening a subgroup stream that carries an
1366 /// explicit subgroup ID and no extensions, per draft.
1367 fn subgroup_stream_type(draft: DraftVersion) -> u8 {
1368 match draft {
1369 DraftVersion::Draft07
1370 | DraftVersion::Draft08
1371 | DraftVersion::Draft09
1372 | DraftVersion::Draft10 => 0x04,
1373 DraftVersion::Draft11 => 0x0C,
1374 _ => 0x14,
1375 }
1376 }
1377
1378 /// Wire bytes of a subgroup header: track alias 1, group 0, subgroup 0,
1379 /// publisher priority 128, no extensions.
1380 fn header_bytes(draft: DraftVersion) -> Vec<u8> {
1381 vec![subgroup_stream_type(draft), 0x01, 0x00, 0x00, 0x80]
1382 }
1383
1384 fn object(object_id: u64, payload_len: usize) -> AnySubgroupObject {
1385 AnySubgroupObject {
1386 object_id,
1387 extension_headers: Vec::new(),
1388 extension_count: None,
1389 status: None,
1390 payload: vec![0xAB; payload_len],
1391 }
1392 }
1393
1394 /// Encode a whole subgroup stream: header plus `objects`, on `draft`.
1395 fn build_stream(draft: DraftVersion, objects: &[AnySubgroupObject]) -> Vec<u8> {
1396 let head = header_bytes(draft);
1397 let mut cursor: &[u8] = &head;
1398 let header = AnySubgroupHeader::decode_stream(draft, &mut cursor)
1399 .unwrap_or_else(|e| panic!("[{draft}] header decode: {e}"));
1400 let mut writer = AnySubgroupObjectWriter::new(&header)
1401 .unwrap_or_else(|e| panic!("[{draft}] writer: {e}"));
1402
1403 let mut out = head;
1404 for obj in objects {
1405 writer
1406 .write_object(obj, &mut out)
1407 .unwrap_or_else(|e| panic!("[{draft}] write object {}: {e}", obj.object_id));
1408 }
1409 out
1410 }
1411
1412 /// A subgroup framer on `draft`, reporting into `counters`.
1413 ///
1414 /// Every framer in this module goes through
1415 /// [`ObjectFramer::with_recorder`] rather than [`ObjectFramer::new`]:
1416 /// it keeps `src/` free of the string standing gate 7 greps for
1417 /// without the gate having to filter comments, and it gives the
1418 /// counter assertions below something to read.
1419 fn framer_for(
1420 draft: DraftVersion,
1421 config: FramerConfig,
1422 counters: &Arc<Recorder>,
1423 ) -> ObjectFramer {
1424 ObjectFramer::with_recorder(DataStreamType::Subgroup, draft, config, Arc::clone(counters))
1425 }
1426
1427 /// What one run of the framer produced.
1428 struct Run {
1429 /// Every byte the framer emitted, in order.
1430 bytes: Vec<u8>,
1431 /// The Object IDs the framer itself resolved while producing them.
1432 ///
1433 /// Objects that left as `Passthrough` are absent — they were never
1434 /// addressable — so this is also the record of *where framing
1435 /// resumed* after one of them.
1436 framed_ids: Vec<u64>,
1437 /// The run's own counters.
1438 counters: Arc<Recorder>,
1439 }
1440
1441 /// Feed `stream` to a framer in `chunk`-sized pieces, dropping every
1442 /// object whose ID is in `elide`, and return what the framer produced.
1443 ///
1444 /// A bypass is a panic rather than a recorded outcome: every test that
1445 /// uses this helper is about an elide surviving, and a stream that
1446 /// quietly stopped being parsed would satisfy the byte assertions
1447 /// while proving nothing.
