edf2csv 0.6.25 → 0.6.27

package/docs/CHANGELOG.md

@@ -3,6 +3,29 @@
 Notable changes to edf2csv. Versions follow [semantic versioning](https://semver.org); while the
 major version is 0, a minor bump may contain breaking changes.
 
+## 0.6.27
+
+### Fixed: the warnings table gave `UNUSABLE_PHYSICAL_RANGE` only its overflow half
+
+The code picks the word from which way the span failed — `too ${underflowed ? 'small' : 'large'}
+to represent` — and the section on the same page explains both at length, including the
+underflow case 0.5.83 was released for. The one-line summary said "too wide to represent",
+which is the half a reader scanning the table would take for the whole rule and the half that
+does *not* apply to a magnetometer, the channel type the section names.
+
+## 0.6.26
+
+### Fixed: the warnings table still said a mixed-rate file means "several output files"
+
+`--layout long` writes one, and the warning's own hint says so — it has two forms, and the one
+it prints under `long` reads "They share one table, each row carrying its own time". The summary
+line at the top of the page kept the pre-0.5.0 answer.
+
+There is a test for exactly this class, added when 0.5.0 landed the flag: it requires every page
+claiming a mixed-rate recording becomes several files to mention the layout that does not. It
+missed this row because its phrase list looks for "several signals files" and the row says
+"several output files" — one word apart from the thing it was written to catch.
+
 ## 0.6.25
 
 ### Fixed: `INPUT_CHANGED` was one row of a summary table and nothing else

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "edf2csv",
-  "version": "0.6.25",
+  "version": "0.6.27",
   "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",