Record checkpoint B results

All five round-trip projects worked on the device. Notes past a
shortened end are kept and come back when lengthened, so the shorten
dialog says so. currentSequence slot vs list position is still open.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

docs/build-plan.md

@@ -253,6 +253,8 @@ Not yet tested: sample rates above 96 kHz or below 22.05 kHz; 64-bit float (refu
 4. Remove a sequence; add one in a different bank.
 5. Shorten a sequence that has notes: note what the device does with notes past the end.
 
+Results (2026-11-11, on the device): all five passed. Notes past a shortened end are kept and play again when the sequence is lengthened. B4 worked but did not settle whether `currentSequence` is a slot or a list position (format doc, open question 7). Details in `docs/checkpoint-b.md`. M8 is done.
+
 ## 8. Builds
 
 - Wails apps are built on each target system. Use one machine or CI runner per system; the Linux VPS alone cannot produce the Mac build.
@@ -267,7 +269,7 @@ Not yet tested: sample rates above 96 kHz or below 22.05 kHz; 64-bit float (refu
 | WAV variants the device rejects or plays wrongly | Largely retired by the WAV test pack (section 7): all eleven formats played correctly. Remaining unknowns are listed there. Akai's spec caps a sample at 20 minutes; refuse longer files. |
 | The app's preview cannot decode a WAV the device accepts | Each system's webview has its own audio decoder, and 8-bit, 32-bit integer or extensible files may fail in some. Try the webview's decoder first; if it fails, fall back to a Go route that decodes the PCM itself and returns 32-bit float samples for the frontend to play. `internal/wav` therefore needs a small decoder as well as the header reader. |
 | Mute group count | Resolved (checkpoint A): none and 1-16. The UI list offers Off, 1-16. |
-| Shortening sequences with notes | Checkpoint B5; until then warn and leave notes untouched. |
+| Shortening sequences with notes | Resolved (checkpoint B5): the device keeps notes past the end and plays them again when the sequence is lengthened. The app still asks before shortening and leaves the notes in the file. |
 | Tuning limits taken from a community toolkit | Resolved (checkpoint A): ±24 semitones and ±90 cents confirmed on the device. |
 | Very long or unusual file names | Keep names as they are; reject only characters illegal on Windows/FAT. Revisit if the device truncates. |
 | File drop quirks per system | Dialog-based assign is always available as the fallback. |

docs/checkpoint-b.md

@@ -51,10 +51,14 @@ On the device:
 
 ## Results
 
+Run on the MPC Sample on 2026-11-11 by bajdev. All five loaded and passed.
+
 | Project | Loads | Result |
 |---|---|---|
-| MpckitB1 | | |
-| MpckitB2 | | |
-| MpckitB3 | | |
-| MpckitB4 | | Sequence selected on open: |
-| MpckitB5 | | Notes past the new end: |
+| MpckitB1 | yes | Worked: plays as DemoFile. |
+| MpckitB2 | yes | Worked: A05 plays the kick; the notes stayed on A01. |
+| MpckitB3 | yes | Worked: the copy in bank C plays; the swapped pads changed places in the sequences. |
+| MpckitB4 | yes | Worked. Sequence 02 (A02) was selected on open. |
+| MpckitB5 | yes | Sequence 02 plays its 2 bars correctly. Lengthened back to 4 bars on the device, the notes in bars 3 and 4 are still there and play. |
+
+B4 does not settle open question 7. In MpckitB4 the stored list is [Bridge (slot 17), Sequence 02 (slot 2)] and `currentSequence` is 1, so "slot 1" and "position 1 in the list" both mean Sequence 02.

docs/xpj-format-and-scope.md

@@ -249,13 +249,23 @@ Four projects written by the Go writer (`tools/checkpoint-a`, built from commit
 
 Every item on the original 0.5.0 list (project name, pad sample assignment, sequence name, BPM, bar length, adding a sequence) plus pad tuning has now been written by these recipes and accepted by the device.
 
