edfcore 0.4.459 → 0.4.461

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 = "0.4.459";
+export declare const VERSION = "0.4.461";
 //# 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 = '0.4.459';
+export const VERSION = '0.4.461';
 //# sourceMappingURL=constants.js.map

package/docs/CHANGELOG.md

@@ -6,6 +6,40 @@ 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.
 
+## 0.4.461
+
+- **Added** the exact-edge cases for `resolveTimeWindow` on a segmented index. The window is
+  half-open — `[start, start + duration)` — and the segmented path decides that on one line:
+  `segmentEndTicks <= windowStartTicks || segmentStartTicks >= windowEndTicks`. Relaxing either
+  comparison to a strict one left the whole suite green, because no window in it had ever landed
+  on a segment edge exactly: they overlapped by a record or missed by ninety seconds.
+- The cost is a duplicated record at every gap. Walking a discontinuous file segment by segment —
+  `[0, 4)` then `[4, 100)` on the fixture here — a relaxed left comparison returns record 3 in
+  both, so a consumer concatenating the two gets one second of signal twice, at the seam where a
+  sleep study is most likely to be scored.
+- Four cases: a segment ending exactly where the window starts is out, a segment starting exactly
+  where the window ends is out, a segment starting exactly where the window starts is IN, and a
+  three-window walk that visits records 0..7 once each. The third is what stops the first two
+  passing on an implementation that excluded every boundary rather than the right one.
+- Found by mutating the comparisons rather than by reading the coverage report, which called the
+  line covered.
+
+## 0.4.460
+
+- **Added** the fourth rung of `timekeepingDefect`'s ladder, which had never run: a timekeeping
+  TAL written `+t 0x14 0x14 0x14 0x00` — two empty texts where the grammar allows exactly one.
+- The other three rungs were covered. Text merged into the timekeeping TAL is the destructive one
+  and has its own file; a stray duration and the widespread `+t 0x14 0x00` shorthand are both
+  benign and both tested. This is the shorthand's opposite — a writer looping over an events list
+  that happens to be empty and terminating each iteration anyway — and deleting the check for it
+  failed nothing.
+- Pinned as BENIGN, which is the part that matters. The two kinds carry separate once-per-call
+  flags, and a rung landing on the destructive side would report every record of a file whose
+  writer does this throughout. The fixture makes all three records malformed and asserts one
+  report, with the "nothing was lost" advice rather than the "EVERY affected record" advice.
+- And the onset still governs the timeline: the span is asserted, so a rung that reported the
+  defect while losing the timing would fail here rather than pass quietly.
+
 ## 0.4.459
 
 - **Added** the two `COMMA_DECIMAL_SEPARATOR` cases that were missing: a comma in the declared

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "edfcore",
-  "version": "0.4.459",
+  "version": "0.4.461",
   "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 = '0.4.459';
+export const VERSION = '0.4.461';