edfcore 0.4.276 → 0.4.278

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 = "0.4.276";
+export declare const VERSION = "0.4.278";
 //# 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 = '0.4.276';
+export const VERSION = '0.4.278';
 //# sourceMappingURL=constants.js.map

package/docs/CHANGELOG.md

@@ -6,6 +6,36 @@ 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.
 
+## 0.4.278
+
+- **Corrected** a sentence in `tests/README.md` that contradicted the rest of the page and the
+  repository. It opened with "No file anyone else wrote is in this repository", and the six
+  committed binaries under `corpus/golden/` were written by pyEDFlib's own `EdfWriter` —
+  `scripts/golden/README.md` says so in its second line, "nothing in `tests/corpus/golden/` is
+  produced by edfcore", and `AGENTS.md` explains they exist "because the parity harness has to
+  compare against bytes another implementation wrote". The page's own later section is headed
+  "Checking against files we did not write".
+- The true claim is narrower and worth stating properly: nobody else's *recording* is committed —
+  the downloaded corpus is gitignored under `tests/corpus/files/` — and those six were generated
+  locally from data this repository specifies. That another implementation wrote them is not an
+  exception to the policy, it is the entire reason they are kept: regenerating them with
+  `support/writer.ts` would make the comparison circular and prove nothing.
+
+## 0.4.277
+
+- **Fixed** a test budget that has now been wrong twice, by making it a different kind of number.
+  `extreme-geometry.test.ts` asserts that a diagnostic per record does not blow the call stack —
+  `TIMEKEEPING_TAL_MISSING` is per record, and `push(...array)` gives up around 125,000 arguments
+  — and it carried a 30 second timeout. How long the sweep takes is not the property. A timeout
+  set just above the observed duration measures the machine instead of the code, and reports a red
+  build in something the test does not touch.
+- It started at vitest's 5 second default, which the 200,000-record sweep landed a few hundred
+  milliseconds under on its own, so it tipped over whenever the rest of the suite ran beside it.
+  Thirty seconds fixed that and repeated the mistake one size up: the suite kept growing — 2,074
+  tests now, one of which spawns a TypeScript compiler over 102 files — and on a machine already
+  busy with unrelated work the sweep took 72 seconds and failed again. It is five minutes now, and
+  named for what it is: an infinite loop still fails, a loaded laptop does not.
+
 ## 0.4.276
 
 - **Added** the check for the other ban `AGENTS.md` lists: "No `Date` anywhere. EDF stores local

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "edfcore",
-  "version": "0.4.276",
+  "version": "0.4.278",
   "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 = '0.4.276';
+export const VERSION = '0.4.278';