+### Checkpoint B (2026-11-11)
+
+Five projects made from DemoFile in the app, exported, and loaded on the device (details and results in `docs/checkpoint-b.md`). All loaded and worked:
+
+- **Unchanged round trip** (B1): plays as DemoFile.
+- **Moving a pad with notes** (B2): the sample plays from its new pad; the notes stay on the old slot, as expected.
+- **Copy across banks, swap pads with notes** (B3): as expected.
+- **Remove the current sequence, add one in bank B** (B4): works; Sequence 02 is selected on open (see open question 7).
+- **Shorten a sequence with notes** (B5): the notes past the new end are kept and play again once the sequence is lengthened (open question 4).
+
 ## 8. Open questions
 
 1. Device-test a moved and a copied pad.
 1a. ~~Confirm the tuning limits on the device.~~ Confirmed 2026-11-10 (checkpoint A): -24 to +24 semitones and -90 to +90 cents.
 2. Level and pan have not been seen changed; confirm they are written to `pad[n].mixable`.
 3. `triggerMode` values (0 and 3 seen) and how they pair with `loopMode`.
-4. When a sequence with notes is shortened, are notes past the new end kept or removed?
+4. ~~When a sequence with notes is shortened, are notes past the new end kept or removed?~~ Answered 2026-11-11 (checkpoint B5): kept. They don't play while the sequence is shorter, and play again when it is lengthened on the device.
 5. Sample rates outside 22.05-96 kHz and 64-bit float are untested. (The extensible header was confirmed in checkpoint A; other WAV bit depths and sample rates are confirmed; non-WAV formats are out of 0.5.0.)
 6. What key and tempo to write for a sample mpckit imports (the device detects them; mpckit cannot yet).
-7. Is `currentSequence` a slot number or an index into `sequences[]`? Every file seen has 0, and slot 0 is stored second, so the files cannot tell. mpckit treats it as a slot (build plan 4.3); checkpoint B4 should remove the current sequence and check which one the device opens.
+7. Is `currentSequence` a slot number or an index into `sequences[]`? Every file seen has 0, and slot 0 is stored second, so the files cannot tell. mpckit treats it as a slot (build plan 4.3). Checkpoint B4 (removing the current sequence) worked, with Sequence 02 selected on open, but in that file slot 1 and list position 1 are the same sequence, so it does not tell them apart. Still open.

frontend/src/lib/SequencePanel.svelte

@@ -25,7 +25,7 @@
     if (seq.hasNotes && l < seq.bars) {
       const ok = await confirm({
         title: "Shorten a sequence with notes?",
-        message: `${seq.name} has recorded notes. Notes after bar ${l} stay in the file; what the device does with them is not yet known.`,
+        message: `${seq.name} has recorded notes. Notes after bar ${l} will not play, but they are kept: lengthening the sequence again brings them back.`,
         items: [],
         action: "Shorten",
       });

pull-requests/m8-load-guard-rails/m8-load-and-guard-rails.md

@@ -18,6 +18,10 @@ Load itself, the unsaved-changes prompt on New and Load, the overwrite confirmat
 
 **Change after the first laptop run** (bajdev, 2026-11-11: checks 1–3 passed): disable the export buttons while a sample is missing. Before, a user could choose a folder and miss the error text.
 
+**Checkpoint B results** (bajdev, 2026-11-11): all five projects loaded and worked on the device, so M8's done-when is met. Recorded in `docs/checkpoint-b.md`, the format doc (section 7, open questions 4 and 7) and the build plan (checkpoint B, risk table).
+- Open question 4 is answered (B5): notes past a shortened end are kept, and play again when the sequence is lengthened. The shorten dialog now says so, instead of "not yet known".
+- Open question 7 (`currentSequence` slot or list position) is still open: in MpckitB4 both readings point at Sequence 02.
+
 ## Testing
 
 On the server: `go vet` and `go test ./...` pass. New tests: