PLN typed a TODO comment into the middle of `d12 $ gF3 $ gM3`, landing the `3`
after the `--`. The line became `d12 $ gF3 $ gM`, and `gM` does not exist:

    Variable not in scope: gM :: Pattern ValueMap -> Pattern ValueMap
    Perhaps use one of these: 'gM1', 'gM2', 'gM3'

One parse error silences the WHOLE do-block, so this was not 'd12 is quiet' —
it was 'do_it_right makes no sound at all', found by silent-eval --seeded while
he was mid-run. Comment moved above the line, where it cannot swallow code.

The TODO itself is preserved verbatim: he wants more of ^18/^19 on this lead.

J-6, the one-before-last KRK rehearsal. Per-track watch list for the 1h30 run,
because ten of the fifteen tracks have something in them PLN has never heard:
seven new d10 risers, the take_5_drops d5 structure fix, ouais_je_funk's d10
finally mutable, lime_tacos' cymbals relocated to d12, and two tracks that were
not in any tool's idea of the set until yesterday.

Also carries the fresh ghost matrix (9 of 14 transitions, 18 orphans) recomputed
on the real 15-track order — and the pleasant surprise that adding the two
missing tracks produced the longest clean stretch in the set,
piment -> ouais -> lime -> perfect.

PLN: "for lime if d11 taken lets push d6 to free d12 :)" — d11 is held by
"Menace de retour", and d10 is now the safe riser, so d12 was the free slot.

The move is genuinely one token: d6 appears exactly once in the file, carries no
`cut 6` and no orbit-bound crushbus, so nothing had to follow it. ^44 stays
shared with d3 DELIBERATELY — that is the idiom PLN ruled correct: one press
re-rhythms hats and cymbals together.

Side effect worth knowing at the desk: PV008 only inspects orbits 1-8, so with
the cymbals on d12 the shared-button warning stops firing entirely. The link is
unchanged; only the lint's field of view moved. Do not read the green as the
coupling having gone away.

WATCH ON THE RUN: Ardour Tidal 12 sits at +6.0 dB and Tidal 06 at -3.8 dB, so
the cymbals land ~9.8 dB hotter than before. Ride Tidal 12 down rather than
re-gaining the pattern until PLN has heard it.

Verified: silent-eval --seeded ok (every declared orbit emits), pvlint clean.

PLN, on lime_tacos' ^44 driving both d3 (hats) and d6 (cymbales sol):

    "lime tacos its ok when two percs are linked, e.g. here, change rythm of
     both 3 and 6. [...] im fine with this, not a NOGO, just a WARN?"

He is right, and the interesting part is WHY the rule was wrong rather than that
it was. PV008 was earned on a real bug — the 2026-07-29 phase-2 remap left d1
holding ^42 after d2 was given ^42, and perfect.tidal shipped one button firing
two orbits. That bug is real and this rule is how it was found.

But the accidental case and the deliberate one are TEXTUALLY IDENTICAL. Two
orbits, one button CC, no way to tell from the source whether PLN meant it. The
rule had quietly encoded "one button drives one orbit" as an invariant of his
style, when it is only a hypothesis about it — and the hypothesis was wrong. He
groups percussion on purpose: one gesture re-rhythms the whole group.

A check that cannot distinguish intent must not hold the gig gate shut. It
reports; PLN decides. So: severity error -> warning, and the detail now states
both readings and what distinguishes them (deliberate percussion grouping vs a
control inherited from a remap) instead of assuming the second.

Effect: tools/gig-up.sh goes NO-GO -> GO. This was the last blocking failure, and
it was blocking on a correctly-written track.

Added test_pv008_is_a_warning_not_an_error, which quotes PLN's ruling, so nobody
"tightens" this back to error without meeting the argument first. pvlint 63/63.

Third time. tools/gig-up.sh's 'tools executable' check (added yesterday for exactly
this) is why it surfaced: setlist.py went into 04fc8c7c with content but mode 100644,
so a fresh clone would get 'permission denied' from a HARD gate check. chmod +x
touches the working tree; git stores the bit separately and needs
`git update-index --chmod=+x`, and even then a pathspec-limited commit can leave
HEAD at 100644 while the index says 100755 — which is what happened here.

Verify with `git ls-tree HEAD <file>`, never `ls -l`.

PLN: "can you add a safe riser on all set tracks that dont have one, and consider
moving the d10 to d9 (eg a synmenace iirc) to have d10 riser always?"

THE CONVENTION. d10 = riser, so the gesture lives on the same column in every track
and his hands stop having to remember per-track. It was already true in three tracks
without anyone naming it — do_it_right (mask "<f!4 t f!3>"), take_5_drops (the
"A FAIR RISER" comment on d10) and vague_de_crime, whose line #33 designates as THE
safe riser because a 1-in-8 mask cannot run away:

    d10 $ gF1 $ gM3 $ mask "<t f!7>" $ "risers:8"

Copied verbatim into the 7 tracks where d10 was free: bombe_dj, wap, lime_tacos,
you_my_sunshine, mafia_sans_serif, desire, electric_hammer. Inserted contiguous with
the existing d-blocks, because all three pre-existing risers live inside the main
block rather than standalone — a riser in its own block would need a separate
ctrl+enter, which is not the gesture. 10 of 15 tracks now have it.

WHY THE RELOCATION IS NOT IN THIS COMMIT. "Consider" did real work here. The two
tracks PLN was remembering are piment_bresilien (d10 = Menace, d9 free) and
ouais_je_funk (d10 = SynMenace, only d12 free) — but moving an orbit is not a
one-line move, because orbit N is simultaneously the LCXL column, the Ardour track
and the cut/crushbus group:

    piment's d10 block carries  -- A6 = d10's effect  on its crushbus CC ^18
    d9's fx CC is ^17, d12's is ^20      -> the CC must move with the sound
    Ardour Tidal 10 = -2.2 dB, Tidal 09 = +6.0 dB, Tidal 12 = +6.0 dB
                                         -> the synth lands +8.2 dB louder

And the argument that settles it: perfect, gimme_acid and the_revolution declare all
12 orbits, so d10 is never free there and the convention caps at 12/15 no matter what.
Relocating buys 2 tracks for two CC remaps, a cut-group change and two synths jumping
+8.2 dB — on material PLN has not rehearsed since, 7 days out. Deferred by choice, not
forgotten; his call to overrule.

#99, d5 of take_5_drops: `>|` takes structure from the RIGHT, so `slow 4` note
patterns were REPLACING the slice rhythm with one event per four cycles — the
"steals structure" symptom, confirmed. Now `|>`, which keeps the left structure and
takes only the note value. Note line 35 immediately above was already doing it that
way, so this makes the block self-consistent rather than inventing an idiom. Marked
TODO @pln verify as asked.

Also ouais_je_funk d10 read `gF3 $ gF3` — the same helper twice, so the orbit had no
family mute and NOTHING on the surface could drop it (pvlint PV013). Now `gF3 $ gM3`.
Also TODO @pln verify.

VALIDATION: all 15 setlist tracks still compile (gig-up "setlist compiles" green,
silent-eval cold-queries every dN), preload still covers the set at 53/53 banks, and
pvlint added zero new findings — still the one pre-existing error, lime_tacos' ^44
driving both d3 and d6. That one is left alone deliberately: ^44 is neither orbit's
home gate (d3 = ^43, d6 = ^58/^90), so fixing it means rebinding a button PLN's hands
already know, which is a taste call and not lint's to make.

PLN's ruling yesterday: "set SSOT is the backlog entry, always. if in the backlog
i add a track or reorder them, its the truth." Then today, the part that decided
the design: "if i added new ones, id not need to tell you to update, it should
flow downstream, ditto if remove one."

So this is not a resolution, it is a pipe.

WHAT WAS ACTUALLY WRONG. Applying the ruling for the first time found the backlog
listing 15 tracks while every tool in the repo believed 13:

    missing everywhere:  ouais_je_funk (120)   lime_tacos (110)
    wrong order:         backlog has you_my_sunshine (166) BEFORE mafia_sans_serif

ouais_je_funk was not a stale wishlist entry — PLN committed a kick rework to it
in 23f60e76, yesterday. And the drift had already cost him something audible:

    preload warmed   50 banks   (13 tracks)
    the set needs    53 banks   (15 tracks)
    would cold-load  dr  gfunk_bass  h2ogmcy

Three banks read from disk on first play. That is the same mechanism PLN heard as
crackle on the take_5_drops transition (#120) — a bug we thought was fixed, still
live, just smaller. check-preload.sh reported "ok" the whole time, because it asked
set-coherence, which was one of the wrong answers. The freshness check had
inherited the staleness it was built to catch.

