Eight commits of rig work that had accumulated on the branch while the shared
checkout sat on it. Reviewed as a diff, not taken on faith:

  perf.sh        stop DWIM-launching Pulsar into /usr/local/sbin -- 183
                 SIGABRTs between 08-19 and 09-05 because BASH_SOURCE resolved
                 to the deploy path, and the perf watcher re-entered every 30s.
  rig_units.py   neither LCXL painter may be auto-started blind; gig-up.sh's
                 generation-aware chooser owns that decision.
  midiviz        its own Wayland app id (desktopFileName, not applicationName),
                 so focus and KWin rules can name this window and no other;
                 plus the CC readout PLN asked for while wiring.
  logs 040/041   the phantom port, and the GIG UP rewire parked behind the
                 Ardour double-launch bug.

The four other branches were already upstream by patch-id.

# Conflicts:
#	TASKS_DUMP.md

PLN ran the root install at 20:06. Verified rather than assumed: the
deployed /usr/local/sbin/perf-audio md5 matches the repo copy, the bug's
fingerprints (nohup pulsar, SCRIPT_DIR) are absent from the deployed
script, and prioritization is retained.

Behavioural confirmation is the part worth keeping: the Bridge's perf
watcher still reasserts every cycle, but now only prioritizes -- no
launch attempts in the journal -- and the last Pulsar SIGABRT in
coredumpctl is 15:58:44, before the deploy. The counter that ran 183 deep
since 2026-08-19 has stopped.

Structured archive entry. The keeper: two bugs hid each other for a month. One
file's module-level sys.exit aborted pytest collection for the whole directory,
so the obvious command ran zero tests while 273 passed file-by-file — and that
invisibility concealed the file's own month-old red check, whose NAME was a
complete bug report about a soundcheck being counted as OPAL's first track.

A test nobody can see failing is worse than one that does not exist, because its
name reads as reassurance. Fixing the harness surfaced a real correctness bug at
no extra cost.

Refreshes the resume state. The R1-R5 block from the first pre-compact is now
partly stale — R2 closed (PLN raised the faders and saved; check-mix reports all
12 orbits reaching master), both R5 test items fixed, and the suffix question
answered by evidence rather than taste.

Three things landed after those resume points that a cold reader needs: pytest
tools/tests/ had been running ZERO tests because one file's module-level
sys.exit aborts collection, and behind that invisibility the OPAL setlist's
first track was a soundcheck, so every tool reading it believed the set had 16
tracks opening on Quand on décolle. Both fixed; suite is 274 green in one
command.

Separated what is genuinely PLN's from what is merely undone, because that
distinction is the whole value of this block. Four decisions block everything
(--go on the 14, /desire specifically, CosmicFest tone, gig page as-performed vs
as-released) plus one ears question about nine faders moving where four were
broken. Four next moves need nobody, and the first is a 15-minute audit of
whether anything consumed the phantom setlist positionally.

Two bugs that were hiding each other, which is why both survived a month.

test_setlist.py ended in a module-level sys.exit(), pytest raises that during
collection, so 'pytest tools/tests/' aborted the whole directory and ran NOTHING
while 273 tests passed file-by-file. And what that invisibility concealed was
test_setlist's own failure, on a check whose name states the bug: OPAL's backlog
opens with '## SOUNDCHECK / -- Quand on decolle', entries() took every item under
the gig heading, and so every tool reading the setlist believed the set had 16
tracks opening on Quand on décolle. The recording has 15, opening on Ceci n'est
pas Une Bombe.

A test nobody can see failing is worse than a test that does not exist, because
its name is a correct bug report that everyone reads as reassurance.

Whole suite now runs in one command: 274 passed, 0 failed.

Two bugs that were hiding each other, which is why both survived a month.

THE SUITE RAN ZERO TESTS. test_setlist.py was a plain script ending in a bare
module-level sys.exit(), and pytest raises that during COLLECTION — so
'pytest tools/tests/' aborted the whole directory with an INTERNALERROR and ran
nothing, while 273 tests passed when invoked file-by-file. Every test in this
repo was invisible to the obvious command. Guarded the exit under __main__ (the
script still works run directly) and added one assertion so pytest actually sees
the checks. Whole suite in one command now: 274 passed, 2 skipped, 0 failed.

AND THE THING IT WAS HIDING: test_setlist had been failing all along, on a check
whose name says exactly what is wrong — 'a deliberation note is not a track (the
quand_on_decolle phantom)'. OPAL 2026's backlog section opens with

    ## SOUNDCHECK
    -- Quand on decolle

and entries() takes every list item under the gig heading, so every one of the
tools reading this setlist believed the set had 16 tracks opening on Quand on
décolle. The recorded set has 15 and opens on Ceci n'est pas Une Bombe
(segments_v3/v4 track 1). The parser now yields 15, in the right order.

The fix is deliberately NOT the section whitelist entries() refuses to build —
that one tried to enumerate which of PLN's prose headings are part of the set and
died on '## Livecoding Techno DNB Nu-jazz'. This is the inverse and far narrower:
naming the one phase that is definitionally not a performance. Three patterns,
and skips print to stderr, because a dropped track and a skipped soundcheck must
never look alike.

Fixed the parser, never backlog.md — the backlog is PLN's own file and it was
right. A soundcheck IS in his set list; it just isn't in his set.

Verified by breaking it: neutered _NOT_THE_SET_RE, watched
test_backlog_setlist_checks_all_pass fail naming the quand_on_decolle phantom,
restored, 274 green.

    the first track was never the first track.
    sixteen tools agreed with each other
    and none of them agreed with the recording.

