edfcore 0.4.474 → 0.4.476

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 = "0.4.474";
+export declare const VERSION = "0.4.476";
 //# 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 = '0.4.474';
+export const VERSION = '0.4.476';
 //# 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.
 
+## 0.4.476
+
+- **Added** the zero-record-duration case to `readTriggers`, the last reader in the package
+  without one. `formatHeader`, `resolveTimeWindow`, `trimToWindow`, `gridSampleStartTicks` and
+  `readEnvelope` all have theirs; BDF permits a zero record duration, and this is the reader whose
+  own comment states what it means — every sample of every record sits at one instant.
+- What that costs is that time can no longer order the events, and `sampleIndex` is the only thing
+  that can. Three trigger runs come back as three events, all at tick 0, at sample indices 0, 2
+  and 5 — so the change detection still works when the times are all equal, which is the part
+  nothing else in the suite says.
+- The all-or-nothing rule is pinned with it: a window starting after that instant yields nothing
+  rather than the tail of the recording, and a zero-length window yields nothing either.
+- The guard the comment sits on is NOT claimed to be covered by this, and the test says so.
+  `ceilDiv(s * 0n, samplesPerRecord)` is 0 for every `s`, so its two branches agree on every
+  input and no test can tell them apart. What is checked is the outcome.
+
+## 0.4.475
+
+- **Added** the boundary cases for `quote`, which is what keeps a `raw:` line on one line however
+  hostile the 80 bytes behind it are. It has four boundaries and only the space at 0x20 was
+  pinned; the other three could each be narrowed with the whole suite still green.
+- All three are about ordinary bytes. `~` is 0x7e, the last printable ASCII character, and it
+  appears in real prefiltering fields (`HP:~0.1Hz`) — printed as an escape it reads as a control
+  byte in a field where the writer typed a tilde. U+00FF is the last character with a two-digit
+  escape, so narrowing that bound spells it `\u{ff}` while every other Latin-1 byte in the same
+  field is spelled `\xNN`, and one header field prints in two notations.
+- The one-line property is asserted directly as well: a field holding a newline, a tab and a
+  carriage return renders as exactly one `raw:` line.
+- And the deliberately different rule on `expected:` and `actual:` is asserted beside it. Those
+  are unquoted values in fixed-width columns, so a control byte becomes one dot through
+  `printable` rather than a four-character escape — the same byte prints two ways on two lines,
+  which is the design and not an inconsistency, and neither half is obvious without the other.
+
 ## 0.4.474
 
 - **Added** all twelve month names to the `dd-MMM-yyyy` subfield tests. `MONTH_ABBREVIATIONS

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "edfcore",
-  "version": "0.4.474",
+  "version": "0.4.476",
   "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 = '0.4.474';
+export const VERSION = '0.4.476';