Debrief · round 2 · 2026-11-04
The developer's 12:12 wake: the twelfth page check, and the online full book stops
Read first: publisher__006 (an info letter to the conductor, copied to me). The 25 figures a page after their paragraph are the layout's trade (LY1) and I am asked to keep reporting them, as information for the writers. Done in developer__046.
Page check on the PDF of 12:22 (142 pages, Appendix A in): 6 tails, 8 gaps (3 borderline), 26 figures a page after their paragraph. I read one new one at the page ends. Mailed to the publisher and designer (developer__046).
A stop: make_editions.py now refuses the full book, because Appendix A prints source addresses in its body (an address off the two hosts, outside a claim's register text). It wrote the Appendix A page before it refused (found by the post-review; the page was in the 12:25 push and is removed, and the builder now builds aside and writes only after every guard: developer__049), and it did not write editions.json, which stays at 11:18 UTC. I kept the rail and asked security (developer__047) with the options; the PDF and the printed book are not affected. Consequence for the page check: it matches the PDF against my 11:18 copy of the chapters, so tails and short pages are unaffected but figure counts may be (3 captions not found); the mail says so. I reverted the print-page map and the QR list the aborted run had partly rewritten, so the committed data files are from the 11:18 state except print-check.json (and the Appendix A page, see above).
Pages of the site: nothing new; no letter answers developer__039; the upload set is still unlinked (git log and grep). Not editing.
Left: security's answer; the last page check around 20:00 London (TG1).