1448 fn run_eliding(
1449 draft: DraftVersion,
1450 stream: &[u8],
1451 chunk: usize,
1452 config: FramerConfig,
1453 elide: &[u64],
1454 ) -> Run {
1455 let counters = Arc::new(Recorder::new());
1456 let mut framer = framer_for(draft, config, &counters);
1457 let mut run = Run { bytes: Vec::new(), framed_ids: Vec::new(), counters };
1458 for piece in stream.chunks(chunk.max(1)) {
1459 framer.feed(piece);
1460 loop {
1461 match framer.poll() {
1462 FramerOut::NeedMore => break,
1463 FramerOut::Header { raw, .. } => run.bytes.extend_from_slice(&raw),
1464 FramerOut::Object { meta, raw } => {
1465 run.framed_ids.push(meta.object_id);
1466 if elide.contains(&meta.object_id) {
1467 framer.note_elided(&meta);
1468 } else {
1469 run.bytes.extend_from_slice(&raw);
1470 }
1471 }
1472 FramerOut::Passthrough(raw) => run.bytes.extend_from_slice(&raw),
1473 FramerOut::Bypassed { reason, fixup_owed } => {
1474 panic!("[{draft}] unexpected bypass: {reason:?}, fixup_owed {fixup_owed}")
1475 }
1476 FramerOut::Error(e) => panic!("[{draft}] framer error: {e}"),
1477 }
1478 }
1479 }
1480 if let Some(tail) = framer.finish() {
1481 run.bytes.extend_from_slice(&tail);
1482 }
1483 run
1484 }
1485
1486 /// Decode `bytes` as a subgroup stream and report the Object IDs the
1487 /// codec's readers resolve from it.
1488 ///
1489 /// The framer used here has the default cap, so every object is framed
1490 /// individually no matter how the producing run chose to emit it.
1491 fn framed_object_ids(draft: DraftVersion, bytes: &[u8]) -> Vec<u64> {
1492 let counters = Arc::new(Recorder::new());
1493 let mut framer = framer_for(draft, FramerConfig::default(), &counters);
1494 framer.feed(bytes);
1495 let mut ids = Vec::new();
1496 loop {
1497 match framer.poll() {
1498 FramerOut::NeedMore => break,
1499 FramerOut::Header { .. } => {}
1500 FramerOut::Object { meta, .. } => ids.push(meta.object_id),
1501 FramerOut::Passthrough(raw) => {
1502 panic!("[{draft}] re-decode fell back to passthrough after {} bytes", raw.len())
1503 }
1504 FramerOut::Bypassed { reason, .. } => {
1505 panic!("[{draft}] re-decode bypassed: {reason:?}")
1506 }
1507 FramerOut::Error(e) => panic!("[{draft}] re-decode: {e}"),
1508 }
1509 }
1510 assert_eq!(framer.buffered(), 0, "[{draft}] re-decode left bytes buffered");
1511 ids
1512 }
1513
1514 /// Eliding an object leaves a stream that still decodes to exactly the
1515 /// objects that survived, on every draft.
1516 ///
1517 /// On drafts 14-23 that is only true because the framer rewrites the
1518 /// next object's leading Object ID varint; on 07-13 the IDs are
1519 /// absolute and the surviving bytes already say the truth.
1520 #[test]
1521 fn eliding_an_object_renumbers_the_rest_of_the_stream() {
1522 for &draft in DRAFTS {
1523 let objects: Vec<_> = (0..4).map(|id| object(id, 16)).collect();
1524 let stream = build_stream(draft, &objects);
1525
1526 let run = run_eliding(draft, &stream, stream.len(), FramerConfig::default(), &[1]);
1527
1528 assert_eq!(
1529 framed_object_ids(draft, &run.bytes),
1530 vec![0, 2, 3],
1531 "[{draft}] elided stream must decode to the surviving IDs"
1532 );
1533 }
1534 }
1535
1536 /// The elide survives an object the framer cannot address
1537 /// individually.
