edfcore 0.5.10 → 0.5.12

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 = "0.5.10";
+export declare const VERSION = "0.5.12";
 //# 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 = '0.5.10';
+export const VERSION = '0.5.12';
 //# sourceMappingURL=constants.js.map

package/docs/CHANGELOG.md

@@ -6,6 +6,48 @@ 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.
 
+## 0.5.12
+
+- **Added** the six mistakes `AGENTS.md` tells a code generator to avoid, run. That file carries two
+  lists: "Things that look like bugs and are not" is bound to tests by `agents-rules.test.ts`, which
+  fails if a rule is added without one; the other list had only `agents-snippet.test-d.ts`, which
+  compiles the fence beneath it. Whether the six sentences are true of a running file was
+  established nowhere.
+- They are the sentences most likely to be acted on without being read carefully, because they are
+  written for something generating code rather than for someone reading. Each describes a mistake
+  that produces OUTPUT rather than an error: microvolts that are really ADC counts, a chunk that is
+  really an array, an index computed from a rate that does not exist, an event compared on the wrong
+  axis, one rate assumed for a file with three, a defect reported as a value and never looked at.
+- The fourth has arithmetic in it and is checked as arithmetic. The two annotation onset fields
+  "differ by record 0's sub-second offset", so the fixture declares a quarter-second one and the
+  difference is asserted against it. On a file with no offset the two are equal, and a test written
+  against such a file would pass while the sentence was wrong.
+- The sixth is checked with the console replaced: the diagnostics arrive as values on the header,
+  nothing throws, and no `log`, `warn`, `error`, `info`, `debug` or `trace` is called.
+- The list is enumerated from `AGENTS.md` and bound by content rather than position, so a reordered
+  list keeps its bindings, a rewritten entry loses one loudly, and a seventh mistake is a mistake
+  with no test until it has one.
+
+## 0.5.11
+
+- **Added** the sentence `api-helpers.md` gives as the entire reason `readEnvelopeAtResolution`
+  exists separately from `readEnvelope`: "Widths that disagree cannot be drawn on one axis." Above
+  it the page lists the two ways that went wrong, each with the numbers it produced — a window of
+  11 s asked at 1 s per bucket coming back as 0.27 s per bucket in one chunk and 0.09 s in the
+  other (0.2.31), and a 100 s run at 30 s per bucket giving four buckets of 25 s while a 60 s run in
+  the same call gave two of 30 s (0.3.9). Both were history, and neither was pinned.
+- The existing envelope tests cover the bucket COUNT — that it is not always what you asked for, and
+  why the two entry points clamp differently — which is the field a caller loops over. The WIDTH is
+  the field a caller draws with. A viewer placing bucket `b` at `startSeconds + b * secondsPerBucket`
+  gets a plot whose x-axis changes scale halfway across if two chunks disagree; nothing looks wrong,
+  the second half of the trace is simply drawn at a different rate from the first.
+- Both ways one call ends up with runs of different lengths are exercised: a window spanning a gap,
+  and a window that does not begin on a record boundary and so produces a first run wider than it
+  asked for. The run lengths are asserted to differ, or the check is one run repeated.
+- The grid is also asserted to cover its run and not sprawl past it — the span of the buckets is at
+  least the chunk's duration and overhangs it by less than one bucket — which is what makes the
+  reported width usable for placing them.
+
 ## 0.5.10
 
 - **Added** the hex dump line `diagnostics.md` prints. The page shows one line — `bytes: 63 61 66

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "edfcore",
-  "version": "0.5.10",
+  "version": "0.5.12",
   "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 = '0.5.10';
+export const VERSION = '0.5.12';