Changes touching this path

• a land whose ferry fails no longer tells the reader to seal a working change that records deletions: ferry_remedy reads the working change as loot status does before it offers loot adopt --seal-wip, and over one that records a deletion it names the deleted paths, counting those sealed from the reader, offers no seal, and sends the reader to loot restore --source HEAD with each path, a plain loot adopt, then the re-land; over one it cannot read it says so and offers no seal either, and over one holding only edits, additions and moves the recovery is unchanged. the read is a disk walk paid only by a ferry that failed. kzknpprm (bd74b64) already stopped a key filed after a tree write from turning the paths it skipped into deletions, which answered the first defect. pinned: a working change of deletions, of edits alone and of both, one that cannot be read, and over a real lane the printed restore putting a deleted path back. red under named mutations, each restored: the deletions ignored (4 failed and 3 passed), no deletion ever found (3 and 4), every row read as a deletion (3 and 4), a sealed deletion named (1 and 6), the restore from @ (1 and 6), an unread working change offered the seal (1 and 6), and the land passing no deletions (1 and 6). cargo test green, 4912 passed over 144 test result lines with 15 ignored. CONTEXT.md says so, and says the sighting was in a landing lane where it said the primary. owes no deploy; lands print it once the release binaries of the primary are rebuilt (kmsukvmw) 5075031c · dbf3dbe6…

Renames are not followed. loot's tree maps a path to an address, so a rename is a delete and an add. This list is the history of the name, not of the bytes.