1538 ///
1539 /// Object 65 is over the buffer cap, so it leaves as `Passthrough` with
1540 /// no `ObjectMeta` for anyone outside the framer to name. A cursor
1541 /// owned by the caller could not renumber it, and its stale delta would
1542 /// shift every later object for the rest of the stream. The IDs are
1543 /// chosen so the correction changes the varint's *width*:
1544 /// `65 - 1 - 1 = 63` is one byte, `65 - 0 - 1 = 64` is two.
1545 #[test]
1546 fn an_oversized_object_after_an_elide_is_still_renumbered() {
1547 for &draft in DRAFTS {
1548 let objects = vec![object(0, 8), object(1, 8), object(65, 4096), object(66, 8)];
1549 let stream = build_stream(draft, &objects);
1550 let config = FramerConfig::new().with_max_buffered_object_bytes(64);
1551
1552 // Whole stream in one feed: the oversized object decodes
1553 // completely and `poll_object` routes it by its own wire length.
1554 let run = run_eliding(draft, &stream, stream.len(), config, &[1]);
1555 assert_eq!(
1556 run.framed_ids,
1557 vec![0, 1, 66],
1558 "[{draft}] object 65 left as passthrough and framing resumed on 66"
1559 );
1560 assert_eq!(
1561 framed_object_ids(draft, &run.bytes),
1562 vec![0, 65, 66],
1563 "[{draft}] oversized successor of an elide, decoded whole"
1564 );
1565 }
1566 }
1567
1568 /// The same, through `poll_oversized`: the object never arrives whole,
1569 /// so it is measured against padding and streamed out in chunks.
1570 ///
1571 /// This is the path where `passthrough_remaining` matters. It is
1572 /// computed from the *source* wire length while the emitted chunk is a
1573 /// byte longer, and the objects are picked so that gap is real.
1574 ///
1575 /// Restricted to the six delta drafts because they are the only ones
1576 /// whose object layout the codec can measure past an unarrived payload;
1577 /// on 07-13 a 4 KiB object with a 64-byte cap is out of measuring reach
1578 /// and the stream bypasses instead, which is a different property.
1579 #[test]
1580 fn an_elide_survives_an_object_streamed_through_in_chunks() {
1581 for &draft in DELTA_DRAFTS {
1582 let objects = vec![object(0, 8), object(1, 8), object(65, 4096), object(66, 8)];
1583 let stream = build_stream(draft, &objects);
1584 let config = FramerConfig::new().with_max_buffered_object_bytes(64);
1585
1586 let run = run_eliding(draft, &stream, 32, config, &[1]);
1587 // The accounting claim: `passthrough_remaining` is denominated
1588 // in source bytes, so the framer consumes object 65 exactly and
1589 // resynchronises on object 66's first byte — even though what
1590 // it emitted for object 65 is a byte longer than what it read.
1591 assert_eq!(
1592 run.framed_ids,
1593 vec![0, 1, 66],
1594 "[{draft}] framing must resume exactly on object 66"
1595 );
1596 assert_eq!(
1597 framed_object_ids(draft, &run.bytes),
1598 vec![0, 65, 66],
1599 "[{draft}] oversized successor of an elide, streamed through"
1600 );
1601 }
1602 }
1603
1604 /// A framer nobody elides on emits its input back, byte for byte.
1605 ///
1606 /// The regression fence for the fix-up code: it must be inert until
1607 /// [`ObjectFramer::note_elided`] is called. This is the property the
1608 /// acceptance suite rests on, restated where the code that could break
1609 /// it lives.
1610 #[test]
1611 fn a_stream_with_no_elide_is_emitted_byte_for_byte() {
1612 for &draft in DRAFTS {
1613 let objects: Vec<_> = (0..4).map(|id| object(id, 16)).collect();
1614 let stream = build_stream(draft, &objects);
1615
1616 for chunk in [1usize, 7, stream.len()] {
1617 let run = run_eliding(draft, &stream, chunk, FramerConfig::default(), &[]);
1618 assert_eq!(run.bytes, stream, "[{draft}] chunk {chunk}: byte identity");
1619 }
1620 }
1621 }
1622
1623 /// A non-minimally encoded Object ID field survives untouched when
1624 /// nothing was elided.