ROOT CAUSE: eight tools each hardcoding `SETLIST = armada/setlist_opal2026.txt`,
and one human keeping that file in sync by hand.

THE SHAPE, and why it is not "make all eight import a module". The .txt stays
exactly where it is, same format, so all eight consumers are untouched — this is
J-7 to a gig and a refactor that breaks a tool costs more than it buys. What
changes is its STATUS: generated, never hand-edited, with a HARD gig-up check
that it still matches the backlog. One parser, one artifact, one answer at a time,
and a gate that fails loudly the moment they drift.

  tools/setlist.py         the parser. Resolves PLN's freeform codenames
                           ("Ceci n'est pas Une Bombe" -> bombe_dj, "Le shifteur
                           marteau ? to finish!" -> electric_hammer) via aliases +
                           token containment, and raises rather than guessing.
  gig-up --converge        regenerates the setlist BEFORE preload. That order is
                           load-bearing: preload reads the setlist, so doing it the
                           other way warms the old set.
  gig-up "setlist vs backlog"   HARD, and placed before every check that reads the
                           setlist — a stale list makes "setlist compiles",
                           "preload covers set" and "transition ghosts" all answer
                           about a set PLN is not playing.

WHAT THE END-OF-SECTION DETECTION TAUGHT ME. First attempt whitelisted the
headings that count as part of the set ({FINALE, Encore, ...}) and died on line
one against "## Livecoding Techno DNB Nu-jazz". PLN's headings are prose; any
whitelist is a list of the ones he happened to write that day. The working rule
ignores heading semantics entirely: run to the next top-level heading, take the
list items. Non-track content does not survive that filter — "> Sunset Forest",
"1h30 - 20h30-22h", "??TRANSITION???" and the trailing url: block are none of
them list items.

VALIDATION, and specifically not "it returns the 15 tracks currently in the
backlog" — a hardcoded list passes that too. tools/tests/test_setlist.py MUTATES
a synthetic backlog and asserts the output moved: add a track (appears, in PLN's
position, not appended), remove one (disappears), swap two lines (output swaps),
unresolvable name / duplicate / missing section (all raise). 18/18.
Then the gate itself, mutated: dropped a line from the .txt -> caught; swapped
two lines, same membership -> caught as "same tracks, DIFFERENT ORDER".

AND IT PAID OFF IMMEDIATELY. With the real 15 in the setlist, pvlint went red on
two tracks that had been invisible to every tool until today: an unmutable orbit
in ouais_je_funk (gF with no gM3 — nothing on the surface can drop it) and button
^44 driving both d3 and d6 in lime_tacos (one press, two orbits). Neither is new
breakage; both were simply never looked at. Left for PLN — they are musical calls.

The 53 banks are warm on the rig now: 54/54 banks OK in 7.0 s.

Reworking d4's texture: the lpf range sweep comes out (resonance stays), crushbus
goes 16->3.5 on ^52, and room 0.4 / sz 0.5 / dry 1.1 opens the tail.

Three closed tasks from the 2026-08-01 session, written for a cold reader. The
through-line worth mining later: the rig SENSES everything and ACTUATES nothing.
scsynth died and sclang logged it, the Bridge warned 63 times, gig-up saw it and
printed GO — and separately, PLN's ears found a preload bug that four green tools
had missed. Plus the session's sharpest lesson: two of my own fixes (a watchdog, a
chmod) were themselves broken in exactly the way they were meant to catch.

PLN's own rehearsal edits, 2026-08-01, saved before a break. Committed as their own
human commit so nothing is lost — IN PROGRESS, not a finished arrangement.

* ouais_je_funk — the kick becomes gate-driven instead of static:
  "[jazz,kick:5]*4" -> midiOn "^41" / midiOff "^41" variants over a layered
  "[techno:0,808bd:2,909,kick:4]" with gain 1.5 # lpf 400. Old "kick:5" kept
  commented rather than deleted.
  Also: the lead drops an octave — note "<b b b g>@7 <cs6!3 fs6>" is now
  ("..." - 12); the superimpose gate moves ^18 -> ^20; crushbus 101 added on ^18;
  and both d11 arpeggio blocks (pinkyup/thumbupdown, diverge/disconverge) are
  commented out rather than removed.
* desire — crushbus 41 (range 16 2.5 "^52").

Left as PLN saved it. He resumes rehearsing Saturday afternoon.

--converge is the CLI big red button: bring the rig to ready, then let the existing
checks prove it. Scope is SERVICES, not apps — it starts systemd units it owns and
REPORTS Pulsar/Ardour with the command instead of launching them, because a script
that opens windows under someone is a script they stop trusting (and Ardour has a
session + crash-recovery dialog, #20). Launching apps belongs to the Bridge (#116).

Safe at any moment: every action is conditional on the thing being broken, so a
healthy rig converges to a no-op — verified, scsynth PID unchanged. It handles the
exact 2026-08-01 failure explicitly: unit ACTIVE but audio server DEAD needs a
RESTART, since `start` is a no-op on an active unit. Distinguishing those two is the
whole lesson. Preload is fixed BEFORE SuperDirt boots, or the warm would not happen
until the next restart. And it waits for scsynth to appear rather than guessing —
guessing is what made the first watchdog flap.

=== THE THIRD TIME, AND THIS TIME IT WAS US ===

gig-up's first run (1219c704) found check-boot.sh and check-tracks.sh at mode 644 —
never runnable as `tools/check-tracks.sh`. That fix turned out to be FAKE: the chmod
only ever touched the working tree. Git has recorded 100644 for both from that day to
this, so every fresh clone still got a broken pre-gig gate.

And gig-up.sh had the same bug, from birth. Committed 100644 by 1219c704 and again by
6ab09eaa. It ran only because the local working tree happened to carry the bit; a
`git checkout master` materialised it at 644 and it died with "permission denied".

That failure was nearly invisible, which is the real story: the caller filtered its
output through sed, so a run that never executed LOOKED like a clean no-op, and I
reported "no-op confirmed" about a script that had not run. A false green produced by
the absence of a program.

So the gate now checks its own hands — "tools executable", HARD, verifying BOTH the
on-disk bit and what git records. It found four real problems on its first run:
    disk=x  git=100644  check-boot.sh          <- day-one "fix" never committed
    disk=x  git=100644  check-tracks.sh        <- same
    disk=-  git=100644  sc-watchdog.sh         <- not executable AT ALL
    disk=-  git=100644  tests/test-sc-watchdog.sh
The third is the one that matters: parvagues-sc-watchdog.service had been ENABLED AT
BOOT with an ExecStart it could not execute. It was stuck `activating`, restart-
looping silently. The supervisor I added to catch silent failure was itself failing
silently, and nothing would have said so until an audio server died at a venue and
did not come back. All six now 100755 on disk and in the index; watchdog restarted
and confirmed logging.

Same family as the day-one finding: a thing verified once, in a context that no
longer holds. The new twist is that "chmod +x" is not a fix — recording it in git is.

Also honest-ing the recovery toast (#122). PLN reported needing a ctrl+enter in
Pulsar to resume after a recovery. Likely ICMP port-unreachable poisoning Tidal's
connected UDP socket (ECONNREFUSED on next send, stream stops until re-evaluated).
Until #122 tests it, the toast says "if the music does not resume on its own,
ctrl+enter in Pulsar" rather than claiming a full recovery — a rig that is alive and
silent is indistinguishable on stage from the original fault.

Validation: 16/16 green on the live rig (9 cold + 7 live). Converge no-op verified on
a healthy rig; converge ACTION verified by stopping the watchdog and watching it come
back, with scsynth untouched throughout.

gig-up proved the set COMPILES and said nothing about whether anything could make
a sound. scsynth was read only to colour a banner (:106,:115) and to skip --audio
(:202) — it never failed. This morning the audio server had been dead for two days
and the gate would still have printed GO.

--live is a MODE, not the default. Cold-by-default remains right: cold checks are
safe mid-set and must work on a train. Severity is chosen per link, deliberately:

  HARD  scsynth alive / sclang :57120 / SC -> Ardour (>=24 pw links)
        no audio server, no listener, or no path to the DAW = there is no gig.
  SOFT  Tidal :6010, LCXL -> SC, SuperDirt warmed, perf mode
        all legitimately absent ten minutes before doors; a gate that reddens for
        them is a gate PLN stops reading.

