edf2csv 0.6.86 → 0.6.88

package/docs/CHANGELOG.md

@@ -3,6 +3,38 @@
 Notable changes to edf2csv. Versions follow [semantic versioning](https://semver.org); while the
 major version is 0, a minor bump may contain breaking changes.
 
+## 0.6.88
+
+### nothing checked that an anchor arrives where it says
+
+The duplicate id fixed in 0.6.87 had been shipping for as long as both headings had existed, and it
+survived every check this site has: it is valid-looking markup, it renders correctly, no link
+404s, and the page reads properly top to bottom. The only symptom is that one link in the contents
+list arrives at the wrong paragraph — which is exactly what proof-reading does not catch and what a
+build can settle in a dozen lines.
+
+Every emitted page is now checked for two things: no id used twice, and no `href="#..."` pointing
+at an element that does not exist. Verified by breaking each one deliberately — removing the
+numbering from the slugger, then misspelling a contents link — and confirming the build stops with
+the page and the id named, before putting both back.
+
+## 0.6.87
+
+### two headings shared an id and the second link went to the first
+
+`NO_SAMPLES` is documented twice on the warnings page, deliberately: once as the warning a single
+empty channel raises, once as the fatal error an entirely empty recording raises, with a paragraph
+between them explaining that the same name means two things. Both headings were given
+`id="no_samples"`. That is invalid HTML, and it fails in the quiet way — the browser scrolls to the
+first match, so the contents entry for the fatal error took the reader to the warning, and 0.6.84's
+permalink beside the second heading handed out a link that goes to the first. Nothing 404s. It
+simply arrives at the wrong paragraph.
+
+Slugs are now unique per page: the first occurrence keeps the plain id, so every link that already
+points at one still lands where it did, and repeats are numbered. The renderer and the contents
+list share one slugger walking the same headings in the same order, because two independent
+implementations of "which one is the duplicate" is how they would disagree later.
+
 ## 0.6.86
 
 ### the contents list omitted every diagnostic code on the page about them

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "edf2csv",
-  "version": "0.6.86",
+  "version": "0.6.88",
   "description": "Convert EDF, EDF+ and BDF biosignal recordings (European Data Format) to CSV from the command line. Local, streaming, and never resamples or alters units.",
   "keywords": [
     "edf",