edfcore 0.4.256 → 0.4.258

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 = "0.4.256";
+export declare const VERSION = "0.4.258";
 //# 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 = '0.4.256';
+export const VERSION = '0.4.258';
 //# 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.
 
+## 0.4.258
+
+- **Added** a check on the `Next:` convention, which `AGENTS.md` states as an absolute — "every
+  thrown message ends with a `Next:` clause naming what the caller should do" — and nothing
+  enforced. All 151 messages keep it today; what was missing is that the 152nd would not have
+  had to. The clause is the part that survives contact with a real user: "byte range [0, 512) is
+  outside the 256-byte buffer" says what happened, and "Next: check that the header and these
+  bytes came from the same file" says what it means.
+- Both halves are covered, which took two passes. `EdfFormatError` is never thrown with `new` — it
+  is built from a diagnostic by `fatalError`, `sink.fatal`, `scalingError` and `toFormatError` — so
+  reading only `throw new` sees 90 messages and misses the 61 that carry the larger share of the
+  contract. And finding where a `throw` ends cannot be done by balancing parentheses:
+  `[${offset}, ${offset + length})` closes one, and the first version of this check reported the
+  two messages using that interval notation as violations of a rule they keep.
+
+## 0.4.257
+
+- **Fixed** `AGENTS.md`'s description of the layering, which had been wrong about nearly every
+  tier. It sketched six — "`bytes`/`text` → `diagnostics` → `header`/`decode`/`tal` → `time` →
+  `io` → entry points" — where the declarations use eight, and grouped modules that are not
+  together: `bytes` is layer 0 and `text` is layer 1, `header`, `decode` and `tal` are three
+  different layers rather than one, and `io` spans two. 0.4.256 reasoned from that sentence to
+  correct a module's layer, which is a good argument for the sentence being right. It is now a
+  table of the eight, and says plainly that each module's own declaration is the source of truth
+  rather than a second definition.
+- **Added** a check that the summary names the layers that exist, and that the count in the
+  sentence above it matches. Only the numbers are compared — a prose list of members is the
+  inventory problem this project keeps deleting, and the declarations already answer membership.
+  What a summary can still get wrong unnoticed is the shape: a tier added or removed in one place
+  and not the other.
+
 ## 0.4.256
 
 - **Corrected** `src/tal/ticks.ts` from layer 3 to layer 1, and started enforcing the direction

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "edfcore",
-  "version": "0.4.256",
+  "version": "0.4.258",
   "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 = '0.4.256';
+export const VERSION = '0.4.258';