edfcore 0.5.22 → 0.5.24

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 = "0.5.22";
+export declare const VERSION = "0.5.24";
 //# 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 = '0.5.22';
+export const VERSION = '0.5.24';
 //# sourceMappingURL=constants.js.map

package/docs/CHANGELOG.md

@@ -6,6 +6,53 @@ 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.
 
+## 0.5.24
+
+- **Added** "the whole flow" at the end of `discontinuous.md`, run, and the diagram at the top of it.
+  That page opens with an ASCII timeline of one file and closes with a four-step program over the
+  same file printing six lines. `discontinuous-page.test.ts` covers the objects in between — the
+  `index.segments` and `index.gaps` blocks, and `locate(13.5)`. Both ends of it were prose.
+- The diagram is the page's premise: six one-second records, a ten-second hole after record 2, and a
+  row of byte offsets — 768, 1400, 2032, 2664, 3296, 3928 — which is "they are still stored back to
+  back on disk" written as numbers. A reader takes those to a hex editor. Both rows are now read out
+  of the diagram and checked, and the same offsets appear again in the four-read block under
+  "openEdf never scans", where they are the ranges the open actually issues.
+- Three of the program's six lines come from `console.log` inside loops the page writes itself, so
+  they are composed here by the same expressions and compared against the page's text.
+  `256 samples from 2 s` and `256 samples from 13 s` are the ones worth having: that is
+  `trimToWindow` applied to a window spanning a gap, and the second is the whole point of the page —
+  a reader trusting the nominal grid would expect the second chunk to start at 4 s.
+- `index.locate(5)` goes with them. It returns `undefined` for an instant inside the hole, which is
+  the answer the page prints and the one a caller has to handle: there is no sample there, and no
+  index can invent one. Step 1 is checked to reach its verdict without issuing a read, which is what
+  "no extra I/O" in its own comment claims.
+
+## 0.5.23
+
+- **Added** the conformance report `validation.md` prints, printed. That page ends with a whole
+  program and the output it produces on one file: a header line, a summary line, two diagnostic
+  blocks, and two per-signal stat lines. `validate-page.test.ts` covers the code table higher up and
+  `validation-report.test.ts` the `report.ok` rule; the transcript — the thing a reader compares
+  their own terminal against — was prose.
+- The stat lines are the reason it is worth running. They are not printed by edfcore at all: the
+  page composes them from `report.signalStats`, `getSignal` and arithmetic of its own, and the
+  paragraph under them is an argument built on the result. "The first channel uses half a percent of
+  the range its header declares, which is legal and lossy." "81,806 of its 153,600 samples fall
+  outside the range the header declares, so that declaration is wrong." A change to
+  `observedDigitalMin`, `outOfDigitalRangeCount` or `sampleCount` would leave the page arguing from
+  figures the library no longer produces.
+- The snippet is now run rather than restated: the stat lines are built by the same expression the
+  page shows and compared against the page's own text character for character, along with the
+  summary line's `no errors, 4 diagnostics, 600 records and 650400 bytes read`.
+- The fixture is built to the page's description and its geometry falls out of the transcript —
+  600 one-second records at 650,400 bytes is 1,084 a record, which is two 256-sample channels and a
+  30-sample annotation region, and 153,600 samples is 600 x 256. The one number that cannot be
+  derived is how many fall outside; that is read off the page and built into the sample generator,
+  which is what makes the count an assertion about the counter rather than about the waveform.
+- The last sentence of the paragraph is checked too. "Any consumer that clamps to it returns
+  different numbers for this file than edfcore does" — a read returns the bytes as written, at
+  -150..150 outside the declared -100..100, and `clampToDigitalRange` returns something else.
+
 ## 0.5.22
 
 - **Added** the scrolling table on `large-files.md`, measured. `large-files-cost.test.ts` runs the

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "edfcore",
-  "version": "0.5.22",
+  "version": "0.5.24",
   "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 = '0.5.22';
+export const VERSION = '0.5.24';