LH034 checkpoint, 9 November 2026
In short. On day 18 after Node.js's fix (8 November), 61.8 per cent of line 22's downloads from nodejs.org were at or above the fixed release, 68.2 per cent of line 24's and 94.4 per cent of line 26's, close to where they stood on day 14; 13.2 per cent came from lines the release does not fix. Node.js's site asks automated readers not to fetch its release list, so from today the study uses the copy it kept on 5 November. Working record, unreviewed.
Read data/read-2026-11-09T1428Z, started 14:28 UTC: the summaries for 5 to 8 November (days 15 to 18), all HTTP 200. Tables: data/tables.json. Written by the scheduled Monday run.
The release list was not fetched (robots.txt)
This was the first read under the keeper's robots.txt rule of 8 November 2026 (scripts/lh_fetch.py). nodejs.org's robots.txt (HTTP 200, sha256 beginning cbcd5e7d85caf581) has Disallow: /dist/ in its * group, so https://nodejs.org/dist/index.json was not fetched and is logged "not fetched". Every earlier read of this study fetched that file; whether the robots.txt said the same on those days was not checked, and those requests are for the rules-of-the-road re-audit (row LH047-reread). The download summaries come from storage.googleapis.com, whose robots.txt allows them.
By the brief's amendment of 9 November, tabulate.py now takes the release list from the newest read that holds it (read-2026-11-05T1414Z, fetched 14:14 UTC on 5 November) and counts separately any version in a day's downloads that the list does not hold. For days 15 to 18 no version in lines 20, 22, 24, 25 or 26 is missing from the list, so no new release in those lines had downloads; the only unlisted version with many downloads is v27.0.0 (79,169 on day 18; the list holds only v27.0.0-alpha.0.0.0), outside the measured lines. Whether Node.js published a new release since 5 November is therefore unknown from the list itself, and would show first as unlisted downloads.
Measure 1: share of the day's line downloads at or above the fixed release (observation, nodejs.org requests only)
| Line | Day 14 | Day 15 | Day 16 | Day 17 | Day 18 | Line downloads, day 18 |
|---|---|---|---|---|---|---|
| 22 | 62.6% | 59.9% | 61.8% | 60.9% | 61.8% | 4.8 million |
| 24 | 69.6% | 69.4% | 70.2% | 67.7% | 68.2% | 4.3 million |
| 26 | 95.0% | 94.7% | 95.2% | 94.9% | 94.4% | 4.3 million |
Days 0 to 14 are unchanged. Days 17 and 18 are a weekend, with about 15.6 million attributed requests a day against 27 to 32 million on the weekdays. The shares have been roughly flat since day 13. No cause is read.
Measure 2: the unfixed tail
Share of version-attributed downloads from lines the release does not fix: 14.7 per cent on day 15, 13.0 on day 16, 12.5 on day 17, 13.2 on day 18 (12.6 on day 14). Line 20: 8.9, 6.7, 6.2, 6.9 per cent; line 25: 0.7 to 0.8. Requests with no version: 0.01 per cent of total or less on each day.
Not done
Measure 4 and its comparators are unchanged by these days. The brief's day-30 read, on 22 November, also re-reads days 0 to 7 once.
Not flagged for review: nothing here changes a public account.
From studies/LH034/checkpoints/2026-11-09.md in the repository, last changed 9 November 2026.