edfcore 0.4.273 → 0.4.275

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 = "0.4.273";
+export declare const VERSION = "0.4.275";
 //# 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 = '0.4.273';
+export const VERSION = '0.4.275';
 //# 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.
 
+## 0.4.275
+
+- **Added** the check for a ban `AGENTS.md` lists under things that look like bugs and are not:
+  `TextDecoder` belongs in `src/tal/` and nowhere else. Verified on Node v24.4.0, every `latin1`
+  label it accepts reports `windows-1252` and decodes byte `0x80` as `U+0080`, while the WHATWG
+  standard mandates `U+20AC` — so a `TextDecoder` on the header path would make the same file
+  produce different strings in Node and in a browser, from a library whose claim is that it reads
+  the same bytes the same way everywhere. `src/tal/` is exempt because annotation text really is
+  UTF-8, the one encoding every runtime agrees on.
+- The stripper removes string literals as well as comments, and that is not tidiness:
+  `header/fields.ts` contains the word `TextDecoder` inside a diagnostic message explaining this
+  rule to a user, so a comments-only sweep reads the file that documents the ban as the file that
+  breaks it. It looked like a live violation until the line was read.
+- A second test asserts the consequence rather than the rule — `0x80` in a real signal label,
+  read back through `openEdf`, comes out `U+0080` and not a euro sign. `latin1.test.ts` pins the
+  decoder in isolation; this is the path a header actually takes.
+
+## 0.4.274
+
+- **Added** the check that edfcore never writes to the console. The README says it twice, and the
+  survey table explaining why this library exists lists what the alternatives do instead —
+  "`console.warn` and `null`, or bare thrown strings". Diagnostics are values on the result
+  precisely so that reporting them is the caller's decision; a library that logs takes that
+  decision away, breaks anything parsing the consumer's stdout, and on a header diagnostic puts a
+  patient's name into whatever collects the logs. One `console.warn` left in during debugging
+  would have shipped, and the only way to find it was to be the person whose output it landed in.
+- Both halves, because neither is enough on its own. A static sweep of `src/` with comments
+  stripped catches a call on a path no test happens to take — the survey table is quoted in
+  comments, so stripping is load-bearing. Running the library with every console method trapped
+  catches one the sweep cannot see, and that was verified rather than assumed: a call written as
+  `globalThis['con' + 'sole'].warn` leaves no literal `console` in the file, passes the sweep, and
+  is caught by the trap.
+
 ## 0.4.273
 
 - **Made** the fuzz suite assert the clause it was missing. `tests/README.md` states the safety

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "edfcore",
-  "version": "0.4.273",
+  "version": "0.4.275",
   "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 = '0.4.273';
+export const VERSION = '0.4.275';