Every check reads the thing that MAKES SOUND — a process, a socket, a real pw-link —
never `systemctl is-active`. That distinction is the whole reason this exists: the
unit reported active/running for two days while the rig was mute, because MainPID is
sclang and scsynth is its child.

ALSO, and it came from PLN's ears: "preload covers set", a HARD cold check.

He heard crackles moving do_it_right -> take_5_drops. Measured in that window: 0
SuperDirt lates, 0 pipewire xruns, 17 lazy soundfile reads — all take5:*. Disk I/O,
not CPU. The preload MECHANISM was healthy and said so ("47/47 banks OK in 1.9 s");
it was warming the wrong list. preload.scd was dated Jul 28, header "10 track(s)
scanned", while the set is 13 — because setlist_samples.py reads setlist_opal2026.txt
(10 tracks) and gig-up/cheat-sheet trust set-coherence.setlist_tracks() (13). The
missing three are do_it_right, take_5_drops, electric_hammer: exactly the pair he was
transitioning between. #120 tracks which source becomes canonical — not guessed here.

Two properties made it invisible, and they are the transferable lesson: preload.scd
is GITIGNORED (correctly — generated), so no diff ever showed it drifting, and
NOTHING COMPARED IT TO THE SET. A generated artifact with no freshness check is a
stale artifact waiting for the worst possible moment; its natural debut was a cold
boot at the venue, on the first play of every track added since Jul 28. It only
surfaced at a desk because SuperDirt had been up since Jul 28 with those banks cached
from a first play weeks earlier, and a restart cleared the cache.

tools/check-preload.sh compares BANK SETS (not bytes — the generator stamps a header
and orders banks, so a byte diff would cry wolf), borrows set-coherence's parser
rather than adding a fourth definition of "the setlist", and has --fix.

Validation, all of it adversarial:
 * check-preload mutation-tested — restore the Jul-28 plan and it names the 7 banks
   that would crackle: daft fbreak120 hammer popkick realclaps superfreak take5.
   hammer and superfreak are independently confirmed: they appear as lazy loads in
   this morning's journal, before any of this was written.
 * --live on the live rig: 7/7 green after fixing MY OWN check (below).
 * THE GATE BITES: run under a PATH shim whose pgrep reports scsynth absent ->
   NO-GO, exit 1, correct fix hint, real rig never touched. That is precisely the
   state that printed GO this morning.
 * cold run unchanged: still GO, live chain not executed.

One check was wrong before it was right, in the session's recurring shape. "SuperDirt
warmed" first warned about a rig that was perfectly warm, because
`journalctl --user -u parvagues-sc` returns lines from a PREVIOUS instance on this
box — it handed back Jul 30 output while the current instance had already logged
"51/51 banks OK". Now filtered by SyslogIdentifier and keyed to the unit's own
ActiveEnterTimestamp, so it asks "did THIS boot warm?" rather than "was there ever a
PRELOAD line", which any previous boot would satisfy forever.

PLN's own live-session edits, made at the desk on 2026-08-01 while rehearsing the
OPAL set. Committed separately from the session's tooling work so the musical
changes stay legible on their own.

* wap — uncomment octersubsubbus 42 / octerbus 41, both on range 0 1.4 "^53".
  The sub layer is live again. (This is the `octersub` want noted in the
  2026-07-31 set-coherence report.)
* bombe_dj — "[808bd:1,808bd:11]" -> "[808bd:11]". One kick sample instead of two
  stacked, on the 1/4/6/9/11 positions.
* do_it_right — mark d3's gM2/gF1 line with a TODO about buttons and variations.

Minutes after test-sc-watchdog exited, PLN got a STICKY desktop notification:

    "3 restarts in 10 min. Giving up — this needs your eyes.
     journalctl --user -u scwd-test-1269280.service -n 50"

scwd-test-1269280.service was the test's own FAKE transient unit. By the time the
notification was read, it had been torn down — so the suggested command returns
nothing, about a failure that never happened, while the real rig was playing fine.

Two defects in one message. The test could reach a real notification daemon at all;
and the give-up path is deliberately urgency=critical/sticky, which is correct for a
real outage and actively harmful for a fake one — it cannot be dismissed by waiting.

Fix: SCWD_NO_NOTIFY=1, honoured by notify() and set by the harness. A test must never
be able to page a human about something that is not real.

Re-run green: 7 passed, 0 failed, and silent. The log still shows the rate limiter
walking (0 prior) -> (1 prior) -> (2 prior) -> GIVING UP, so suppressing the toast
did not suppress the evidence — the assertion reads the log, not the notification.

Same family as the flap it was written to catch: a supervisory tool whose side
effects escape the scope it was reasoning about.

The rig went silent overnight and NOTHING recovered it. Seven layers watched it
happen. From the journal:

    10:37:36  PM: suspend entry (s2idle)
    10:48:38  systemd-coredump: scsynth terminated abnormally
    10:48:38  sclang: "Server 'localhost' exited with exit code 0."
    11:18:47  Bridge: "scsynth process not found"   x63, to the journal only

scsynth is a CHILD of sclang and systemd's MainPID is sclang, so the unit reported
`active (running)` for two days while there was no audio server at all. A green unit
is not sound.

Two findings behind it:

* `parvagues-sc.service` was DISABLED. Six user units start with the session —
  gig-log, bridge, perf-tray, tidal-ardour-autoroute, lcxl-leds-watch,
  midi-autoconnect — and the only one that produces audio was not among them. It had
  been alive since Jul 28 purely as a manual leftover, which is why its death was
  permanent. Now enabled, and symlinked to the repo like the others so it cannot
  drift from the tracked copy.
* The d8346672 fix (QT_QPA_PLATFORM=offscreen) worked exactly as designed: sclang
  survived the suspend perfectly. Suspend killed scsynth instead. The mitigation had
  been applied one layer too shallow — and because sclang stayed up, the unit never
  failed, so no Restart= setting of any kind would have fired.

So: Restart=on-failure (bounded by StartLimitBurst=3/5min, honouring the unit's own
"audio gear should never flap" comment) for the sclang case, plus an external
watchdog for the scsynth case. Deliberately outside the boot path — a bug in
start_and_midi.scd is gig-fatal, a bug in a watchdog is merely annoying. Folding
this into an sclang ServerQuit handler is the right long-term shape; not seven days
before a gig.

THE FIRST VERSION FLAPPED ON PLN'S DESK, and the bug is the interesting part.
The unit had `PartOf=parvagues-sc.service`, so that stopping SuperDirt would stop
the watchdog. But PartOf propagates RESTARTS. Three lines, 90ms apart:

    11:35:46.265  sc-watchdog[1102323]: scsynth GONE — restarting
    11:35:46.321  systemd: parvagues-sc.service: Consumed 1min CPU time
    11:35:46.354  sc-watchdog[1123545]: watching scsynth        <- NEW PID

The watchdog killed itself issuing the restart, Restart=always revived it, and its
in-memory rate-limit counter reset to zero. MAX_RESTARTS was unreachable BY
CONSTRUCTION: the state enforcing the limit was destroyed by the action it was
limiting. It would not have stopped at four notifications; it would have gone
forever. A supervisor must outlive its supervisee.

Fixed three ways, defence in depth:
  - no PartOf on the watchdog unit (the actual fix)
  - restart log persisted to $XDG_RUNTIME_DIR, so being killed cannot launder it
  - await the server's return (up to 100s > TimeoutStopSec=90s) instead of sleeping
    a guessed 25s and judging a boot that was still in flight

Also: every notification was urgency=critical, which Plasma treats as never-expire,
so recovery buried the screen in sticky popups. Stickiness is now the signal —
normal+timeout for "handled itself", critical+sticky only for "this needs your
hands".

Validation. tools/tests/test-sc-watchdog.sh runs the real script against a FAKE unit
holding a FAKE child server, reproducing the green-unit-no-sound topology exactly.
7/7 pass in ~90s and scsynth's PID is unchanged across the run — it touches no real
audio. It pins the flap directly: "watchdog survived issuing the restart", and the
log shows (0 prior) -> (1 prior) -> (2 prior) -> GIVING UP. Plus a 25s --dry-run
against the live unit: zero false positives on a healthy rig.

That testability is the real lesson. v1 could only be exercised by killing the
actual audio server five days before a gig, so its flap was found by a human hearing
four notifications instead of by a test. UNIT and PROC are now injectable.

The watchdog is committed but NOT enabled: it needs one real-rig recovery test
before it earns a place in the boot set.