1625 ///
1626 /// QUIC varints may be written wider than they need to be, and a
1627 /// producer that does so is still emitting a legal stream. Re-encoding
1628 /// such a field would change bytes the proxy was asked to forward
1629 /// unchanged — so the fix-up must be reached only when an elide
1630 /// actually armed it, not merely when the draft delta-encodes.
1631 #[test]
1632 fn a_widened_object_id_field_is_not_re_encoded() {
1633 for &draft in DELTA_DRAFTS {
1634 let stream = build_stream(draft, &[object(0, 8), object(1, 8)]);
1635
1636 // Object 0 starts right after the five header bytes, and its
1637 // leading field is the one-byte varint 0. The two-byte form of the
1638 // same value differs by draft: `0x40 0x00` under RFC 9000, and
1639 // `0x80 0x00` from draft-17, where `0x40` is the one-byte 64.
1640 let head = header_bytes(draft).len();
1641 assert_eq!(stream[head], 0x00, "[{draft}] object 0's ID field is not where expected");
1642 let two_byte_zero: [u8; 2] =
1643 if draft.uses_moqt_varint() { [0x80, 0x00] } else { [0x40, 0x00] };
1644 let mut widened = stream[..head].to_vec();
1645 widened.extend_from_slice(&two_byte_zero);
1646 widened.extend_from_slice(&stream[head + 1..]);
1647
1648 let run = run_eliding(draft, &widened, widened.len(), FramerConfig::default(), &[]);
1649 assert_eq!(run.framed_ids, vec![0, 1], "[{draft}] widened field must still decode");
1650 assert_eq!(run.bytes, widened, "[{draft}] widened field must be forwarded verbatim");
1651 }
1652 }
1653
1654 /// No draft-16 subgroup header the decoder accepts sets both the
1655 /// explicit-subgroup-ID bit and the first-object bit, so the proxy
1656 /// never has to choose between them.
1657 ///
1658 /// Setting `0x04` (explicit subgroup ID) and `0x02` (subgroup ID is the
1659 /// first object's) together puts bits one and two at `0b11`, and
1660 /// draft-16 Section 10.4.2 reserves that Subgroup ID mode. The decoder
1661 /// refuses all eight bytes that spell it, so no such header decodes,
1662 /// and the sweep below covers the whole reserved mode rather than one
1663 /// byte of it.
1664 ///
1665 /// Why the proxy reads the field through the uniform accessor rather
1666 /// than off the header's bits: those two bits are one two-bit field,
1667 /// and a predicate reading `0x02` alone answers `true` for the reserved
1668 /// mode as well — at the same time as one reading `0x04` alone, for a
1669 /// state one two-bit field cannot be in.
1670 /// [`AnySubgroupHeader::subgroup_id`] reads the mode as a field
1671 /// instead: it answers `None` for the reserved mode — the header
1672 /// determines no Subgroup ID — and reads the explicit field only for
1673 /// mode 2, which is the only mode the decoder reads one for.
1674 #[test]
1675 #[cfg(feature = "draft16")]
1676 fn draft16_refuses_the_reserved_subgroup_id_mode() {
1677 let draft = DraftVersion::Draft16;
1678 // Bits one and two are the Subgroup ID mode; 0b11 is reserved.
1679 // These are the eight subgroup types that set it, with and without
1680 // the extensions and priority bits.
1681 for ty in [0x16u8, 0x17, 0x1E, 0x1F, 0x36, 0x37, 0x3E, 0x3F] {
1682 // Track alias 1, group 0, subgroup 42, publisher priority 128 —
1683 // a complete body, so a refusal is of the type and never of a
1684 // short buffer.
