edfcore 0.6.31 → 0.6.33

package/dist/constants.d.ts

@@ -109,5 +109,5 @@ export declare const SIGNAL_FIELD_BLOCK_OFFSETS: {
     readonly reserved: 224;
 };
 /** Published package version. Kept in sync with package.json by a test. */
-export declare const VERSION = "0.6.31";
+export declare const VERSION = "0.6.33";
 //# sourceMappingURL=constants.d.ts.map

package/dist/constants.js

@@ -79,5 +79,5 @@ export const SIGNAL_FIELD_BLOCK_OFFSETS = {
     reserved: 224,
 };
 /** Published package version. Kept in sync with package.json by a test. */
-export const VERSION = '0.6.31';
+export const VERSION = '0.6.33';
 //# sourceMappingURL=constants.js.map

package/docs/CHANGELOG.md

@@ -6,6 +6,37 @@ alone does not tell you whether you were affected.
 edfcore is pre-1.0. Patch releases have carried behaviour changes where the old behaviour was a
 defect; those are called out below.
 
+## 0.6.33
+
+- **Added** a sixteenth shape to the `AWKWARD` matrix: BDF+D, 24-bit samples and a discontinuity in
+  one file. The matrix held both halves and never together — every BDF shape in it ran end to end,
+  and every discontinuous shape was 16-bit EDF — so a three-byte stride and a chunk boundary that
+  is not the nominal grid had never met.
+- They meet in one arithmetic. A read across a gap resolves a record range per contiguous run, and
+  each run's bytes are addressed as `headerByteLength + record * recordByteLength +
+  signal.recordByteOffset`, where every term is a multiple of `bytesPerSample` — 3 here and 2
+  everywhere else in the matrix — and a sign extension that is right for a 16-bit sample is wrong
+  for a 24-bit one. An off-by-one in either would show on this file and on no other.
+- Thirty sweeps run over the matrix and all of them pass over it. `twenty-four-bits-across-a-gap.
+  test.ts` checks the crossing is real rather than nominal: two segments, a seven-second hole, and
+  the same samples through `readWindow` across the join as through the whole record range.
+
+## 0.6.32
+
+- **Added** a fifteenth shape to the `AWKWARD` matrix: a file with two annotation channels of
+  different widths. EDF+ permits more than one and is specific about the asymmetry — the FIRST
+  carries the timekeeping TAL that states each record's onset, and any of them may carry events.
+  Two scorers writing into one file, or a device separating its own markers from a technician's
+  notes, produce exactly this.
+- `secondary-annotation-signal.test.ts` covers the decoding. What the matrix had never held is a
+  file where the answer to "the annotations channel" is two channels of different widths, so a
+  helper that finds one and stops loses half the events silently and a fixed stride through the
+  record is wrong for one of them. That is the failure a sweep over a whole file can see and a
+  targeted test cannot. Every sweep passes over it.
+- `two-scorers-two-channels.test.ts` also pins the property the sweeps do not state: one
+  `readAnnotations` call with no `signalIndices` returns both scorers, sorted together rather than
+  one channel after the other, each carrying the `signalIndex` that says which scorer wrote it.
+
 ## 0.6.31
 
 - **Added** a fourteenth shape to the `AWKWARD` matrix: a file whose last record is half there.

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "edfcore",
-  "version": "0.6.31",
+  "version": "0.6.33",
   "description": "Modern, typed, zero-dependency reader for EDF, EDF+, BDF and BDF+ biosignal files. Works in browsers and Node with true random access.",
   "keywords": [
     "edf",

package/src/constants.ts

@@ -93,4 +93,4 @@ export const SIGNAL_FIELD_BLOCK_OFFSETS = {
 } as const;
 
 /** Published package version. Kept in sync with package.json by a test. */
-export const VERSION = '0.6.31';
+export const VERSION = '0.6.33';