Two entries. #44's lesson is the one worth mining later: wiring six working
checks into one caller found two of them dead, because every check had been
verified once, interactively, by a human who supplied the missing piece without
noticing. A human types `zsh tools/foo.sh` out of habit and never learns the
file is not executable. An orchestrator has no habits.

#114 records that the overnight constraint list did real work rather than being
ceremony: "never write to a setlist .tidal" is what kept three ranked-"needed"
tasks off the queue, and drawing that boundary first is why the night was
productive instead of risky.

The sweep landed after the brief was written. Replaces the placeholder with the
measured figure and, more importantly, with the right DENOMINATOR: study/ is
documentation prose, so 79+30 of 799 overstates it the same way the harness bug
did. Also surfaces the CRASH class to him directly, since those are the ones
worth his time and the ones I deliberately did not touch.

The clean re-run after the harness fixes. Three numbers have now been attached
to this question and the first two were both about the tool:

    #112 said        14   (a baseline taken during the mute migration)
    pre-fix run     128   (the column-0 `$` artifact, ~49 of them fake)
    clean run        79 BROKEN + 30 CRASH, out of 799 files

And 799 is the wrong denominator too. Split by scope:

    OPAL setlist        13 files    0 BROKEN   0 CRASH
    live/              694 files   51 BROKEN  21 CRASH   ~10%
    study/              59 files   13 BROKEN   3 CRASH
    sandbox/blocks/test 21 files    5 BROKEN   3 CRASH

`study/` is Tidal DOCUMENTATION with inline examples — prose that was never
meant to compile as a track. Counting it inflates the figure the same way the
harness bug did, just more quietly. The number worth quoting is 51+21 of 694.

THE GIG IS UNAFFECTED. Zero setlist tracks in either category, and gig-up is GO.

The CRASH class is new and it is the valuable one. These tracks COMPILE, then
throw when queried — invalid mini-notation inside a string, which is exactly why
the type checker waved them through:

    "[d d] <d d d*2 d*4>]"     unbalanced bracket
    "~ ulgab?:1"               `?` before `:`
    "[0 .. 7 0]"               a range with a trailing element
    "<0.75 .. 0.8 0.8 .. 0.65>"  a sequence on the left of `..`

Every one would fail live, because the scheduler queries every cycle. Until this
run they were all reported as "FAIL — 0 orbit(s) SILENT".

Deliberately NOT fixed here. None are in the set, and each one needs a decision
about what PLN meant — "[0 .. 7 0]" has at least two readings and only he knows
which. Triage is the deliverable; the edits are his call.

Also records a counting caveat rather than hiding it: silent-eval prints a
track's STEM, not its path, so stems that exist in several directories are
marked as ambiguous instead of guessed. The real fix is repo-relative paths in
silent-eval's output — noted on #112, not silently papered over.

Leads with 'nothing needs your hands', because the previous two mornings led
with a fader he had to raise and that framing sets the whole read. Then the two
things that shipped, then the one finding that changes something he was already
told: #112's '14 tracks do not compile' was a harness artifact, and the harness
was calling his own dominant writing style a parse error.

Deliberately does NOT quote a corpus-wide count. The clean re-run was still
going; the number lands on the task instead. Quoting an unmeasured figure is
exactly how #112 got a wrong one.

The previous commit fixed three bugs in the tool whose entire job is catching
failures that make no noise. That makes its own failure mode the worst one
available: saying OK about a track that does not work, or BROKEN about a track
that does. It was doing BOTH, corpus-wide — and it stayed hidden because the 13
setlist tracks, the only ones anyone ever pointed it at, happened to dodge all
three.

Seven tests, all structural — no ghc, no rig, 0.06s:

  1. a column-0 `$` is a CONTINUATION (PLN's dominant corpus style), plus a
     guard that indenting does not FLATTEN relative structure below it, plus
     one for the ordinary inline style — which gets a test precisely because
     dodging the bug is what let the bug live.
  2. a non-zero exit with no SILENT line is a CRASH, and its stderr survives
     into the detail, because stderr is the only thing that says why.
  3. the exit code answers "is anything wrong", never "is anything silent".

MUTATION-CHECKED, because green tests on their own prove nothing. Each fix was
individually reverted and the suite re-run:

  remove the continuation indent  -> test_dollar_at_column_zero_is_indented FAILS
  key the return to silence only  -> test_exit_code_is_not_keyed_to_silence_alone FAILS
  both restored                   -> 7 passed

Test 3 is the one that matters. When crashes were split out of the silence
count, the return value was briefly left keyed to `silent_only`, so a corpus
where every single track failed to compile would have printed "OK — every
declared orbit emits events" and exited 0. That is a false green from the gate,
and a false green here is indistinguishable from a working set right up until
the downbeat.

Full suite: 229 passed.

#16 was ranked "bonus, can be 80/20ed". The cheat-sheet IS the 80/20: at OPAL
he is alone in a field with a laptop and a 48-cell control surface. The FOH
chain and the go-bag he can hold in his head. Twelve columns of CC numbers he
cannot.

GENERATED, NOT AUTHORED. A hand-written card is a copy of tools/lcxl_grid.py,
and a copy drifts silently the moment the grid changes — which it did twice
this fortnight. This reads the grid and the setlist directly, so a card that
disagrees with the rig is impossible by construction. `--md` rewrites
docs/GIG-CHEATSHEET.md; regenerate rather than edit.

THE EMERGENCY BOX GOES FIRST
The moment you reach for this page is the moment something is wrong. A card
that opens with a reference table makes you read past your own panic to get to
the answer. So the top of the page is eight failure modes and their remedy,
each one a symptom he has actually hit: LEDs dark but knobs alive (USB OUT
endpoint stall — replug), a track making no sound (it did not compile; one
block, one error), an orbit that will not die (it is the previous track's
pattern), mixer silent while Tidal looks fine (a fader; the desk wins on touch).

TWO LAYOUT BUGS THE FIRST RENDER CAUGHT
  * the title bar was hand-padded and off by one — now ljust'd like every
    other line, so it cannot drift again
  * gF3 and gM3 hold 7 and 8 orbits, which overran a side-by-side two-column
    family layout and silently misaligned the block. Stacked instead.

AND ONE COLUMN DELETED FOR SAYING NOTHING
The set table first carried a "roles" column. Every track in the set carries
kick+perc+bass+lead, so it printed the identical string 13 times and pushed the
page past 80 columns. Replaced with the orbits each track does NOT declare —
which is the fact worth having on stage, because an undeclared orbit is not
silent, it keeps playing whatever the previous track left there. That column is
the transition ghost list (#77), per track, at a glance. It also shows `perfect`
declaring all 12, i.e. the one track in the set that cannot inherit a ghost.

Verified: 0 lines over 80 columns, renders from a clean checkout, and the
family map printed matches lcxl_grid.filter_family/mute_family exactly.

#112 said "14 corpus tracks do not compile". Running the harness over the whole
corpus said 128. Neither number was about the music.

BUG 1 — column-0 `$` continuations reported as parse errors
  PLN writes a lot of the corpus like this:

      d1
      $ whenmod 128 129 (…)
      $ s "<k k*2 <k*2 k> k>"

  Pulsar sends a block wrapped in GHCi's `:{ … :}`, which suspends the layout
  rule, so that `$` at column 0 is a continuation and the track plays fine.
  `orbit_bindings` rewrites `dN` into `name = idcp` and emits the block into a
  real MODULE, where a column-0 `$` opens a NEW top-level declaration — a hard
  parse error. The harness then printed, in these words:

      ✖ THE TRACK DOES NOT COMPILE — it will fail live too.

  about code that has never once failed live. The intent to indent was already
  written in the comment above the line ("becomes `name = idcp` + `  $ x`");
  the indent itself was never applied. Two spaces on lines[1:] restores it and
  preserves every relative indent below.

  Why it stayed hidden: ZERO of the 13 setlist tracks use that style, and the
  setlist is what anyone actually runs. The harness was green on everything it
  was ever pointed at. Same shape as the chmod bug in the previous commit — a
  thing verified once, in a context that no longer covers the corpus.

BUG 2 — a crashed probe reported as "FAIL — 0 orbit(s) SILENT"
  The Haskell probe sets its failure flag ONLY when it prints a SILENT line, so
  a non-zero exit with no SILENT line never measured silence at all — it threw
  while QUERYING. `check_track` mapped any non-zero return to "SILENT" and
  discarded stderr, so the one string that said what went wrong was thrown away
  and replaced by a self-contradicting verdict. New CRASH verdict keeps stderr.