1685 let head = vec![ty, 0x01, 0x00, 0x2A, 0x80];
1686 let mut cursor: &[u8] = &head;
1687 assert!(
1688 AnySubgroupHeader::decode_stream(draft, &mut cursor).is_err(),
1689 "type {ty:#04x} sets the reserved Subgroup ID mode and must be refused",
1690 );
1691 }
1692 }
1693
1694 /// The cursor advances lazily, one object behind the framer.
1695 ///
1696 /// An object's disposition is not known when it is emitted — the caller
1697 /// decides afterwards — so `last_forwarded_id` moves at the top of the
1698 /// *next* `poll_object`. The fix-up flag is the part that must be
1699 /// visible immediately, because it is what a bypass has to report as
1700 /// still owed.
1701 #[test]
1702 fn the_elide_cursor_tracks_the_last_object_actually_forwarded() {
1703 for &draft in DRAFTS {
1704 let delta = DELTA_DRAFTS.contains(&draft);
1705 let objects: Vec<_> = (0..3).map(|id| object(id, 8)).collect();
1706 let stream = build_stream(draft, &objects);
1707
1708 let counters = Arc::new(Recorder::new());
1709 let mut framer = framer_for(draft, FramerConfig::default(), &counters);
1710 framer.feed(&stream);
1711
1712 assert_eq!(
1713 framer.elide_cursor(),
1714 ElideCursor { last_forwarded_id: None, fixup_pending: false },
1715 "[{draft}] fresh framer"
1716 );
1717
1718 assert!(matches!(framer.poll(), FramerOut::Header { .. }));
1719
1720 // Object 0, forwarded.
1721 let FramerOut::Object { meta, .. } = framer.poll() else {
1722 panic!("[{draft}] expected object 0");
1723 };
1724 assert_eq!(meta.object_id, 0);
1725 assert_eq!(
1726 framer.elide_cursor().last_forwarded_id,
1727 None,
1728 "[{draft}] object 0's disposition is not known yet"
1729 );
1730
1731 // Object 1, elided. Polling for it commits object 0 first.
1732 let FramerOut::Object { meta, .. } = framer.poll() else {
1733 panic!("[{draft}] expected object 1");
1734 };
1735 assert_eq!(meta.object_id, 1);
1736 assert_eq!(
1737 framer.elide_cursor().last_forwarded_id,
1738 Some(0),
1739 "[{draft}] object 0 committed as forwarded"
1740 );
1741 framer.note_elided(&meta);
1742 assert_eq!(
1743 framer.elide_cursor(),
1744 ElideCursor { last_forwarded_id: Some(0), fixup_pending: delta },
1745 "[{draft}] the elide is visible at once, and only delta drafts owe a fix-up"
1746 );
1747
1748 // Object 2 clears the fix-up as it is emitted, and the cursor
1749 // still names object 0 until the object after it is asked for.
1750 let FramerOut::Object { meta, .. } = framer.poll() else {
1751 panic!("[{draft}] expected object 2");
1752 };
1753 assert_eq!(meta.object_id, 2);
1754 assert_eq!(
1755 framer.elide_cursor(),
1756 ElideCursor { last_forwarded_id: Some(0), fixup_pending: false },
1757 "[{draft}] the fix-up is spent on the object that carried it"
1758 );
1759
1760 assert!(matches!(framer.poll(), FramerOut::NeedMore));
1761 assert_eq!(
1762 framer.elide_cursor().last_forwarded_id,
1763 Some(2),
1764 "[{draft}] object 2 committed as forwarded"
1765 );
1766 }
1767 }
1768
1769 /// A bypass is reported once, as a zero-byte item, on the poll *after*
1770 /// the one that decided it.
1771 ///
1772 /// A draft-19 fetch stream is the cleanest case: the header decodes,
1773 /// so it is emitted, and only then is the stream found to be one whose
1774 /// objects are not addressed. `latch_bypass` cannot return the
1775 /// `Bypassed` item from that poll — the poll already owes a `Header` —
1776 /// so it is deferred.
