ABR overhaul Phase 1 (planning PR #20, RFC §2.1–2.4 — signed off): the motion-onset field case's transmission-plane half. Static content used to strip both defenses — the microburst allowance was gigabit-sized so Wi-Fi frames never paced at all, and adaptive FEC decayed to 1 % — and the first motion frame died unprotected.
Pacing (§2.1–2.3):
The unpaced allowance becomes time at the pace rate — clamp(pace_rate × 10 ms, 16 KiB, 256 KiB) (send_pacing::auto_burst_bytes, pure + unit-tested). One constant lines up both proven ends: ~19 KiB at the field case's 15 Mbps pace (its discriminator was PACE_BURST_KB=16) and ~112 KiB at 90 Mbps LAN pace ≈ the old 128 KiB floor, so LAN latency does not regress. PUNKTFUNK_PACE_FACTOR=0 keeps the legacy fraction-of-frame rule; PUNKTFUNK_PACE_BURST_KB still pins.
A streamed AU consumes one allowance across its block flushes (StreamedOpen.burst_left) instead of a fresh 128 KiB per block — the latent multiplication the handoff flagged.
One frame's paced spread is bounded to ~2 frame intervals (live refresh off stats.mode) so a paced IDR can't back the encode|send sync_channel(3) up into cadence_degraded, which refuses every climb. MAX_PACE_SPREAD stays the absolute ceiling — a 5 fps virtual mode re-licenses nothing.
FEC (§2.4):
The adaptive decay gains a burned floor (FecFloor, pure + unit-tested): any window reporting real shard loss raises the floor to 5 %; ~2 clean minutes re-earn 1 %; a re-burn inside the step-down's probation doubles the requirement (bounded ~16 min); a step-down that survives resets it. No session-permanent latch — the encode stand-down's house rule. PUNKTFUNK_FEC_PCT pin semantics and the GameStream plane are untouched.
Docs: the FEC_PCT row now mentions the burned band.
Verified:cargo fmt --check clean; the pure logic (FecFloor, auto_burst_bytes, native_budget old+new assertions) proven green natively via the verbatim-copy harness. The workspace compile + tests lean on ci / rust — local Docker was occupied by a parallel job's build (the OOM-each-other trap). Field validation (the A/B static→motion rig) is tracked in the RFC's Phase 0 opens.
ABR overhaul **Phase 1** (planning PR #20, RFC §2.1–2.4 — signed off): the motion-onset field case's transmission-plane half. Static content used to strip both defenses — the microburst allowance was gigabit-sized so Wi-Fi frames never paced at all, and adaptive FEC decayed to 1 % — and the first motion frame died unprotected.
**Pacing (§2.1–2.3):**
- The unpaced allowance becomes **time at the pace rate** — `clamp(pace_rate × 10 ms, 16 KiB, 256 KiB)` (`send_pacing::auto_burst_bytes`, pure + unit-tested). One constant lines up both proven ends: ~19 KiB at the field case's 15 Mbps pace (its discriminator was `PACE_BURST_KB=16`) and ~112 KiB at 90 Mbps LAN pace ≈ the old 128 KiB floor, so LAN latency does not regress. `PUNKTFUNK_PACE_FACTOR=0` keeps the legacy fraction-of-frame rule; `PUNKTFUNK_PACE_BURST_KB` still pins.
- A **streamed AU consumes one allowance across its block flushes** (`StreamedOpen.burst_left`) instead of a fresh 128 KiB per block — the latent multiplication the handoff flagged.
- One frame's paced spread is **bounded to ~2 frame intervals** (live refresh off `stats.mode`) so a paced IDR can't back the encode|send `sync_channel(3)` up into `cadence_degraded`, which refuses every climb. `MAX_PACE_SPREAD` stays the absolute ceiling — a 5 fps virtual mode re-licenses nothing.
**FEC (§2.4):**
- The adaptive decay gains a **burned floor** (`FecFloor`, pure + unit-tested): any window reporting real shard loss raises the floor to 5 %; ~2 clean minutes re-earn 1 %; a re-burn inside the step-down's probation doubles the requirement (bounded ~16 min); a step-down that survives resets it. No session-permanent latch — the encode stand-down's house rule. `PUNKTFUNK_FEC_PCT` pin semantics and the GameStream plane are untouched.