  It immediately paid for itself. `live/chill/dub.tidal` compiles, plays three
  orbits, then throws:

      Syntax error in sequence:
        "<0.75 .. 0.8 0.8 .. 0.65>"
                     ^

  A real mini-notation bug, precisely located — and one that WOULD fail live,
  since the scheduler queries every cycle. It had been sitting behind a message
  that said zero orbits were silent.

BUG 3 (latent, found while fixing 2) — a false green in the exit code
  Splitting the summary so crashes stop being counted as silence revealed that
  the return was keyed to the silence count alone. A corpus where every single
  track failed to compile would have printed "OK — every declared orbit emits
  events" and exited 0. The exit code now answers "is anything wrong", which is
  the only question a gate may answer — this is the one tool whose entire job
  is catching the failure that makes no noise.

Verified: setlist still 13/13 green and gig-up still GO (so no regression on
the path that matters); dub now reports THROWS WHEN QUERIED and exits 1;
previously-"broken" tracks compile. The corpus-wide count is deliberately NOT
restated here — it needs a clean re-run, and quoting a number I have not
re-measured is how #112 got a wrong one in the first place.

A fortnight of debugging produced a shelf of good checks: check-boot,
check-mix, silent-eval, fix-mute-roles, pvlint, orphan-orbits. Every one of
them worked. Every one of them was also a command PLN would have to REMEMBER —
and on Sat 8 Aug at ~19:00 he is in a field, alone, with a soundcheck window
and no assistant. A check you have to remember under pressure is a check you
do not run.

So gig-up.sh contains no new checking logic at all. It is a running order, an
exit code, and a decision about which failures are GO-BLOCKING. Its whole value
is that there is now one thing to type.

WHAT IS HARD vs SOFT, and why the line is drawn there
  HARD  boot helpers    one parse error in BootTidal.hs silences EVERY track
                        at once, invisibly, while a stale ghci holds the old
                        definitions. First check for that reason.
  HARD  ardour faders   earned it three sessions running, a DIFFERENT set of
                        tracks each time: 07-28 had 05/06/08/10 at -inf,
                        07-30 had 06/10/12, 07-31 had 10. Three recurrences is
                        not bad luck, it is a property of MIDI-learned faders
                        that drift from the saved session in both directions.
  HARD  compiles / mute map / pvlint
  SOFT  transition ghosts   real (#77) and known-open. A gate that goes red for
                            work you have consciously deferred is a gate that
                            gets ignored.
  SOFT  LCXL present        legitimately unplugged at a kitchen table. At the
                            venue, a warn here means the set has no hands.

TWO BUGS FOUND BY THE FIRST RUN — which is the argument for building it
  1. check-boot.sh and check-tracks.sh were mode 644. NOT EXECUTABLE. The
     pre-gig audio gate has never once been runnable as `tools/check-tracks.sh`
     since it was written; it only ever ran when someone typed `zsh tools/...`.
     Every green memory of it is a green memory of a different invocation.
  2. pvlint's `--setlist` is a MODIFIER ("these paths are ordered"), not a
     mode — it still requires paths. `python3 -m pvlint --setlist` exits 2 on
     an argparse error, which from the outside looks exactly like a lint
     failure. gig-up now borrows set-coherence's setlist_tracks() rather than
     teaching a fourth file to parse the setlist.

Both are the same shape and it is this rig's signature failure: a thing that
was verified once, in a context that no longer holds. Assembling the checks
into one caller is what re-executed them under fresh conditions.

DESIGN NOTES
  * Cold by default. Makes no sound, sends no MIDI, writes no .tidal. Safe to
    run mid-set, at 3am, or backstage. The audio gate is --audio, and it skips
    itself if the cold gate is already red — booting 13 tracks to confirm what
    a typecheck just told you is 10 wasted minutes at a venue.
  * Never touches CC 77-84 (Ardour track gains; 77 down is total silence) and
    never sweeps 93 (gPanic).
  * Scope is the SETLIST, not the corpus. 14 corpus tracks do not compile
    (#112) and must never make the gig gate red.
  * Failures print FIRST and carry their fix command inline, because the read
    order in a field is "what is broken / what do I type".
  * GO states its own limit every time: check-mix reads the SAVED session, so
    green means "Ardour will BOOT with this fader up", never "the mixer IS
    correct". Saying it in those words is how the -inf bug stops hiding behind
    a clean report.

Verified: NO-GO path observed for real (2 failures, both genuine, both fixed);
GO path 7/7 green; --help, --quiet and unknown-flag paths exercised.

feat(tools): set-coherence — 108 controls sit idle under his hand, and three of them are systematic (#113)

PLN, 2026-07-31, having walked the OPAL set by eye:

  "when i went across set i saw some unmapped effects/button, a 'set coherence
   mapping review' and quick wins suggestions (bass octersub, lead slice, etc etc)
   could be good ways for you to fill-in suggest"

He was right, and the number is larger than "some": **108 idle slots across 117
declared orbits in the 13 setlist tracks.**

## Why this needed a tool rather than a read-through

Absence has no line number. The surface is a fixed grid — each orbit owns a known set
of cells (effect knob, maybe a second, gate button, maybe a second) — so a track that
declares d5 and never references ^33 leaves d5's knob inert for that track's whole
duration. You cannot grep for a control that isn't there; you can only compute it, by
subtracting what the track USES from what the orbit OWNS.

The tool reports both directions of the same coherence question, deliberately:
  IDLE     orbit owns a slot the track never uses   -> unused affordance (new)
  STARVED  track drives more controls than it owns  -> already covered by PV008
                                                       and the FIXME(#54)/(#94) markers
Fixing one can create the other, which is why they belong in one report.

## Three systematic findings, not a flat list

1. **The drums' own effect knobs are the deadest controls on the board.**
   d2.fx (B2 ^30) idle in 12/13 tracks, d1.fx (B1 ^29) 9/13, d3.fx (B3 ^31) 8/13.
   d1-d3 own exactly ONE per-orbit knob each — their row-C cells are the family
   filters — so this is not a spare slot going unused, it is their only knob. The
   KICK has no per-orbit effect knob bound anywhere in the entire set.

2. **Row A (d9-d12 effects, A5-A8) is ~90% idle.** d11 6/6, d12 5/5, d10 6/7, d9 5/8.
   That is #98's proposal (pool A5-A8 in order of appearance) with its supporting
   numbers: four knobs almost never used, on exactly the orbits that overflow.

3. **d6 has gates but NO knobs, 5 of 5 tracks.** Fallout from the d7->d6
   consolidation: ^58/^90 came along, ^34/^54 did not. And vague_de_crime's d6
   references nothing at all — no knob, no gate.

## Suggestions are grounded, not invented

Measured what PLN actually binds to a per-orbit effect knob across the set — perc
legato(4); bass lpf(3), gain(1), bandf(1); lead width(2), modIndex, filterRange,
cutoff, resonance; kick nothing — and the suggestion table echoes that idiom before
extending it (memory feedback_self_reference_rhythm_study: extract the grammar from
his own tracks rather than importing generic Tidal advice).

Worth recording because it inverts the ask: his two named wants sit OUTSIDE his
measured idiom. He binds lpf to bass knobs and never octersub; slice lives on his
buttons, not his knobs. So "bass octersub, lead slice" are additions he is reaching
for, not corrections of drift — and saying so is more useful than pretending the
corpus already implied them.

## Deliberately not applied
Every suggestion changes how a track sounds and he is rehearsing on these files this
weekend. Output is docs/2026-07-31-set-coherence.md to skim and cherry-pick. The tool
also never suggests binding CC 77-84 / 13-16 (Ardour track gains), 93 (gPanic), or the
family cells 49/50/51 + 73/74/75.

Also confirms the #94 remap held: "borrowed" CCs — ones an orbit drives without owning
— are 0 or 1 across every track in the set.

`# width (range 0.2 0.8 "^33") -- /!\ FIXME EXPLOSIF /!\` in gimme_acid's acid synth.
PLN has had a hazard marker on that line for a while. The cause is mechanical and
checkable without any audio:

^33 is an A/B knob, so BootTidal's #55 seed gives it 0. The seed policy states its own
convention — "range <neutral> <extreme>, so 0 == neutral for effect knobs" — but 0.2 is
NOT neutral for `width`: the preset reference PLN pasted two lines above the call says
nominal is width = 0.51. So the line boots the synth at an extreme and the knob sweeps
it THROUGH nominal rather than away from it, which is a good candidate for whatever
"explosif" means to his ears.

