edfcore 0.4.381 → 0.4.383

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 = "0.4.381";
+export declare const VERSION = "0.4.383";
 //# 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 = '0.4.381';
+export const VERSION = '0.4.383';
 //# 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.
 
+## 0.4.383
+
+- **Added** a check for the last paragraph of `data-sources.md`, which says what survives
+  `close()`: the header and the timeline stay readable, and `readRecords`, `readWindow`,
+  `readAnnotations` and an unmemoised `index.locate` fail. Every `fileSource` example on the site
+  is wrapped in `try { … } finally { await source.close?.() }`, so a recording outliving its handle
+  is the ordinary shape of a program here, and that paragraph was the only statement of which half
+  of the object still works.
+- The carve-out in it turned out to be narrower than it reads. A probed index memoises records 0
+  and n−1 at open and every record a search walks over, so `locate(0.5)` answers after the close
+  only if it already ran before it — on a freshly opened file the binary search probes a midpoint
+  first. Both cases are pinned, since a later memoisation change makes that quietly more generous.
+- It does not claim the rejection is an `EdfError`. It is whatever the source raises: for
+  `fileHandleSource` that is Node's own `Error: file closed`, and `isEdfError` returns false for
+  it. Closing your own handle is not a file defect.
+
+## 0.4.382
+
+- **Fixed** the docblock at the top of `scripts/release.mjs`, which described the scheme that was
+  replaced in 0.4.326: "this script only moves the version forward and creates the GitHub Release;
+  `publish.yml` sees that release and publishes to npm". It has created no release and watched no
+  release since — the tag it pushes is what publishes — and that paragraph is the first thing
+  anyone reads before running the script.
+- `release-model.test.ts` has asserted the trigger, the ordering and the absence of `gh release
+  create` since 0.4.327, and every one of those checks passed the whole time. They read what the
+  script calls, and a comment calls nothing.
+- So the docblock is checked too, by the affirmative rather than the negative: it has to name the
+  pushed tag as what publishes, say outright that the script creates no release, and point at
+  `npm run announce`. A docblock rewritten back to the old scheme has to delete those sentences to
+  read coherently, and deleting them fails.
+
 ## 0.4.381
 
 - **Added** the closure this batch earned: every documentation page has to be named by a test that

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "edfcore",
-  "version": "0.4.381",
+  "version": "0.4.383",
   "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 = '0.4.381';
+export const VERSION = '0.4.383';