Both were pre-existing, unrelated to today's check-mix.py/gig-log.py work
(spotted while running the suite for that change).

test_orphan_orbits.py: 9 of 41 tests failed (16-row setlist vs a frozen "13",
and 6 HAND_MEASURED orbit sets vs the parser). Root cause confirmed by git log
+ grep, not a parser bug: the setlist grew from 13 to 16 tracks since the
fixture was written, and commit 2376e431 (2026-08-01, "d10 is the riser") added
a d10 safe-riser idiom to 5 of the hand-measured tracks (bombe_dj, wap,
you_my_sunshine, mafia_sans_serif, desire) after the 2026-07-28 hand
measurement date; piment_bresilien went the other way (d10->d9, PLN's own
request, commit 71bb9bc2/5e5a37ad). Fix: the setlist-length assertions are now
derived from an independent sed-based row count of the setlist file (same
pipeline check-tracks.sh uses) instead of a literal that rots every time PLN
adds a track; the HAND_MEASURED dict and the vague_de_crime->bombe_dj orphan
assertion are re-measured by hand (grep on each file's dN lines, not by
copying the parser's own output) against the corpus as it stands today. No
assertion loosened — 41/41 pass, same invariants, current facts.

test_lcxl3_display.py: pytest collected 0 items. Cause: this was written as a
plain script (module-level `check()` calls, run via `python3
tools/tests/test_lcxl3_display.py`), like test_setlist.py, with no `test_*`
function pytest could find — a file that cannot fail under pytest, i.e.
silent coverage loss. Converted to 30 ordinary pytest test_* functions (one
assert group per logical check), same assertions, nothing dropped or
loosened. Verified live: broke rec_line's hour rollover (divmod 3600->3601)
and watched test_rec_line_formats_hmmss_past_an_hour fail with the exact
wrong value (REC 1:02:02 vs 1:02:03), then reverted.

Suite: 234 passed/9 failed/2 skipped -> 273 passed/2 skipped (ignoring
test_setlist.py, a separate plain-script file, out of scope here — it already
breaks whole-directory pytest collection via a module-level sys.exit and
predates this fix). test_gig_log.py untouched: 169 passed before and after.

Ran the new generation lens proactively against Cosmic26_master, before that
release goes anywhere: 12 OK, zero stale, zero strays. The exact opposite of
OPAL, for two reasons that are lessons rather than luck — only one render
generation exists on disk, and every permalink is gig-scoped via
--permalink-suffix, so no CosmicFest upload could collide with the OPAL cut of
the same score. Four of these titles exist at both gigs.

That answers an open board decision with evidence instead of taste. It asked
whether to keep OPAL unsuffixed for consistency; the suffixed release has 0
permalink pathologies and the unsuffixed one has 14. Consistency with a broken
scheme is not a virtue, so new OPAL uploads get suffixed. The 6 correct
unsuffixed tracks stay — they are right, and renaming spends a permalink for
nothing.

Found in passing: CosmicFest is missing exactly 2 tracks, and they are almost
certainly the two already-known vanishers that logged '✓ live at', passed an API
check, then 404'd. A plain idempotent re-run converges. Which is the standing
lesson landing on its feet: verify the shelf AFTER a run, not only during it —
'N uploaded, 0 failed' was never evidence that N tracks exist, and it took a
lens three weeks later to notice two were gone.

check-mix reports OK, all 12 orbits reach master. Third recurrence of the
saved--inf bug, closed by PLN's hands.

Two things fell out of verifying it. check-mix resolves the saved session with
Ardour CLOSED — it reads the file, not the process, so the fader check can run
in a pre-gig script before anything is launched. And the diff is bigger than the
fix: orbits 02-06 were in a coherent -8 to -10 dB trim set and are now all near
unity, so nine faders moved where four were broken. Master is still -0.2 so
nothing clips, but the balance between orbits shifted ~9 dB on five of them.

Recorded as a QUESTION, not a finding. The LCXL3's faders are absolute-position
controls and touching one snaps the DAW value to the pot — which also means the
board's 'v3 retires pot-pickup' claim covers the endless ENCODERS only; the
eight faders are still absolute and that whole class survives the upgrade. But
PLN may simply have dragged them on purpose, and a cause nobody asked about is
not a cause.

Found while shutting down: unplugging the LCXL took midiviz's ALSA
source with it, the process exited cleanly, and Restart=on-failure did
not fire. The permanent lens is only permanent while the board stays
plugged -- which is the same resolve-once stale-binding shape as the
zombie driver this session was opened to fix. lcxl-leds.py already has
the fix pattern (re-resolve per send, cache invalidated on failure).

The rewire (tray Launch Gig -> POST /api/rig, arm-then-confirm) is
written and loads clean, but testing it end-to-end killed Ardour:
launchers.launch('ardour') reported 'launched' while Ardour was already
running, a second instance collided with the first over the session
lock, and both exited -- taking all 12 orbit links with them.

A false negative in is_running() wired to a tray button is worse than
the mislabeled button it replaces, so it does not ship until launch()
fails closed. Parked as a patch beside the note rather than committed
(unvalidated) or left loose in the worktree (this checkout is shared and
another session commits into it by pathspec).

Carries the rest of the session's open ends too: the perf.sh redeploy
that is committed but not live, gig-up's four pre-existing NO-GO
blockers, perf-tray's 2.6GB RSS, and why gig-down is blocked on the
Freebox being down.

Captain's log for the gear-failure pass: one zombie lcxl3-driver holding a
stale binding to an unplugged mk3 explained three of PLN's five symptoms,
the fourth was a mislabeled tray button, and the fifth was tooling that
had been running invisibly all along.