This is the exact shape BootTidal.hs already documents as an authoring bug, using
piment_bresilien's `legato (range 0.05 2 "^52")` as its worked example, and its stated
remedy is to fix the range in the TRACK rather than special-case the seed.

Deliberately NOT applied. The candidate is `range 0.51 0.8 "^33"` — boots at nominal,
knob only adds — but it costs the 0.2-0.51 downward half of his sweep, and that is a
decision about how the track sounds. The analysis goes in the file next to his marker
so the next reader inherits the diagnosis instead of re-deriving it; the sound stays
his.

Did remove one unambiguous thing: `# filterRange (range 0 8 "^53")` appeared TWICE in
consecutive lines, same value, so the second was a no-op paste.

MORNING.md rewritten for today. Leads with the finding that actually matters at J-7:
check-mix.py reports Tidal 06/10/12 at -inf dB in the SAVED Ardour session, and
cross-referencing those against the orbits each setlist track declares shows **9 of 13
tracks would lose at least one orbit on a relaunch** (#62). The live faders are
probably fine — PLN has been playing — which is precisely the bug: the fix is not
persisted, so it dies with the process. Ctrl+S is the whole remedy, and it is his
hands, not mine: the D row is MIDI-learned to CC 77-84 and writing from outside
desyncs it from the physical desk.

Also flagged in MORNING.md rather than guessed: yesterday's file said "OPAL Day-5" and
today's framing was "J-7". There is no OPAL 2026 entry in Web/www/content/lives/2026/,
which CLAUDE.md names as the canonical source for gig metadata, so the countdown is
unresolvable from here and is asked rather than invented.

refactor(corpus): 917 lines move the kick onto its own mute button — every track, not just the set (#105, #70)

PLN, mid-migration, when I had scoped this to the 13 setlist tracks:

  "good, and not only in the set am i right? did we not do the whole refact on the
   other .tidal files on my repo?"

He is right, and the reason is muscle memory: his fingers do not know which folder a
track lives in. A half-migrated corpus is worse than either extreme, because it makes
the surface unpredictable exactly when he reaches for an old track mid-set.

## What moved

913 lines across 213 files, plus 7 orbits given a mute they never had:

     505+4  gM1 -> gM2      percs vacating the kick's button (the BULK of the work —
                            not the kick itself, which is only 133 lines)
    127+1   gM2 -> gM1      the kick, arriving
      66+4  gMute2 -> gM1   same, in the legacy long spelling
        64  gM1 -> gM3
        58  gMute3 -> gM3   spelling normalisation only
     41+2   gM2 -> gM3
      12+2  gMute2 -> gM2 / gMute1 -> gM2 / gMute1 -> gM1 / gM3 -> gM2  the tail
         7  FILL            orbits with NO mute at all, in the SET only

Only the setlist was filled. PLN: "ensure we have that consistently in all set, and
flag in other non-set tracks for me to see." Outside the set a missing gate may be a
composition choice, so docs/2026-07-30-orbits-without-a-mute.md REPORTS the rest and
writes nothing — split into the cut that makes 2397 findings actionable:

    A · partial  110 files   321 blocks   uses family mutes but missed some
                                          <- the oversight he meant
    B · never    354 files  2076 blocks   no orbit has one; predates the convention
                                          <- a style era, not a mistake

## Also, in this diff: #70

`live/midi/nova/dnb/liquid/you_my_sunshine.tidal` had d5 (the voice) and d11
(`no_sunshine:4/4` chopped) both on `# cut 5`. A cut group is monophonic, so every
d11 event killed the vocal mid-word. That is ALSO the `-- FIXME VOICE OFF` PLN had
left on d5 line 59: one root cause wearing two labels, found by pvlint PV004 on
2026-07-28 and confirmed by reading the file today. d11 moves to `# cut 11`, matching
the group-N-equals-orbit-N convention d3/d6/d8/d9 already follow. The comment says
how to revert in one line if the vocal was meant to be monophonic across both layers.

## How this was verified, given it touches 213 files PLN has to play in 7 days

1. **Convergence** — `--apply` twice; second pass proposes 0 rewrites and 0 fills.
2. **Structural proof, not spot-checks** — diffed every changed line against HEAD and
   asserted the two sides are identical once mute tokens and whitespace are stripped.
   908 of 917 changed lines are provably mute-plumbing-only; the other 9 are listed
   and each accounted for (4 fills onto head lines with no `$`, 3 trailing-whitespace,
   1 FIXME annotation, 1 the cut 5->11 fix). Nothing else in the corpus moved.
3. **Compile-invariance argument** — gM1/gM2/gM3 are the same type and arity in
   BootTidal, so substituting one for another cannot change whether a file compiles.
   Only the 7 fills alter expression structure, and all 7 are in the set.
4. **silent-eval --seeded on the set: 13/13 ok**, before and after.
5. Suite: 222 passed.

THE HONEST LIMIT, stated because a green check here is misleading: every button CC
seeds to 0 at boot and gMute<N> is `midiOn "^7N" (mask "f*16")`, so all three mutes
are equally inert until pressed. silent-eval therefore proves nothing got silenced at
boot and NOTHING about behaviour under a press — which is the only thing this change
alters. The real gate is PLN's ear, or pv-at (#74).

NOTE FOR PLN: 213 files changed on disk. Anything open in Pulsar now has a stale
buffer, and Pulsar saves the BUFFER — reload before evaluating or your next Ctrl-S
silently reverts the migration in that file.

PLN settled the gM question on 2026-07-30, and the answer was not the one this task
had assumed for a day:

  "I want the kick, now on d1/fader1, to have its mute on F1 botrow button. I want
   the C1 knob to djf all percs, C2 only bass djf, C3 melodies DJF. however i want
   the mutes consistent across all all tracks: m1 d1 / m2 other percs / m3 all
   bass+melodics"

## The thing I had wrong

#105 was framed as "align gM<N> to gF<N>" — make the mute index mirror the filter
index so an orbit's knob and button share a column. I measured 35 mismatches in the
set under that invariant and was ready to fix them.

That invariant is WRONG, and the measurement was answering the wrong question. The
filters and the mutes are two different gestures:

  FILTERS want the rhythm section as ONE BLOC     — you sweep the drums together
  MUTES want the kick SPLIT OUT on its own button — you drop the kick alone

So gF groups {d1,d2,d3,d8} and gM splits d1 from {d2,d3,d8}. Under the real map the
edit count is 55, not 35 — and `d4: gF2+gM3`, which the mirror-invariant called a
mismatch in 11 tracks, is CORRECT. Had I shipped the "alignment" I would have moved
11 correct lines and left the kick sharing a button with the hats.

Recorded as tools/lcxl_grid.py `_FILTER_FAMILY` / `_MUTE_FAMILY` — in the ONE
authored grid (#97), with his words above it, so the asymmetry reads as deliberate
rather than as drift. `as_dict()` now exports `orbit_family` so the HUD can say which
knob AND which button own an orbit; `--generate` rewrote tools/lcxl_grid.json and the
HUD package's own copy (outside this repo, in Tools/pulsar-parvagues-hud).

## tools/fix-mute-roles.py

Convergent migrator: `--apply` rewrites, `--check` exits 1 on drift (a pre-gig gate),
`--fill` gives a mute to a block that has none, `--report-missing` writes a markdown
report instead. Reads the map from lcxl_grid rather than restating it.

## PV012 / PV013, so it cannot re-drift

PV012 flags an orbit whose gM does not match its ROLE — deliberately NOT "gM index
!= gF index", and there is a regression test asserting `d2 $ gF1 $ gM2` is clean,
because a rule written to the mirror invariant would flag that valid line. PV013
flags an orbit with no family mute at all: PLN's "no mute is oversight!" — nothing on
the surface can drop such an orbit, so taking it out means editing live.

## PV013 immediately caught a bug in the migrator that wrote it

Minutes after PV013 existed, it flagged `d5` in you_my_sunshine — an orbit the
migrator had walked straight past. Cause: fix-mute-roles had its OWN orbit regex
requiring `^dN $`, and that file's head line is `d5  -- The Voice of Love`, no `$`,
with the gates on a commented continuation. pvlint's regex allows `dN` followed by
`$`, `--`, or EOL. Two parsers disagreeing about what an orbit IS is the same class
of bug as two copies of the grid, so the migrator now imports pvlint's parser and
owns none of its own. Re-running found 14 more rewrites and 7 more missing mutes it
had silently skipped — the miss was under-application, not corruption, but it would
have left exactly the inconsistency this task exists to remove.

