edfcore 0.4.345 → 0.4.347

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 = "0.4.345";
+export declare const VERSION = "0.4.347";
 //# 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 = '0.4.345';
+export const VERSION = '0.4.347';
 //# sourceMappingURL=constants.js.map

package/docs/CHANGELOG.md

@@ -6,6 +6,39 @@ 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.
 
+## 0.4.347
+
+- **Added** an execution of the `bigint` ticks decision on `design-decisions.md`. It is defended
+  with a specific, hard-to-notice failure — two onsets that are one instant on disk comparing
+  unequal after a round trip through a binary fraction, an averaging window landing a sample off
+  for a subset of trials — and that is arithmetic, so it can be shown instead of asserted.
+- Worth showing, because the cost paragraph underneath is a standing invitation to convert to
+  seconds and be done with it. Three tenths is now demonstrated to be three times one tenth to the
+  tick and not in seconds, and a seven-decimal onset survives exactly while its float does not.
+- The path itself is checked from the source. `tal/ticks.ts` states that `parseFloat`,
+  `Number(text)` and float arithmetic appear nowhere on it, and nothing made that true — it is one
+  careless conversion away from being false while every existing test passes, because the two agree
+  on almost every value. Scanned with comments stripped, since the docblock making the claim quotes
+  the very calls being looked for.
+
+## 0.4.346
+
+- **Added** an execution of the `strict` section of `design-decisions.md`, which is where the
+  odd-looking choices are defended and where two pieces of load-bearing behaviour had no test that
+  read the page.
+- The `info` exemption is the first. Without it, `strict: true` would reject nearly every real
+  recording — `DATE_CLIPPED_TO_1985_2084` is on almost all of them — and it would do it while
+  looking correct, because the mode is called strict and rejecting is what it is for. Removing it
+  reads as tightening. A conforming file now has to open under strict and still carry that
+  diagnostic.
+- The second is the list of conditions that throw "either way". Three of the four are properties of
+  the header and are checked in both modes. The fourth, record onsets that go backwards, gets its
+  own section, because choosing a fixture where it throws at open would have hidden something real:
+  `openEdf` reads records 0 and n-1 and nothing else, so a reversal between them throws immediately
+  and one in the middle is not something it has seen. That file opens with a probed index and
+  throws the moment `buildRecordIndex` reads the records — "either way" is about the two modes, not
+  about throwing before the bytes are read.
+
 ## 0.4.345
 
 - **Added** a check that the six single-interface field tables on `api-types.md` list exactly the

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "edfcore",
-  "version": "0.4.345",
+  "version": "0.4.347",
   "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 = '0.4.345';
+export const VERSION = '0.4.347';