Written for mining later — the sharp bit is that name-based identity is
the actual bug, and it caught the investigation as well as the rig: a
subagent grepped aconnect for LCXL3, hit the driver's own phantom virtual
port, and reported the hardware was back.

TASKS_DUMP has accreted amendments since 2026-08-16 and the oldest sections are
now the least true — but a reader arrives at the top, where A3 still says
/take-five-drops uploaded '24s short', that every other track 'matched to the
second', that 8 of 15 are up, and that no artwork exists. All four are wrong,
and the first two are the exact reasoning that let six stale uploads pass a
duration check in August.

So: a READ THIS FIRST banner naming what is superseded and sending the reader to
the resume points at the end, plus an inline stop sign at A3's 'TWO DEFECTS'
where someone reading that epic top-down actually hits it. The wrong diagnosis
is kept rather than deleted, because the reasoning being auditable is the point —
it just no longer gets to be the first thing anyone reads.

A board whose front page is a year of stale confidence is worse than no board.
Nobody rewrites 1700 lines; they act on line 90.

lcxl3-driver was policy "login", so rig_units.ensure() started it
whenever it was not already active -- and ensure() runs inside
gig-up.sh --converge, which is what the Bridge's RIG UP calls.

With the CLASSIC LCXL plugged that is actively destructive: the v3
driver has Conflicts=lcxl-leds-watch, so a converge would stop the
painter that was correctly lighting the classic board, then fail to bind
(there is no LCXL3 port) and crash-loop at RestartSec=5. That is exactly
what happened between 11:26 and 14:36 on 2026-09-05 -- 86 restarts, the
board dark the whole time -- and today's repair would have been undone
by the next press of GIG UP.

lcxl-leds-watch was already "manual" for the mirror-image reason, stated
in the comment right below this one. The rule generalises: neither
painter may be auto-started blind, because "which painter" is a question
only the plugged hardware can answer. gig-up.sh's leds() is the
generation-aware chooser -- it greps aconnect for LCXL3 and restarts the
right unit -- and it now owns the decision outright.

Verified: `rig_units.py --ensure` (dry) reports "nothing to do" with the
driver inactive and the v2 painter running, instead of queueing a start
that would have gone dark.

Trade-off worth knowing: on a mk3 rig the driver no longer comes up at
login, so faders reach Ardour only after gig-up runs. That is the price
of not guessing the generation, and gig-up is the documented bring-up.

Structured archive entry for the session, written as standalone learning rather
than a board line. Fourteen findings, of which the four that will outlive the
task: SoundCloud accumulates generations and never replaces (so a stale upload
is permanent until deleted); the wrong cut wore the cover art, which is why
PLN's ear beat every tool to it; the 'matches to the second' tolerance was the
bug, because full_duration is millisecond-exact and a loose window made two
different cuts identical; and roll call cannot find a stray, because a check
that only looks where it expects a thing never discovers it elsewhere.

Written for a reader months out with no memory of any of it.

Captain's log 040 and the pre-compact resume state.

The log's finding, in one line: OPAL-26 sounded broken on SoundCloud while every
rendered file was correct to the millisecond, because SoundCloud will not replace
audio behind a permalink and our uploader is idempotent — so every re-upload
under a fresh slug ADDED a copy. The whole Aug-12 first cut was still up beside
the six correct v4 ones. And the tell PLN was actually reacting to: eight of the
fourteen stale uploads wear cover art and not one of the correct six does. The
superseded audio looked finished. His ear was reading the artwork.

Two lens lessons in it, both recurring. 'Matches to the second' was a tolerance,
not a verdict — SoundCloud reproduces full_duration to the MILLISECOND, so at the
±1.5s a transcode seems to deserve, the Aug-12 and v4 cuts of /wap are the same
track, which is how six stale uploads passed a duration check in August. And roll
call cannot find a stray, because a check that only looks where it expects a
thing can never find that thing somewhere else.

Plus the git one, which cost four commits: this is a shared checkout, another
session moved HEAD between two of my commits, and 'git push origin master' from a
non-master HEAD is a silent no-op that exits 0. The PUSHED in my own log was
something I wrote myself.

The board now carries R1-R5 written for a reader with zero context: exact
commands with paths, the artifacts already on disk so nothing is regenerated,
the three decisions that are PLN's and not a machine's, and the shared-checkout
hazard spelled out before anyone commits into it again.

183 Pulsar SIGABRTs between 2026-08-19 and 2026-09-05, 80 of them inside
one 70-minute window, all from this one block.

perf.sh launched Pulsar in "the script's directory",
$(dirname "${BASH_SOURCE[0]}") -- but perf.sh is deployed root-owned as
/usr/local/sbin/perf-audio (perf-audio.sudoers, for scoped NOPASSWD), so
BASH_SOURCE resolved to /usr/local/sbin and Pulsar got /usr/local/sbin as
its project directory. Coredump argv:
  /opt/Pulsar/pulsar --executed-from=/home/pln --no-sandbox /usr/local/sbin

The Bridge's perf watcher (bridge.py:41 -> perf.py:run_watcher ->
reconcile -> set_mode) shells out here every ~30s, and every mode flag
routes through set_priorities, so the retry was unbounded: crash ->
pgrep finds nothing -> next tick tries again. No backoff, no lock, no
exit-status check. And the "✓ Pulsar launched" line proved nothing: the
verification was a 0.5s pgrep presence poll that caught the doomed
Electron process alive during startup and returned before it aborted.