1777 #[test]
1778 #[cfg(feature = "draft19")]
1779 fn a_bypass_is_reported_once_on_the_poll_after_the_item_that_decided_it() {
1780 let draft = DraftVersion::Draft19;
1781 // Fetch stream type 0x05, request ID 42, then four bytes of
1782 // whatever the publisher was sending.
1783 let stream = vec![0x05u8, 0x2A, 0xDE, 0xAD, 0xBE, 0xEF];
1784
1785 let counters = Arc::new(Recorder::new());
1786 let mut framer = ObjectFramer::with_recorder(
1787 DataStreamType::Fetch,
1788 draft,
1789 FramerConfig::default(),
1790 Arc::clone(&counters),
1791 );
1792 framer.feed(&stream);
1793
1794 let FramerOut::Header { raw, .. } = framer.poll() else {
1795 panic!("expected the fetch header");
1796 };
1797 assert_eq!(&raw[..], &stream[..2], "the header is still emitted in full");
1798 assert!(framer.is_bypassed(), "the bypass was decided before the header was handed out");
1799
1800 assert!(
1801 matches!(
1802 framer.poll(),
1803 FramerOut::Bypassed {
1804 reason: BypassReason::FetchGroupOrderUnknown,
1805 fixup_owed: false
1806 }
1807 ),
1808 "the bypass follows the header rather than replacing it"
1809 );
1810
1811 let FramerOut::Passthrough(rest) = framer.poll() else {
1812 panic!("expected the remainder to be forwarded uninterpreted");
1813 };
1814 assert_eq!(&rest[..], &stream[2..], "no byte is lost to the bypass");
1815
1816 assert!(matches!(framer.poll(), FramerOut::NeedMore));
1817 assert!(matches!(framer.poll(), FramerOut::NeedMore), "reported exactly once");
1818
1819 let c = counters.snapshot();
1820 assert_eq!(c.framers_created, 1);
1821 assert_eq!(c.framer_header_polls, 1);
1822 assert_eq!(c.framer_object_polls, 0, "a bypassed stream never enters the object path");
1823 assert_eq!(c.streams_not_shapeable, 1);
1824 assert_eq!(c.objects_not_addressable, 0, "no object was ever decoded");
1825 }
1826
1827 /// A bypass while a fix-up is owed says so, because the bytes still in
1828 /// flight decode to Object IDs one ahead of what was delivered.
1829 ///
1830 /// Objects 0 and 1 are written normally and object 1 is elided, which
1831 /// arms the fix-up. The third object is a hand-written all-ones varint —
1832 /// `2^62 - 1` in eight bytes under RFC 9000, `2^64 - 1` in nine from
1833 /// draft-17. Either is legal and its delta resolves to an Object ID above
1834 /// the ceiling, so `read_object_meta` returns `InvalidField` rather than
1835 /// an incomplete-input error and the framer abandons the stream with the
1836 /// fix-up unspent.
1837 #[test]
1838 fn a_bypass_that_strands_a_fix_up_reports_it_as_owed() {
1839 for &draft in DELTA_DRAFTS {
1840 let mut stream = build_stream(draft, &[object(0, 8), object(1, 8)]);
1841 // All-ones is the largest varint the draft's encoding can express:
1842 // eight bytes under RFC 9000, nine from draft-17.
