Reading findings

Every check Atoo runs reports in the same shape: a finding, with a severity, the thing it is about, what it found, and a code. This page says what each part means and what to do about it.

The three severities

Severity	Means	What to do
ERROR	The desk cannot be put in that state: a socket the desk does not have, a stereo pair that starts on an even socket, a value out of range.	Fix it. A plan with an error is refused by apply.
WARNING	It can be written, but it is probably not what you meant: the critic disagrees with the plan, a DCA is at the floor with live channels in it.	Read it, then decide. Warnings never stop anything.
INFO	Something you should know but need not act on, most often a choice the plan made that your files did not state.	Skim. If a choice is wrong, state it next time.

INFO: a choice made, and said

Your conventions will not say everything. When they are silent, the plan still has to decide, because a desk cannot be left half-set: a bus is mono or stereo, a Matrix has a number. Atoo makes the usual choice, and tells you at INFO, with what it was decided from:

busses[MON5, MON6, Drums, …] (11 busses): width set to stereo – the conventions did not state it; decided by the generator: Instructions add MON5 and MON6 as stereo IEM mixes … / Mix busses conventionally stereo / FX busses conventionally stereo (inferred.bus-width)

A wrong choice is visible at soundcheck and costs a moment. If it is wrong, add a line to your conventions (“monitor busses are mono”), and the next build will not have to guess.

The verdict

atoo verify, and the pre-flight on the Apply screen, end with one line that sums up the findings:

Verdict	Means
VERIFIED: no findings. Plan is applicable.	nothing to report
APPLICABLE: notes only -- choices made where the conventions were silent.	INFO only
APPLICABLE: warnings only.	warnings, no errors
BLOCKED: do not apply.	at least one error

For example, on the Demo show, from an Atoo checkout:

atoo verify --plan starter/demo-show/plan.json --input starter/demo-show/input.csv

0 error(s), 0 warning(s), 4 note(s)

  [info] plan: 12 channel(s) carry a preset's values, written at the processing stage and read back like every other write: ch 1 Kick In (83), ch 2 Snare (83), ch 3 H Tom (83), ch 4 Fl Tom (83), ch 5 OH (83), ch 6 Bass (83), ch 7 E Gtr (83), ch 8 E Gtr (83), ch 9 Keys (83), ch 10 Vocal (80), ch 11 Vocal (80), ch 12 Vocal (80)  (preset.expanded)
  [info] plan: 3 bus(ses) carry a bus preset's processing, written at the processing stage and read back like every other write: Drums: 'Mix Bus' (55), Vocals: 'Mix Bus' (55), Vox Verb: 'Vox Verb' (55)  (preset.bus-expanded)
  [info] plan: 1 FX insert(s) are loaded and switched on, each read back: Vox Verb: V-REV on FX5 (saved bypassed, switched on)  (fx.inserts)
  [info] channels[Music]: aux 1 'Music' is not on the input list; built because instructions.md says 'Walk-in music ("Music") on Aux 1, stereo, source LCL 3-4. Fader at -20dB', on LCL 3-4 as it states  (reconcile.from-instructions)

APPLICABLE: notes only -- choices made where the conventions were silent.

The same message for several objects is one line naming them all, so a count can be larger than the number of lines: the first note here names twelve channels.

Codes

The code at the end of each finding, such as inferred.bus-width or patch.stereo-pair-not-odd-even, stays the same from version to version. Search these pages for it, or quote it in a bug report. Its first word says which check found it:

Starts with	Found by
inferred., reply., build.	the build: a choice made, or how the model’s draft was filled in
rebuild.	the browser, from what a rebuild changed against the plan it replaced
critic.	the critic pass, a second model call that checks the plan against your conventions
patch., strip., group. and others	verify: the plan against the desk’s limits and your input list
apply.	the Apply pre-flight and the apply run
a letter code and three digits, such as R005 or CV016	the checks audit and Watch run on the desk itself; the Rules reference lists every one
WATCH:	one of your own watch conditions

A critic finding starts with a short claim, such as Keys bus created, but there is only one keyboard, and then gives its reasons. The critic is a model giving a second opinion, so its findings are WARNING at most and never stop an apply. When it finds one of your lines left out and quotes it word for word, the build’s fix-up round carries it out; the critic’s warning then becomes a note, build.critic-repair, quoting your line; where code cannot see whether it was carried out, the note asks you to confirm it in the plan. A warning that ends still not carried out after the fix-up round is a line to add by hand, or to state more plainly. One that starts On the first plan: was about the plan the fix-up round replaced, and may already be mended.

In the browser

The Findings panel down the right of Build and Apply opens with a verdict, in a large tag: green Ready to apply when nothing is an error, red Blocked with the number of errors beside it, or gray No plan yet when there are findings but nothing has been built. Under it are three counts: the things to check, the critic’s advice and the notes on choices the build made. A count that is not zero is in bold. If the plan is older than the files it was built from, the tag is a yellow Stale instead of Ready to apply: The input list changed since the last build: rebuild before applying. A Rebuild button sits beside it, and the build’s and the critic’s notes are about the plan as it was built. A plan with errors still reads Blocked, with the same note under it. Nothing stops you applying a stale plan, and the Apply screen says it beside the plan; it is the plan as it was built, without your edit.

The number beside Findings, also on the panel’s tab when it is folded away, counts the errors and warnings. It is gray at 0, yellow when there are only warnings, and red when any of them is an error.

Below the verdict, each finding is one line: its severity tag (ERROR, WARNING or INFO), where it is, named the way the desk names it (CH 23 Crowd, AUX 2 HOST), and what it says. Click the line for the rest: the explanation, the fix, the code, which check found it, and links to the row, tile or line it is about. One finding about several strips is one line, such as 3 strips built from instructions.md, not on the input list, with each strip listed inside.

• To check: errors, then warnings. Findings about strips come before findings about files. With warnings and no errors the verdict reads, for example, Ready to apply · 1 to check. After a rebuild, each strip that changed although nothing it is made from changed is one of these, rebuild.unasked-change: Not asked for: gain 20 → 26 dB on CH 6 Bass DI, with links to its row and to the step 04 list of changes (What a rebuild changed).
• Advice: the critic’s warnings, marked with a dashed edge. They are a second opinion and never block an apply.
• ▸ Choices: the INFO notes that are choices, folded: what the build decided where your conventions were silent, and what a full apply does beyond the plan’s own strips.
• ▸ Build details: notes on how the build read the model’s draft, such as the order it numbered channels in, and notes that restate your conventions rather than choose for them: what the Category table filled in, which preset values the plan carries, an effect loaded and read back, and what a full apply resets first (apply.init). They are kept for debugging, and verify and build still print them (build prints verify’s copy of one only as a count, since it has just printed its own). A warning or an error is never put here.

Tick a finding’s box once you have dealt with it, and it moves to the folded Addressed list at the foot of the panel, its tag gray, so a ticked error no longer reads as live. Watch has its own table instead (Watch the desk during a show).