Fixed by removing the launch, not by resolving the directory better.
Setting priorities is this script's job; starting applications is not.
Ardour, two branches up, has always been prioritize-if-present with no
launch branch -- Pulsar is now symmetric with it. Both the tray and the
Bridge have real Pulsar launchers that pass the right project directory
and surface failures. #116 also wants the hot path unable to start
things by accident, and a root script spawning a GUI editor through
`sudo -u $USER env ...` every 30s is the opposite of that.

NOT LIVE YET: /usr/local/sbin/perf-audio is a root-owned copy and still
carries the bug. It needs
  sudo install -m755 -o root -g root perf.sh /usr/local/sbin/perf-audio
which is outside the sudoers whitelist, so PLN has to run it. Dormant
until then -- the block only fires when no Pulsar is running.

The argv-only bug is fixed by extracting the 3-rung ladder into
tools/ardour_session.py, imported by both gig-log and check-mix. gig-log's
'keep check-mix in sync' comment is gone because there is nothing left to keep
in sync — one parser per concept satisfied by construction rather than by
somebody remembering. 169 gig-log tests pass unchanged.

Then it ran for real and reported Tidal 07, 08, 10 and 12 at -inf dB in the
saved session. Third recurrence of that bug (04-08+12, then 02+10, now these
four). PLN's parked two-minute check went from unrunnable straight to failing,
which is the whole argument for wiring a check into a caller: it gets run under
conditions the unit test never sees.

Also logged: two pre-existing test failures the suite run made visible
(test_orphan_orbits 16 rows vs 13 expected; test_lcxl3_display collects zero
tests, and a test file that cannot fail is not a passing test), and the
uncommitted duplicates left sitting on claude/rig-streamline so whoever rebases
it does not double-apply them.

    the fader was down in the file, not in the room.
    nobody could hear the difference
    because a closed fader and a silent instrument
    look identical on a post-fader meter.

PLN, after the first look at it: "the midi mon is beautiful! so lets
streamline, now it should always oppen, with always on top, all desktops
displayed when i switch from desk 1 to 2 3 4" -- and then "would help to
see the chan num: 33 or 53 etc so as i press i can wire".

Always-on-top needed no change: the WindowStaysOnTopHint at :293 is
unconditional and KWin honours it. The other three did.

IDENTITY. The Wayland app_id comes from desktopFileName, not
applicationName, so KWin saw resourceClass "python3" -- verified by
querying workspace.windowList(), not assumed. Two consequences: a window
rule keyed on the app id would have matched EVERY python GUI on the box,
and launchers.focus_window("midimon") could never match, so clicking the
Bridge/tray button on an already-running lens silently did nothing. Set
desktopFileName + ship a .desktop (which also silences the portal's "App
info not found" warning), and point the launcher's focus at "midiviz".

ALL DESKTOPS is a compositor property, not a Qt flag, so it lives in a
KWin rule (~/.config/kwinrulesrc, wmclass=midiviz, desktops= empty with
desktopsrule=2/Force). Declarative on purpose: it applies at map time,
every time, with no post-launch retry racing the window's existence.
Verified on the live window rather than inferred from the key names --
onAllDesktops=true keepAbove=true.

CC NUMBERS, two layers. Per-cell dim decimals are the map you consult
while wiring; they render in their own pass because f_micro is already
current there, and a per-cell setFont would add 96 font switches a frame
to a path this file keeps to a counted number of drawText calls. They
are skipped when a cell is too narrow to hold them beside the value. And
the header now prints the last-touched control as ^53=127 -- decimal,
with the corpus's caret, so it is the literal token to type into a
pattern instead of a hex value to convert mid-set.

Autostart as a systemd --user unit bound to graphical-session.target,
not default.target: it needs WAYLAND_DISPLAY/XDG_RUNTIME_DIR imported
into the user manager first. Registered "login" in rig_units.py.

Paint path exercised by the offscreen selftest: 116 paints, PASS.

check-mix.py resolved the live session from argv only, so any Ardour
launched from the GUI's recent-sessions list (bare argv) went unresolved
and check-mix silently fell back to a guessed path. gig-log.py's FilesLens
already solved this with a 3-rung evidence ladder (argv > cwd > open fds,
commit abad7515) and had a sync comment flagging check-mix as needing it.

Extracted the ladder into tools/ardour_session.py — one resolver instead
of two, per feedback_one_parser_per_concept — and made both gig-log.py and
check-mix.py import it. gig-log's FilesLens keeps its instance-level
caching/logging but delegates all evidence-gathering to the shared module;
the now-unnecessary "keep in sync" comment is gone.

Verified: gig-log's 169 pytest tests still pass unchanged (they exercise
all 3 rungs against synthetic /proc fixtures). Live: Ardour was running
("Tidal Live" session, launched with argv), so check-mix.py resolved it
for real and reported per-orbit gains — surfacing the same recurring bug
this fix was meant to catch: Tidal 07/08/10/12 sit at -inf dB in the
currently SAVED session. Rungs 2 (cwd) and 3 (open-fds) were proven against
synthetic fixtures matching the live-rig gap, since the running process's
argv already carried the session path.

master sc delete --manifest exists: dry-run by default, identity verified in
BOTH directions (a permalink that now resolves to a different sc_id is a
refusal, not a warning), public tracks refused, a reason required per entry
because a deletion without provenance is not allowed, batch capped before any
API call, and a receipt that keeps every field plus the on-disk paths of the
archived artwork and the source FLAC. The old test-namespace gate was not
widened — a second gate was added beside it, and a test proves the old one is
unreachable from the new path.

Dry run: 14 deletable, 0 refused. All private, all uploaded in the same minute
on 2026-08-12, and every one still has its FLAC in Opal26_master/tracks/, so
deleting them loses a URL and not a recording. --go is NOT approved yet.

