Each published crate (irtt-rs, irtt-client, irtt-server, irtt-stats,
irtt-proto) now points crates.io at its own README instead of
inheriting the root project README via the workspace-wide readme
default, which is removed. Each crate README is a short landing page:
what the crate is, who should use it, which sibling crate to use
instead if this isn't the right abstraction level, and links to
docs.rs and the main repository.
Add real Cargo example targets teaching irtt-client's blocking/Tokio/
managed API tiers, a basic and a multi-listener irtt-server example,
and an irtt-stats example contrasting finite vs continuous retention.
irtt-proto intentionally has none. Update each crate's top-level
rustdoc to explain its API-tier choices and point at the examples, and
enable docs.rs rustdoc example scraping for irtt-client and
irtt-server (via doc-scrape-examples on each example target, needed
because the crate has dev-dependencies).
The client examples are caller-paced against next_send_deadline()
rather than sending on every loop iteration, bound their receive wait
by the next send/timeout deadline instead of awaiting indefinitely,
poll timeouts on a short fixed cadence, and surface send/receive/close
errors instead of discarding them via unwrap_or_default()/if-let-Ok --
so they demonstrate correct pacing and terminate/report cleanly rather
than racing the server, hanging on a lost reply, or silently dropping a
failed final close.
Add proptest-based generative tests for irtt-stats's stateful rolling
windows and IPDV pairing.
The rolling_count and rolling_time bounds turn out to be two
independent VecDeques (rolling.rs), not a single window bounded by
both, so the main property test proves that independence: it drives a
StatsCollector configured with both bounds simultaneously through
generated operation streams, checking after every operation that each
window matches a from-scratch reference model applying only its own
bound (and that the cumulative snapshot matches a full-history
replay). Since both the production window and the reference replay
the same events through the same deterministic aggregation code, the
comparison is exact equality, not a tolerance.
Also add two smaller property tests for IpdvTracker's bounded FIFO
state and at-most-once adjacent-pair emission under arbitrary
(including non-monotonic and duplicate) sequence streams, complementing
the existing strictly-increasing hand test.
No discrepancies between production and the reference model were
found.
timeout_at is pre-send anchored (F-04), so it can precede the same
probe's post-send sent_at when probe_timeout is small relative to the
send-call duration. Rolling-window eviction assumes non-decreasing
event timestamps, so clamp the internal StatsEvent::Loss windowing
value without changing the public EchoLoss.timeout_at field. Also
correct an AGENTS.md claim that sent_at feeds the downstream one-way
delay, which it does not.
Rolling snapshots are recomputed with `SampleMode::RunningOnly` regardless
of the configured sample mode, so the new `SampleMode::Exact` wording — that
every timing metric a `Snapshot` exposes reports a median — was not true of
`rolling_count()` and `rolling_time()`.
Scope the promise to the cumulative snapshot and say plainly on both rolling
accessors that they use running statistics only and report no median, since
a rolling window is a bounded recent view rather than the run's retained
history. A test pins both halves so the wording cannot quietly go stale.
Rolling retention behavior itself is unchanged.
`SampleMode::Exact` retained exact samples for eight of the eleven public
timing metrics. Send call, timer error, and server processing always built
their `TimeMetric` with retention disabled, so a finite run reported no
median for them even though they are the same public `TimeStats` type and
appear in the same summary table with a Median column.
Derive one `retain_exact` decision from the sample mode in `CoreStats::new`
and apply it to every timing metric, so exact mode means the same thing for
all of them. Running-only mode is unchanged and still retains none of them.
The retention estimator counted eight exact sample streams per measurable
reply. Rename it to the eleven exact samples an ordinary successful probe
can retain and document the wider optional-measurement upper bound, which
raises the finite estimate on this target from 668 to 764 bytes per probe.
A test counts the metrics a fully measured probe actually retains and
compares it against that constant, so retention and its estimate cannot
drift apart unnoticed.
No new median-capability type was needed: `TimeStats::median_ns` already
distinguishes a retained median from an absent one, and an optional
measurement the server never supplied still produces no samples at all.
The TUI suppressed late-reply timing by rebuilding the event with sent_at,
rtt, server_timing, one_way and received_stats erased. That leaked stats
policy into event construction and, as a side effect, reclassified a matched
late reply as an untracked late reply.
StatsConfig now carries an explicit LateReplyMode. Measure keeps today's
behavior for every existing caller. CountOnly keeps the reply a late unique
reply and keeps packet, byte and late counters updating, while suppressing
timing, one-way, server-processing and IPDV contributions and reporting
contributed_sample == false. Cumulative and rolling snapshots share the
policy. Genuinely untracked late replies stay untracked in both modes.
The TUI selects CountOnly and passes the original event straight to the
collector.
An audit of the server-reported receive window found no defect in the
statistics pipeline: the bitmap is decoded raw, stored raw, never
interpreted, and never feeds the directional loss estimates, which use
the cumulative server receive count.
The public rustdoc was silent on all of that, leaving consumers of
PacketCounts free to read a window of 0x1 as a 63-packet loss burst.
Document the field's contract, name the count as the only server input
to LossStats, and cover the window-only case with a regression test.
Allow one-way delay samples to represent negative values caused by wall-clock skew between client and server. Update CLI output and statistics collection to correctly handle and report these signed durations.
Use wrapping arithmetic instead of checked arithmetic to ensure inter-packet delay variation (IPDV) is correctly calculated when sequence numbers wrap from u32::MAX to 0.
Hide internal fields to improve type safety and provide helper methods for unit conversions and negative value checks. Fix several feature-gated test compilation issues.
Simplify RttSample by merging signed and unsigned fields into a single SignedDuration for adjusted and effective RTT. This ensures accurate reporting and statistics when server processing time exceeds raw RTT.
Simplify the statistics model by merging signed and unsigned duration types into a single structure that uses i128 to consistently handle both positive and negative time offsets.
It was an attempt at handling integer wraparound more cleanly keeping in mind that we have a continuous mode,
and so we can more easily reach seq > u32::MAX
However after running the numbers, if we run with a interval of 10ms, it's still gonna take ~497 days before we hit integer overflow, so for now it doesn't seem worth to have this extra logic to handle something that'll in 99.9% of cases be purely an edge case
Keeping in mind autorate usage, in sqm-autorate-rust each send loop is 500 ms by default, and the sleep is then 500 ms / 5 reflectors, so our interval is 100 ms. that means it'll take 4970 days or ~13.6 years to reach the limit. Not worth dealing with IMO
Expose multiple jitter updates when a single event fills a sequence gap, ensuring accurate IPDV reporting for out-of-order packets in human-readable output.
Nils Andreas Svee