Docs: the `FEC_PCT` row now mentions the burned band.
**Verified:** `cargo fmt --check` clean; the pure logic (FecFloor, auto_burst_bytes, native_budget old+new assertions) proven green natively via the verbatim-copy harness. The workspace compile + tests lean on `ci / rust` — local Docker was occupied by a parallel job's build (the OOM-each-other trap). Field validation (the A/B static→motion rig) is tracked in the RFC's Phase 0 opens.
Static content used to strip every defense before motion: the unpaced
microburst allowance was an absolute max(128 KiB, wire/4) — sized for
gigabit LAN, so at Wi-Fi bitrates every frame went out back-to-back —
and adaptive FEC decayed to 1 %, so the first big motion frame died
unprotected (the 2026-08-26 field case; RFC design/abr-stack-overhaul.md
in punktfunk-planning, §2.1–2.4).
- The allowance is now TIME at the pace rate: 10 ms, clamped to
[16 KiB, 256 KiB] (send_pacing::auto_burst_bytes). One constant lines
up both proven ends — ~19 KiB at the field case's 15 Mbps pace (its
discriminator was PACE_BURST_KB=16) and ~112 KiB at 90 Mbps LAN pace,
so LAN latency does not regress. PACE_FACTOR=0 keeps the legacy rule.
- A streamed AU consumes ONE allowance across its block flushes instead
of a fresh one per block (the latent per-block multiplication).
- One frame's paced spread is bounded to ~2 frame intervals so an IDR
can't back the encode|send sync_channel(3) up into cadence_degraded,
which refuses every climb; MAX_PACE_SPREAD stays the absolute bound.
- Adaptive FEC gains a burned floor: any window with real shard loss
raises the decay floor to 5 %; ~2 clean minutes re-earn 1 %, an early
re-burn doubles the requirement (bounded), a durable step-down resets
it — no session-permanent latch, per the encode stand-down's house
rule. PUNKTFUNK_FEC_PCT pin semantics and the GameStream plane are
untouched.
Pure logic (FecFloor, auto_burst_bytes, native_budget) is unit-tested;
the burned-floor band is noted in the FEC_PCT docs.
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
ABR overhaul Phase 1 (planning PR #20, RFC §2.1–2.4 — signed off): the motion-onset field case's transmission-plane half. Static content used to strip both defenses — the microburst allowance was gigabit-sized so Wi-Fi frames never paced at all, and adaptive FEC decayed to 1 % — and the first motion frame died unprotected.
Pacing (§2.1–2.3):
clamp(pace_rate × 10 ms, 16 KiB, 256 KiB)(send_pacing::auto_burst_bytes, pure + unit-tested). One constant lines up both proven ends: ~19 KiB at the field case's 15 Mbps pace (its discriminator wasPACE_BURST_KB=16) and ~112 KiB at 90 Mbps LAN pace ≈ the old 128 KiB floor, so LAN latency does not regress.PUNKTFUNK_PACE_FACTOR=0keeps the legacy fraction-of-frame rule;PUNKTFUNK_PACE_BURST_KBstill pins.StreamedOpen.burst_left) instead of a fresh 128 KiB per block — the latent multiplication the handoff flagged.stats.mode) so a paced IDR can't back the encode|sendsync_channel(3)up intocadence_degraded, which refuses every climb.MAX_PACE_SPREADstays the absolute ceiling — a 5 fps virtual mode re-licenses nothing.FEC (§2.4):
FecFloor, pure + unit-tested): any window reporting real shard loss raises the floor to 5 %; ~2 clean minutes re-earn 1 %; a re-burn inside the step-down's probation doubles the requirement (bounded ~16 min); a step-down that survives resets it. No session-permanent latch — the encode stand-down's house rule.PUNKTFUNK_FEC_PCTpin semantics and the GameStream plane are untouched.Docs: the
FEC_PCTrow now mentions the burned band.Verified:
cargo fmt --checkclean; the pure logic (FecFloor, auto_burst_bytes, native_budget old+new assertions) proven green natively via the verbatim-copy harness. The workspace compile + tests lean onci / rust— local Docker was occupied by a parallel job's build (the OOM-each-other trap). Field validation (the A/B static→motion rig) is tracked in the RFC's Phase 0 opens.