Two caveats recorded loudly rather than buried. The delete TRANSPORT is
unproven: no DELETE has ever been issued from this codebase against api-v2, and
a clean dry run is not evidence the deletes will land. And the generator's first
version lined stale uploads against the newest cut BY TRACK NUMBER — dropping
Desire renumbered REVOLUTION from 15 to 14, so the manifest confidently compared
Desire against REVOLUTION's duration. An unnamed reference, again.

Also logged: four of my own commits landed on a concurrent session's branch
because the shared checkout moved under me, and 'git push origin master' from a
non-master HEAD is a silent no-op that exits 0. Recovered via worktree
cherry-pick without touching their branch or the live tree. The worktree rule in
CLAUDE.md has no orchestrator exemption, which is what I had assumed.

    two sessions, one working tree,
    and a push that said yes to a question
    it had never actually been asked.

The regen moved timecodes by up to 21s (Piment +21.0, REVOLUTION +19.0, Gimme
Acid +16.7). That file is not just website furniture — Slopmotion cuts visuals
to it, so the release videos were pointed a third of a minute off the music on
a third of the record.

The more interesting finding is a parser conflict the regen walked into.
build_gig_tracksjson.py drafts NEW gig pages and never computes the
styleDistribution that LiveEvent.js renders; gig-metadata.py regenerates
existing ones and does. They also disagree outright on REVOLUTION's style
(jazz vs nujazz). Two parsers, one concept — so the board's line that
build_gig_tracksjson 'is now the way to make any gig page' is too strong and
is corrected here rather than left to bite the next regen.

Left explicitly to PLN: whether the OPAL page documents the set as PERFORMED
(15 tracks, Desire included) or as RELEASED (14). Desire has no v4 boundary
because it was cut, so its v3 timecode rides through unverified. A page can
honestly be either thing; only he knows which one he wants a visitor to get.

The first version of this amendment said one OPAL track on SoundCloud was
stale. Then I read the account instead of reasoning about it, and it is
fourteen: the entire Aug-12 first cut is up, thirteen of it under junk
permalinks (/opal26-wap, /12-mafia-12, /ceci-nest-pas-une-bombe) that roll
call could never have found, because roll call only looks where it expects a
track to be. Same error the board made before me, one size smaller.

The finding that explains PLN's ear exactly: eight of the fourteen stale
uploads carry cover art and not one of the six correct v4 tracks does. The
wrong cut is the one that looks finished.

Also corrected: 'every other track matches to the second' was never a verdict,
it was a tolerance. SoundCloud reproduces full_duration to the MILLISECOND, so
at the ±1.5s a transcode seems to deserve, the Aug-12 and v4 cuts of /wap
(361.0 vs 361.6) and /perfect (315.0 vs 314.6) are the same track. That is how
six stale uploads passed a duration check in August.

And two coincidences caught before they could become deletions: /saria-demo
(2025) and /au-revoir-lord-toyota (2020) match OPAL to 50ms by chance on a
132-track account. Duration is a strong discriminator and still not provenance.

    the ones with the cover art were the wrong ones.
    six months of a set played once, and the tell
    was never in the sound -- it was in which
    of the twenty ghosts looked ready.

CLAUDE.md pointed at ../../Web/www/next/content/lives/, which has not existed
since www hoisted the app out of next/ in 6e51fff. The correct path is
../../Web/www/content/lives/{year}/{slug}.md.

This is a worse bug than a typo in docs. The instruction it sits inside is
'Mastering != metadata. Never invent gig metadata... Verify there and cite the
path — if a fact isn't in the canonical source, ask.' A reader who follows the
documented path gets a false 'not found' from the one source they were told to
trust, and the honest next step from there — asking, or worse, filling the gap
from an intermediate script — is exactly the failure the rule was written to
prevent. A stale read is how a Write becomes a delete.

The tide-table scripts already had it right (build_catalog_view.py's LIVES
carries a comment naming the hoist commit); only the instruction was behind.

- process-palette.json wires the editor's command palette to the rig's own
  entrypoints (Init MIDI Channels -> init_midi.sh, and the rest). The whole
  'one entrypoint, fewer files' goal is only real if the editor knows where
  the entrypoints are, so this is rig config, not editor preference.
- sketch.properties points the Processing sketch at its main .pde.
- Logo.xcf is the editable source of the ParVagues mark, from 2025-05. The
  armada design system is documented down to its type stack and its reserved
  brand magenta, and the one file you cannot regenerate from any of that is
  the layered original. 1.2 MB to never have to redraw it.

Deliberately still untracked: TODO.md. It is a paused 2026-06-22 perf/thermal
board, and TASKS_DUMP.md has been the board since 2026-08-16. Committing it
would put a second answer to 'what is next' in the tree, which is the shape
that keeps biting us. PLN's call: fold its two live items into TASKS_DUMP and
delete it, or keep it as a scratch file outside git.

`sc_generation_lens` already knows which OPAL-26 uploads are the Aug-12 first
cut; nothing turned that verdict into something the new `sc delete --manifest`
could execute. This does, and it is the last checkpoint before a destructive
call, so it refuses rather than warns.

In: verdicts starting STALE/ORPHAN (right permalink, wrong cut) plus every
stray (an older cut under another slug). Out, and asserted out: verdict OK
(that IS the ear-verified cut), NOT UPLOADED (nothing to delete), anything
without an sc_id, and everything in `coincidences_discarded` — the lens finds
strays by duration and on a 132-track account two unrelated tracks will share
a length to 50 ms, which is how /au-revoir-lord-toyota (2020, and PUBLIC) and
/saria-demo (2025) both matched an OPAL render. The date floor is re-applied
here instead of trusting the file's own bucket names: two chances to catch a
mistake whose cost is permanent.

