edf2csv 0.6.21 → 0.6.23

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.
 
+## 0.6.23
+
+### Fixed: "Record `n` starts at `n * record_duration` seconds, and always will"
+
+Said of EDF+C, and false since 0.4.9 made the first record's timekeeping TAL the point a
+recording is timed from. `fractional-start.edf` is a continuous fixture in this repository whose
+first row is `0.500`; `output-files.md` states the correct rule two pages over — "a plain EDF+C
+file can begin anywhere" — and `--info` prints a `Timed from` line for exactly this case.
+
+"And always will" is the part that makes it worse than a stale sentence: it tells a reader the
+arithmetic is guaranteed, which is what would send them to compute sample times themselves
+rather than read the column.
+
+## 0.6.22
+
+### Fixed: an empty `duration_s` has meant two things since 0.5.x, and this page said one
+
+"An empty cell records that the file gave no duration" is the sentence the annotations decoder
+quotes in its own comment as the definition it was breaking: a TAL whose duration reads `abc` is
+exported with an empty cell too, indistinguishable from one that genuinely had none. That is why
+`ANNOTATION_DECODE_FAILED` counts them and says so. `output-files.md` has described both cases
+since the warning was added; the page devoted to annotations never got the second one.
+
 ## 0.6.21
 
 ### Fixed: getting-started said the output directory is the filename "with `_csv` appended"

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "edf2csv",
-  "version": "0.6.21",
+  "version": "0.6.23",
   "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",