Also fixes #109: the orphan-orbits hand-measurement for you_my_sunshine still
expected d7, which PLN moved to d6 in ce887b78. Re-derived by reading the file's
column-0 declarations, not by pasting the parser's output — a hand measurement that
quotes the thing it checks is a tautology.

Suite: 222 passed, 0 failed (8 new PV012/PV013 tests).

Twelve files PLN had been working live, committed as his own change before any
tooling touches them. Not a refactor: this is the set moving under his hands.

The through-line is the #94 remap paying off. The remap had COMMENTED OUT every
control line that fired across columns — honest, but it left orbits playing with
their hands tied. He is now restoring them onto the CCs the orbit actually owns:

  gimme_acid  d1's kick variations were disabled because they rode ^42, which the
              remap gave to d2. Restored on ^41, d1's own gate — so the kick keeps
              its midiOn/midiOff variation without borrowing a neighbour's button.
  perfect     d5's SUPERSTAR chain was 14 commented lines. Rebuilt live on ^53/^33
              with a nested `midiOff "^89"` mask, so the slice logic is gated by a
              button instead of hardcoded.
  desire      the compile blocker is gone: `# pan 0.42plz /se` (line 46) is fixed.
              That one typo, in a file with no blank lines — hence ONE do-block —
              was killing every orbit in the track. Found by silent-eval learning to
              blame the track instead of itself; `silent-eval --seeded` now reports
              13/13 ok, the first time the whole OPAL setlist has built cold.

Also: d7→d6 in you_my_sunshine (with ^91→^90, ^59→^58) — the same consolidation he
made in do_it_right, freeing an orbit slot. This invalidates a hand-measured fixture
in tools/tests/test_orphan_orbits.py; the parser is right, the fixture is stale.

Orbit naming throughout (BASSE DU 3e cercle, Synth lapin speed, SuperStars, Breaks
divins), gain tuning, and backlog notes. Structural changes stay in the next commit
so this one reads as pure performance work.

Four tasks closed 2026-07-29 evening: the silent-eval harness fix (and the real bug
it uncovered), the take-pack sidecar+archive tool, and the desire rhythm study.
Written for a cold reader per the archive's own convention — commit hashes are in
git log, this is the narrative and the numbers that don't show up in a diff.

The existing MORNING.md led with a warning about ctrl+enter fading to silence over
4 bars (#79) — fixed and confirmed by ear hours ago ("no sample goes oblivion!! 👏").
Leading with a solved problem buries what actually needs review tonight.

Replaced with five things worth his eyes/ears, ordered by urgency: the desire
compile break (#108, a real gig blocker), the corrected take-94 mix (v2, the trim
threshold fix), silent-eval now covering all 13 setlist tracks, the slop renderer's
first real output (not yet good enough to send, flagged as such), and the four
desire bass-rhythm alternatives to A/B. Everything else from the session is
mechanical and needs no listening, so it's kept to one short list at the bottom
rather than mixed in with what actually needs his judgment.

PLN: "i gotta solve desire bassline is not great rythmics coulld be cooler more nto
more ete_a_mauerpark or insouciance or haunted_house i guess."

Read the three references he named and extracted what they share. The diagnosis turns
out to be mechanical rather than a matter of taste. desire's d4 is

    note (scale "aeolian" "0 . 0*<1 2> . <0 3> . 0*<1 1 4>" + 2 - "[24,36]")

— four EQUAL groups, the same degree in nearly all of them, never a rest, never
displaced. The only variable is density, applied to a repeated note. That is a
metronome with a stutter, and no amount of filter movement makes it a riff.

The six mechanisms both references use and desire has none of:
 1. uneven durations via `@` (insouciance 5:1:1:1 of 8; mauerpark 1:1:14 of 16) —
    one long note holds the harmony, short notes carry the motion;
 2. the pitch pattern IS the rhythm pattern, so a new note is a new onset, instead
    of a static degree with density stacked on top;
 3. calage — `(0.125 <~)` displaces the whole line so it pushes against the kick
    rather than doubling it;
 4. accent follows rhythm — `|* gain ("1@5 0.95 0.98 0.95")` carries the SAME
    weighting as the notes;
 5. space — rests inside the phrase and whole bars masked out;
 6. the phrase is longer than the bar (4-bar alternation, plus `slow 2`).

Four proposals in study/desire_bass_rhythm.tidal. A, B and C each isolate ONE
mechanism so what he hears is attributable; D combines all six. Pitch design is held
constant everywhere — aeolian, +2 root, "[24,36]" double — because the key was
measured at corr=0.903 and is not what he is unhappy with.

All four verified through silent-eval --seeded: they compile and every one emits
events cold. Measured onsets per bar: B 3.0, A 4.0, C 4.5, D 5.75, against ~5.5 for
what desire has today. D is deliberately DENSER than the current line and still
reads as a riff — which is the control that shows the problem was never "too many
notes", it was where they fall and how long they last.

desire.tidal is untouched: it is in his dirty working tree, and it also does not
currently compile at all (#108).

PLN has said repeatedly he is bad at socials and hoped slop would carry that weight.
The plan for this has been complete since the clip planner landed; the renderer had
never been written, so nothing could actually be produced. It can now.

`visuals/slop/render_slop_clip.mjs` takes either a planned cut (`--idea desire --cut 90`,
from clip_ideas.json with its measured tilt/kick-density/novelty and playset choice) or
an arbitrary window of any file (`--audio ... --start ... --dur ...`), and renders a
vertical 1080x1920 reel with Slopmotion reacting to that audio.

DESIGN NOTES WORTH KEEPING

- It lives in THIS repo, not hexa. hexa is Kevin's project and its main is clean; the
  script borrows its node_modules through createRequire instead of being committed
  into someone else's tree. Its scratch files go in hexa/public/__slop and are removed
  in a finally block — nothing is left behind in his working copy.
- Capture is REALTIME via CDP screencast, not per-frame screenshots. Hydra's animation
  clock is wall time, so frame-stepping plays motion back ~3x fast and drifts away
  from the audio it is supposed to be reacting to. Screencast frames arrive timestamped
  and become a vfr->cfr concat, so a dropped frame is a slightly longer one — invisible
  — rather than a wrong clock.
- The bands are computed from the REAL audio through an AnalyserNode, not from the
  synthetic pump the FX-preview script uses. Kick is a fast rise in the sub band above
  its own slow running average, because an absolute threshold would need retuning per
  track while a relative one rides the mix.
- Backdrop clip choice is a hash of the job name, not Math.random: two renders of the
  same cut must be comparable, which is the entire point of a review pipeline.
- The viewport is 9:16 natively rather than a 16:9 scene centre-cropped afterwards,
  which would throw away most of the motion.

TWO ENVIRONMENT PROBLEMS FOUND AND FIXED WITHOUT COLLATERAL

- The hexa checkout could not boot at all: 9 declared dependencies were missing from
  node_modules, so vite 500'd on @supabase/ssr and @vercel/speed-insights. `npm ci`
  refused (lock out of sync with package.json), and plain `npm install` would have
  rewritten package-lock.json — a TRACKED file in Kevin's repo. Installed the 9 with
  --no-save --no-package-lock instead; his tracked files are byte-identical after.
- Playwright 1.60 wants chromium build 1223 and the cache has 1217. Rather than pulling
  ~150 MB into PLN's shared browser cache unasked, the script finds the newest full
  chromium already present. Full chromium, not chrome-headless-shell, which ships
  without the GPU/ANGLE stack this needs.

WHAT IS NOT DONE: frame rate. 7.0 fps at 1080x1920 under swiftshader — the page itself
renders at 7 fps, so it is the GL backend and the machine, not the capture. The two
follow-up measurements (hardware EGL, and half resolution) both came back at 1.1 fps,
which is not a verdict on either: load average was 12.5 with Pulsar burning a core from
the #7 marker leak. They measure the machine. Recorded in FEEDBACK.md with an explicit
warning to re-measure on a quiet box before concluding anything — a false workaround
adopted from a contended benchmark is very hard to remove later.

Deliberately not pushed further tonight: a sustained software-GL render is a CPU load
test, and those do not run at night here while the rig is up.

Four of thirteen setlist tracks could not be built, so their orbits had never been
verified cold. Two causes, both the same shape: THE HARNESS DIVERGED FROM THE BOOT
ENVIRONMENT, which is what #93 predicted would keep producing new errors.