Every entry carries a reason citing the evidence, the artwork the pull saved,
and the local render the upload came from — all fourteen resolve to a FLAC in
Opal26_master/tracks/, so deleting these loses a URL and not a recording.

One bug found while writing it, and it is the reason the reason strings matter:
the first version lined a stale upload up against the newest cut BY TRACK
NUMBER, and the Aug-12 cut had fifteen tracks where v4 has fourteen. Dropping
Desire renumbered REVOLUTION from 15 to 14, so the manifest claimed
"#14 Desire, 388.1s vs v4 205.6s" — 205.6s is REVOLUTION's length. A
measurement is a comparison; an unnamed reference cites the wrong track. The
lookup is keyed on the title slug now, and a track absent from the newest cut
says so instead of borrowing a neighbour's duration.

Dry-run against the live account: 14/14 resolved, both directions, all private,
nothing refused. Nothing deleted — that is PLN's call.

The regen moved timecodes by up to 21s (Piment +21.0, REVOLUTION +19.0, Gimme
Acid +16.7). That file is not just website furniture — Slopmotion cuts visuals
to it, so the release videos were pointed a third of a minute off the music on
a third of the record.

The more interesting finding is a parser conflict the regen walked into.
build_gig_tracksjson.py drafts NEW gig pages and never computes the
styleDistribution that LiveEvent.js renders; gig-metadata.py regenerates
existing ones and does. They also disagree outright on REVOLUTION's style
(jazz vs nujazz). Two parsers, one concept — so the board's line that
build_gig_tracksjson 'is now the way to make any gig page' is too strong and
is corrected here rather than left to bite the next regen.

Left explicitly to PLN: whether the OPAL page documents the set as PERFORMED
(15 tracks, Desire included) or as RELEASED (14). Desire has no v4 boundary
because it was cut, so its v3 timecode rides through unverified. A page can
honestly be either thing; only he knows which one he wants a visitor to get.

SuperDirt's boot runs MIDIIn.connectAll (start_and_midi.scd:19), so SC
subscribes to every hardware MIDI input it can see -- the LCXL included.
The surface then fed SC twice: once raw, once through Midi Through. And
MIDIFunc.cc takes `src` and never checks it (start_and_midi.scd:51), so
every CC fired the handler twice and every button noteOn double-fired
into ~lcxlChordCheck. A double noteOn on a toggle is a no-op: that is how
a mute button silently stops working.

lcxl3-driver's prune() handled this for the v3 board, but only while it
runs -- and it was stopped today (stale binding to an unplugged mk3,
Conflicts=-killing the mk1 painter). A hand-run `aconnect -d` fixed it
live and then SC restarted and re-armed it, which is the whole argument
for putting it in the reconcile loop that already owns the wiring.

Rule 3 matches the RAW board on type=kernel, never on name alone:
lcxl3-driver publishes a virtual port called 'ParVagues LCXL3' that
matches the same name candidates, and cutting the driver's translated
feed would kill the path we actually want. Only surface->SuperCollider
links are cut; -> Midi Through and the aseqdump taps that gig-log and the
Pulsar HUD read all survive, verified live.

Tested by arming the bug (aconnect 20:0 130:3, 20:1 130:4) and watching
the loop cut both and settle: lcxl-path.py went from "LCXL -> SC direct"
back to "LCXL -> Midi Through -> SC", logged once, no per-tick spam, and
gig-log kept counting touched controls.

Also logs PLN's ear-feedback on the classic-LCXL play-test (rose_rouge
works great; midiviz landed) to the archivist's notes.

The first version of this amendment said one OPAL track on SoundCloud was
stale. Then I read the account instead of reasoning about it, and it is
fourteen: the entire Aug-12 first cut is up, thirteen of it under junk
permalinks (/opal26-wap, /12-mafia-12, /ceci-nest-pas-une-bombe) that roll
call could never have found, because roll call only looks where it expects a
track to be. Same error the board made before me, one size smaller.

The finding that explains PLN's ear exactly: eight of the fourteen stale
uploads carry cover art and not one of the six correct v4 tracks does. The
wrong cut is the one that looks finished.

Also corrected: 'every other track matches to the second' was never a verdict,
it was a tolerance. SoundCloud reproduces full_duration to the MILLISECOND, so
at the ±1.5s a transcode seems to deserve, the Aug-12 and v4 cuts of /wap
(361.0 vs 361.6) and /perfect (315.0 vs 314.6) are the same track. That is how
six stale uploads passed a duration check in August.

And two coincidences caught before they could become deletions: /saria-demo
(2025) and /au-revoir-lord-toyota (2020) match OPAL to 50ms by chance on a
132-track account. Duration is a strong discriminator and still not provenance.

    the ones with the cover art were the wrong ones.
    six months of a set played once, and the tell
    was never in the sound -- it was in which
    of the twenty ghosts looked ready.

CLAUDE.md pointed at ../../Web/www/next/content/lives/, which has not existed
since www hoisted the app out of next/ in 6e51fff. The correct path is
../../Web/www/content/lives/{year}/{slug}.md.

This is a worse bug than a typo in docs. The instruction it sits inside is
'Mastering != metadata. Never invent gig metadata... Verify there and cite the
path — if a fact isn't in the canonical source, ask.' A reader who follows the
documented path gets a false 'not found' from the one source they were told to
trust, and the honest next step from there — asking, or worse, filling the gap
from an intermediate script — is exactly the failure the rule was written to
prevent. A stale read is how a Write becomes a delete.

