Pinky WWW: the UUU reject fix — kill-cause survives the terminal + the README line-start sweep (v1.79.0) #111
Loading…
Reference in a new issue
No description provided.
Delete branch "oss-oo/pinky-www-uureject-fix"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
REJECT remediation for the UUU verdict (2 MAJOR).
W1 — V4 fold-dead-canonical: the terminal-phase kill-cause gate keyed the breadcrumb against the FULL activity map, which includes the 🔨 free path relief write (completed_at = click time, always newer) — phase 2 dropped the cause phase 1 attached. The gate now keys on the PRE-TERMINAL basis (sessions_preterminal_activity_ms: message ∪ part ∪ turn started_at, blind to every completed_at): stable across the relief write, later-life-guarded by started_at. Two-phase rig drives the REAL seams (fixture DB, kill-log, real _free_orphaned_running_turn, fetch_sessions twice with the facts memo cleared — the restart stand-in); it reproduces the verdict measurement on pre-fix code and retains the cause on the fix. No-false-cause controls: no breadcrumb, earlier-life kill, genuine owner cancel.
W2 — V5 MD018 README:1639: the U2 heading wrap put "#14).** …" at a line start (ATX shape). Fixed the reflow; swept README + docs/ROADMAP.md for the class (found + fixed one more of the same class in ROADMAP). Station lint recipe (six lanes) green on the final committed tree, with an MD018 tripwire proving the lane fires.
G3: full suite 1233 OK (1 pre-existing skip); version pins bumped; v1.79.0.