edfcore 0.4.343 → 0.4.345

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 = "0.4.343";
+export declare const VERSION = "0.4.345";
 //# 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 = '0.4.343';
+export const VERSION = '0.4.345';
 //# 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.
 
+## 0.4.345
+
+- **Added** a check that the six single-interface field tables on `api-types.md` list exactly the
+  members those interfaces have, in declaration order. That page is the field-by-field reference
+  and where someone goes to learn what is on an `EdfHeader` without opening `types.ts`, and all six
+  tables are hand-typed copies of a declaration.
+- Both directions of drift are silent. A member added to `types.ts` and not to the page is a field
+  nobody can discover — which is how `declaredRecordCount` or `recordCountSource` would go unused
+  by exactly the caller who needed them. A row for a member that no longer exists is worse: it
+  reads as an API, the reader writes `header.something`, and the answer is `undefined` rather than
+  an error. `docs-coverage.test.ts` proves every export is mentioned; nothing had looked inside a
+  type.
+- One table spans three interfaces and cannot have its rows attributed to any one of them, so it
+  is skipped by name and the number of skipped tables is asserted — an exemption that cannot
+  quietly grow.
+
+## 0.4.344
+
+- **Added** a check over what `data-sources.md` publishes about `cachedSource`. `cache.test.ts`
+  covers the behaviour thoroughly — eviction order, deduplication, copies, clamping, abort — but
+  every one of those tests passes its own sizes in order to be deterministic, so the two DEFAULTS
+  a caller inherits by writing `cachedSource(inner)` were invisible to all of them.
+- Those defaults are what almost every caller actually gets. They are checked twice: against the
+  private constants they come from, and behaviourally — one byte asked for, one whole 1 MiB block
+  fetched, and the rest of that block then free.
+- Also executed: the worked recipe at the end of the section, which is arithmetic a reader is told
+  to do themselves and which makes a claim about cost. `readHeader` costs two reads against the
+  uncached source, the derived block size is a whole number of records, and `openEdf`'s second pass
+  pulls in block 0 — the block the first record read needs anyway.
+
 ## 0.4.343
 
 - **Added** `tests/property/scaling.test.ts`, which measures the bound 0.4.342 asserted. Random

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "edfcore",
-  "version": "0.4.343",
+  "version": "0.4.345",
   "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 = '0.4.343';
+export const VERSION = '0.4.345';