The tide-table scripts already had it right (build_catalog_view.py's LIVES
carries a comment naming the hoist commit); only the instruction was behind.

- process-palette.json wires the editor's command palette to the rig's own
  entrypoints (Init MIDI Channels -> init_midi.sh, and the rest). The whole
  'one entrypoint, fewer files' goal is only real if the editor knows where
  the entrypoints are, so this is rig config, not editor preference.
- sketch.properties points the Processing sketch at its main .pde.
- Logo.xcf is the editable source of the ParVagues mark, from 2025-05. The
  armada design system is documented down to its type stack and its reserved
  brand magenta, and the one file you cannot regenerate from any of that is
  the layered original. 1.2 MB to never have to redraw it.

Deliberately still untracked: TODO.md. It is a paused 2026-06-22 perf/thermal
board, and TASKS_DUMP.md has been the board since 2026-08-16. Committing it
would put a second answer to 'what is next' in the tree, which is the shape
that keeps biting us. PLN's call: fold its two live items into TASKS_DUMP and
delete it, or keep it as a scratch file outside git.

Written earlier today by the gear-repair session and left uncommitted, which
for this particular file is the one failure mode that matters: performance_notes
IS the archive, so an uncommitted entry is not a note, it is a memory.

Two signals, both about the rig rather than the audio:

- *"ok just played rose_rouge works great!"* — first confirmation that the
  CLASSIC LCXL (mk1, 1235:0061) is a fully viable performing surface once
  lcxl-leds-watch owns the paint. Worth recording precisely because the rig had
  drifted into treating the mk3 as the only real board; the classic is a
  fallback, not a downgrade.
- *"and the midi mon is beautiful!"* — on first sight of tools/bridge/midiviz.py.
  A taste signal about the rig's own UI, and the follow-up ask (always-open,
  always-on-top, every desktop) says it plainly: midiviz is an instrument of
  the rig, not a debug tool.

CLAUDE.md asks for every reaction captured durably because this corpus feeds
sampling, effect choices and the auto-mastering heuristics. Durably means
committed.

Four kinds of orphan, all real source, none of it generated:

- tools/sweep_state.py — 'the surface needs a sweep', as a fact the rig can
  act on. Companion to the LCXL3 driver thread.
- armada/tide-table/kick_bass.py — the deterministic kick-vs-bass tiebreaker,
  the one split CLAP cannot do reliably because both live low. Exactly the
  'validate by analysis, never infer role from the name' rule in code.
- armada/tide-table/calibration.stem.json — joins calibration.json and
  calibration.whole-mix.json, which were already tracked. A calibration set
  missing one of its three members is a trap for whoever reads the other two.
- install_superdirt.sclang — pins SuperDirt v1.7.2 and recompiles. The boot
  path is exactly where 'simple, non-brittle, one entrypoint' matters.
- live/collab/benoit/tous_ensemble.tidal and live/midi/nova/nujazz/
  nouvelle_couleur.tidal — two scores. nouvelle_couleur carries a '-- url:'
  header naming its backdrop source, per the scene standard.
- visuals/{maze,nightly_repairs,there,there0}.js — four hydra sketches.

Python entrypoints staged 100755 via --chmod, not left to the working tree.

    the drawer is empty now.
    everything in it was something,
    which is why it took three weeks to look.

Ten shell tools plus FORMAT.md and MIX.md, written 2026-07-29 and untracked
since: stem-stats, energy-profile, onset-detect, silence-detect,
transition-detect, mix-meta, mix-render, mix-normalize, mix-master, lufs-check,
and lib/progress.sh. FORMAT.md defines a <project>.mix.json metadata contract
whose tools 'read existing metadata and augment (don't overwrite unless told)'.

Committed with the exec bit set IN THE INDEX, which is the part that bites:
3fdf5b97 had to restore it on 63 scripts, and MIX.md invokes every one of these
as ./stem-stats. A file that is executable in the working tree and 100644 in
git is executable exactly until someone else clones it. Note also that
'git commit -- path' silently drops a staged mode flip (the trap 1667bb02
documented), so this went in without a pathspec after proving nothing else
was staged.

mix/take94's 833 MB stayed out (a5a5b37). The scripts are the artifact; the
take is the input.

⚠️ Flagged, not resolved: this overlaps ../tidal-ears' mastering CLI, which is
the canonical recipe per MASTERING.md. Two toolboxes that both master a set is
the shape feedback_one_parser_per_concept warns about. I am not merging them
blind — committing first so the overlap is visible in one tree instead of
half-invisible in an untracked directory. PLN's call which one is the spine.

    ten small knives in a drawer nobody opened,
    sharp the whole time.

Written 2026-08-14, uncommitted for three weeks: the -inf fader forensics on
Tidal Live.ardour, the corpus triage that found '51 broken tracks are actually
22 — the harness was the biggest bug', and the 192 kHz master-format audit
('probe, never assume'). All three are marked shareable: true, which is the
whole point — these logs are the source material for the posts and videos,
and they were sitting one bad rm away from not existing.

The rule in armada/tasks/README.md is 'write the entry when the task lands,
while it's fresh'. It worked: the writing happened same-day. It was the
committing that didn't, because git status had become unreadable (see a5a5b37).
Documentation discipline is not one habit, it is two, and the second one is
the one that survives a hard drive.