1843 if draft.uses_moqt_varint() {
1844 stream.extend_from_slice(&[0xFFu8; 9]);
1845 } else {
1846 stream.extend_from_slice(&[0xFFu8; 8]);
1847 }
1848
1849 let counters = Arc::new(Recorder::new());
1850 let mut framer = framer_for(draft, FramerConfig::default(), &counters);
1851 framer.feed(&stream);
1852
1853 assert!(matches!(framer.poll(), FramerOut::Header { .. }));
1854 assert!(matches!(framer.poll(), FramerOut::Object { .. }), "[{draft}] object 0");
1855 let FramerOut::Object { meta, .. } = framer.poll() else {
1856 panic!("[{draft}] expected object 1");
1857 };
1858 framer.note_elided(&meta);
1859 assert!(framer.elide_cursor().fixup_pending, "[{draft}] the fix-up is armed");
1860
1861 assert!(
1862 matches!(framer.poll(), FramerOut::Error(_)),
1863 "[{draft}] the third object must fail to decode outright"
1864 );
1865 assert!(
1866 matches!(
1867 framer.poll(),
1868 FramerOut::Bypassed { reason: BypassReason::DecodeError, fixup_owed: true }
1869 ),
1870 "[{draft}] the stranded fix-up must be reported"
1871 );
1872
1873 assert_eq!(counters.snapshot().streams_not_shapeable, 1, "[{draft}] one latch");
1874 assert_eq!(
1875 counters.snapshot().object_ids_rewritten,
1876 0,
1877 "[{draft}] the fix-up never got the chance to be written"
1878 );
1879 }
1880 }
1881
1882 /// The counters move where — and only where — the slow path ran.
1883 ///
1884 /// This is the falsifiable half of the `Interest::NONE` claim: the
1885 /// framer is the slow path, so a framer that counts nothing makes the
1886 /// proof pass by measuring nothing.
1887 #[test]
1888 fn the_framer_counts_the_slow_path_work_it_did() {
1889 for &draft in DRAFTS {
1890 let delta = DELTA_DRAFTS.contains(&draft);
1891 let objects: Vec<_> = (0..4).map(|id| object(id, 16)).collect();
1892 let stream = build_stream(draft, &objects);
1893
1894 let run = run_eliding(draft, &stream, stream.len(), FramerConfig::default(), &[1]);
1895 let c = run.counters.snapshot();
1896
1897 assert_eq!(c.framers_created, 1, "[{draft}]");
1898 assert_eq!(c.framer_header_polls, 1, "[{draft}] one poll decoded the header");
1899 // Four objects plus the poll that returned `NeedMore`.
1900 assert_eq!(c.framer_object_polls, 5, "[{draft}]");
1901 assert_eq!(
1902 c.object_ids_rewritten,
1903 u64::from(delta),
1904 "[{draft}] one rewrite, and only where IDs are deltas"
1905 );
1906 assert_eq!(c.objects_not_addressable, 0, "[{draft}] every object was framed");
1907 assert_eq!(c.streams_not_shapeable, 0, "[{draft}] nothing was bypassed");
1908 }
1909 }
1910
1911 /// An object too big to buffer is counted as unaddressable exactly
1912 /// once, on both of the paths that produce one.
1913 ///
1914 /// The whole-stream feed routes object 65 through `poll_object`'s
1915 /// oversized branch; the 32-byte feed routes it through
1916 /// `poll_oversized`, which emits it as several `Passthrough` chunks
1917 /// and must still count *one* object.
1918 #[test]
1919 fn an_unaddressable_object_is_counted_once_per_object_not_per_chunk() {
1920 for &draft in DELTA_DRAFTS {
1921 let objects = vec![object(0, 8), object(1, 8), object(65, 4096), object(66, 8)];
1922 let stream = build_stream(draft, &objects);
1923 let config = FramerConfig::new().with_max_buffered_object_bytes(64);
1924
1925 for chunk in [stream.len(), 32] {
1926 let run = run_eliding(draft, &stream, chunk, config.clone(), &[1]);
1927 let c = run.counters.snapshot();
1928 assert_eq!(
1929 c.objects_not_addressable, 1,
1930 "[{draft}] chunk {chunk}: one object, however many chunks it left in"
1931 );
1932 assert_eq!(c.object_ids_rewritten, 1, "[{draft}] chunk {chunk}");
1933 assert_eq!(c.streams_not_shapeable, 0, "[{draft}] chunk {chunk}: no bypass");
1934 }
1935 }
1936 }
1937}