1. Ambiguous `cutoff`. BootTidal defines `let cutoff = pF "cutoff"`. In ghci a let
   binding SHADOWS an import, quietly and legally. The harness dedents that block to
   module top level, where there is no shadowing — GHC reports "Ambiguous occurrence"
   and the track becomes unverifiable for a difference that does not exist on the rig.
   Fixed by hiding every name the generated module defines from the Context import,
   with the hide-list DERIVED from the generated text rather than listed, so it cannot
   drift when BootTidal gains a helper. 235 names on the current boot.

2. Overlapping IsString instances on a chord literal
   ("<gb3'maj db3'maj bb2'min bb2'maj>" could be ParseBP's `Pattern a` or Simple's
   `ControlPattern`). ExtendedDefaultRules and NoMonomorphismRestriction are ON BY
   DEFAULT IN GHCI, which is where these patterns actually run. Turning them on
   matched the environment; annotating each literal would have treated the symptom
   one track at a time forever.

=== And then the third failure turned out not to be a harness limit at all

`live/collab/raph/desire.tidal` line 46 reads `# pan 0.42plz /se`. It is committed,
not a stray working-copy edit. `plz` is not a Tidal function and the file has NO
blank lines, so it is one single do-block: that one typo means every orbit of desire
is dead on ctrl+enter. It is on the OPAL setlist, six days out.

The tool had been reporting it as "BUILD FAILED — harness limit, NOT a track verdict".
That message was written to protect the music from the tool, and it was right twice
and catastrophically wrong the third time: it told the reader, in bold, to ignore a
broken setlist track. A parser miss must never masquerade as a data conflict, and the
inverse is worse.

So the verdict is now computed, not asserted. Errors whose line numbers fall inside
the track's own generated lines get a distinct BROKEN verdict — "THE TRACK DOES NOT
COMPILE — it will fail live too" — and count as a gig blocker. Errors in the boot
helpers or the scaffolding stay a harness limit. Line numbers are also mapped back to
the real .tidal by matching the offending line's TEXT, because a generated-module line
number is useless to whoever has to fix the file and looks authoritative while being
so; unmatched or ambiguous lines are left alone rather than guessed at.

Result: 12/13 build and every declared orbit emits events cold. The 13th is a real
bug in the music, reported as one, with a file:line you can jump to.

NOT FIXED HERE deliberately: desire.tidal is in PLN's dirty working tree and Pulsar
saves the BUFFER, so a disk edit made behind his back can be silently reverted by his
next ctrl+S. It is item 1 of MORNING.md instead, with the one-line fix, to be applied
in the editor where it will actually stick.

Take 94 came back with no track trail, and there was no way to tell from the logs
whether the watcher had seen the track changes, failed to bind them, or simply
never run. It had been running for 2h46m. It had emitted nothing but systemd's own
start/stop lines the entire time.

Two independent causes, and either alone is enough to blind the unit:

1. Python BLOCK-buffers stdout when it is a pipe, and under systemd it always is.
   Even the lines that were being printed would have sat unflushed in a 4 KB
   buffer. Fixed with Environment=PYTHONUNBUFFERED=1 in the unit.

2. `--watch -q` suppressed the track-change line itself, so there was nothing to
   flush. That is the wrong thing to make quiet. `quiet` should mean "no per-CC
   chatter" — thousands of lines an hour — not "never say what you are painting",
   which is a handful of lines an hour and the entire purpose of the daemon. The
   track-change log is now unconditional.

Verified live rather than by reading: restarted the unit, wrote a different track
to ~/.cache/parvagues/current-track, restored the original, and read the journal
back. Both transitions appear with their binding counts (desire 28 controls,
vague_de_crime 20), and a re-parse of the same file is labelled distinctly from a
real track change — that distinction matters because an edit-and-save must keep
the touched-state while a genuine track change must clear it.

Same family as the stale-binding pattern this rig keeps producing: something
resolved once, invalidated by an event, with no error anywhere. Here the error
reporting itself was the thing that had been silently resolved to /dev/null.

Two things in one file, both from the same evening's practice recording.

=== PART 1: `pack` — the performance must travel with the audio (#106, #107)

"ensure recordings have the midi too, keep this from blowing up tracking either
ardour rec or compressing to avoid storing hours of no-move?"

Today the audio is permanent and the performance is not. Ardour's sources live
forever under the session; the MIDI and the track timeline sit in a gig-log
directory that prune() deletes on a timer. So the 29862 CC events that explain
take 94 were on a countdown while the audio they explain was not. `pack` writes a
sidecar next to the audio: meta.json (window, sr, subtype, per-orbit peak dBFS,
silent orbits, xrun delta, gear + boot state), midi.jsonl (every cc/note/track/mark
record in the window, copied verbatim out of the log so pruning cannot reach it),
and edl.json.

The sidecar states its own honest limit in a field: gig-log COALESCES CC per
(port, channel, controller) per second, keeping count + first/last/min/max. That is
a faithful summary of a knob sweep and is what makes the log cheap, but a .mid
rendered from it would be a reconstruction, not a recording. The JSONL is the truth.

`--compress` then wavpacks every source and verifies each one with `wvunpack -vm`
before reporting a single byte saved — an archiver that reports a ratio without
checking is reporting a hope. The .wav is NEVER deleted; removal stays a separate
human step after the take has been heard. FLAC was rejected outright: it cannot
store 32-bit float, and routing through s32 would hard-clip the 8 orbits sitting
over 0 dBFS while looking like a win on the silent one.

Two measurement bugs found by running it on real data rather than trusting it:
 - xrun is a RUNNING TOTAL in gig-log's `s` records, not a per-tick delta (its own
   CUMULATIVE tuple exists for exactly this mistake). Summing it reported 58,668,469
   xruns for a 21-minute take. Rebased to last-minus-first: 230, a real number.
 - Records are not events. A `cc` record is a one-second bucket carrying `n`, so
   printing the record count understated the performance 25x (1167 vs 29862).

=== PART 2: the trim threshold was protecting every bad transition

PLN heard the first machine-trimmed mix and found the hole: "silence around 18 is
not normal its a bad transition ahah in a real mixing work wed cut it proper and
fade or at least trim most". The gap he heard WAS detected and reported, and
survived anyway. SILENT_RUN_S was 20 s; the three gaps that made it into the mix
measured 17 s, 18 s and 11 s. Every one sat just under the bar. A threshold chosen
so the trimmer could never eat a musical break had silently become a threshold that
preserves every failed transition instead.

Three changes, all of them his rule rather than a tuning:

1. CLAMP, don't binary-cut. The bar drops to 10 s, but a detected silence is no
   longer removed whole — the last 2 s before the music resumes is kept as a
   breath. That is "or at least trim most" implemented literally, and it makes the
   lower threshold safe: the worst case if the rule fires on something musical is
   now a shortened rest, not a missing one. classify() computes both the detected
   region and the removed region so the report and the mixer cannot disagree.

2. FADE every seam. Every cut this tool had ever made was a butt-join, which in a
   live room is a click. Each removal now fades out into the cut and back in out of
   it over 250 ms, applied sample-accurately — the old mixer decided keep/drop per
   one-second block by its midpoint, quantising every boundary to +-0.5 s.
   Verified on the re-rendered take: max sample-to-sample step at the five seams is
   0.0001-0.0029 against a p99.99 of 0.0562 for the take as a whole, i.e. the joins
   are 20-100x below ordinary musical transients.

3. BRIDGE blips. A dead patch is rarely clean end to end, and one orbit's tail
   crossing -60 dBFS for a second splits it into pieces that each fall under the
   bar. Only single-orbit blips are bridged: a real hit lights several orbits, and
   that is what keeps this from swallowing a one-shot.

Result on take 94: 143 s trimmed instead of 119 s, all five gaps closed including
the one at +18.0 min he named, output 18.75 min, peak -3.36 dBFS.

=== `selftest`

Added because two of these rules could not be validated on real material — the
bridge fires zero times on take 94 and a correct fade is inaudible by construction.
A rule you have never seen fire is a hope, not a rule. It found a bug on the first
run, in the test's own arithmetic: asserting "no span survives a multi-orbit hit"
was wrong because the left half is exactly SILENT_RUN_S long and qualifies by
itself. The property that actually matters is that no trimmed span ever CONTAINS a
moment when the music was playing, and that is what it now asserts.

Correction to a verification, worth recording: the first check of the re-rendered
mix reported an 11 s silence still present. It did not exist. I measured the output
against -60 dBFS while the mix carries a uniform -16 dB, so I was testing at an
effective -44 dBFS at source. At the matched threshold exactly one rest over 4 s
remains, 9.25 s, correctly left alone as musical. The right lens per control, again.