Left for PLN, one line: all three carry 'log: 025' in their frontmatter, which
025-fifteen-intros-are-not-a-set.md already owns. They also use dated filenames
like 2026-08-03-j5-audio-dropout.md rather than the README's NNN-slug.md.
Renumbering is a call about his own archive's spine, not mine — and the dates
are 2026-08-14, which lands them BEFORE 039, so the next free numbers would
read out of order.

    the log was kept. the log was not carried.
    a ship's record that never leaves the desk
    is a memory, not a record.

Three weeks of untracked working tree, and the reason nothing was committed is
that 'untracked' had stopped meaning anything: 957 MB of take audio and scene
mp4s sat in the same status output as three finished captain's logs, so the
whole list read as noise and got skipped as one.

Sorted by what each thing actually is, against the precedent already in this
file and in the index: mix/'s ten scripts are source and go in, mix/take94's
833 MB does not. The .js scene sketches are source, the 124 MB of rendered
mp4 is not. calibration.json and calibration.whole-mix.json are tracked, so
calibration.stem.json belongs with them; bounds_*.json is generated output
whose builder is already tracked, so it does not.

No new policy here — just applying the one written at the top of this file to
a pile that had grown too mixed to read.

    the tide went out three weeks ago
    and left everything on the sand together:
    charts, driftwood, a gigabyte of water.
    sorting is not cleaning. sorting is deciding what was ever yours.

PLN opened the day hearing it: 'our private cuts of OPAL are badly cut
begin/end unaligned'. The board had one theory (a 24s truncation) and it was
wrong. This lens found the real shape: the ENTIRE Aug-12 first cut is still
on SoundCloud — all 15 tracks — living beside the 6 correct v4 ones under a
litter of ad-hoc permalinks (/opal26-wap, /12-mafia-12, /ceci-nest-pas-une-
bombe), and eight of the stale ones are the ones wearing artwork. So the
wrong cut is the one that looks finished. Of course he heard it.

Duration is the discriminator, and it is far sharper than assumed: measured
on this account, SoundCloud's full_duration reproduces the source to the
millisecond (477.600 uploaded, 477.600 reported). Tolerance is 50ms, not the
seconds you would allow for a transcode — a loose tolerance here buys no
safety, it only invents ambiguity between cuts that differ plainly.

Three things I got wrong first, each left in as a guard:
  - filenames are not authoritative. 'Perfect <3' uploads to /perfect-3, so a
    filename-derived slug reported two live tracks as NOT UPLOADED. Titles
    come from the plan now, which is what the uploader actually used.
  - streaming and club masters share every duration, so offering both as
    rival candidates marked all 14 AMBIGUOUS. Identical generations collapse.
  - a duration collision is not provenance. /au-revoir-lord-toyota (2020) and
    /saria-demo (2025) both matched OPAL to 50ms. Anything uploaded before
    the oldest render on disk cannot be a cut of it, and a coincidence must
    never reach a delete manifest — so the date floor is a hard gate and the
    discards are printed, not silently dropped.

The second pass is the one that earned its keep. Roll call only looks where
it expects a track to be, so it can never see a stale cut hiding under a
different slug. Strays scans the whole account. Absence needs proof.

    fifteen ghosts of a set played once,
    each wearing the cover art of the real thing --
    the lens does not listen, it only counts seconds,
    and seconds were enough.

catalog.generated.json was stale since 2026-06-23 (73 tracks / 23 gigs) and
missing opal-festival-2026 + cosmicfest-2026 entirely — it never got rebuilt
after catalog_view.json picked those gigs up on 2026-09-02. Re-ran the
pipeline (build_catalog_view.py then build_catalog.py):

  catalog_view.json:      81 tracks / 25 gigs / 63 takes (unchanged — was
                           already current, only as_of + minor EDA-coverage
                           deltas moved)
  catalog.generated.json: 73→81 tracks, 23→25 gigs, now carries both missing
                           2026 gigs, both validate clean against models.Catalog,
                           and track-id sets match view↔generated exactly
                           (no silent drops either direction)

Root cause of the as_of staleness: build_catalog_view.py hardcoded
`AS_OF = "2026-06-06"` as a literal instead of stamping the actual build
date, so every regeneration since kept reporting the same frozen date no
matter when it ran. Now computed as `date.today().isoformat()`. Checked the
other suspected bug (a next/content/lives hardcoded path) — not present;
LIVES already points at .../www/content/lives without next/ since 6e51fff.

PLN opened the session hearing badly-cut begin/end on the private OPAL
uploads. Measured all three render generations against what the board
recorded on the account: /take-five-drops sits at 6:57 = 417 s, which is
exactly the Aug-12 first cut — four days and two boundary revisions before
the ear-verified v4. The complaint had a mechanical cause.

Ruled out before believing anything: tracks_v4_streaming matches
segments_v4.json to the millisecond on all 14 (CRIME's 2s rest and
REVOLUTION's reverb tail included), the upload plan's titles/files/durations
are aligned and exact, and the ear file's 19:33 mtime is the signoff record
being written, not a later boundary edit. Nothing needs re-cutting; the
approved audio is already on disk. This is a re-upload problem.

Two things stand in the way and one of them is PLN's call: SCW.delete() is
hard-gated to the test namespace so we cannot remove our own stale uploads,
and SoundCloud skips rather than replaces an existing permalink. Also: no
sc_state.json was ever written, so the account is the only record of what
landed — sc pull --artwork recovers ids, permalinks, durations AND the covers
PLN asked to save, read-only, in one pass.

Plus the catalog answer he asked for (generated.json is 2 gigs stale behind
the view, 14 gigs uncatalogued), www's OPAL tracks.json still on segments_v3
where Slopmotion reads it, the CLAUDE.md next/ path that no longer exists,
and two stale-opens git says already shipped.