Official release dates from MangaUpdates; peers lower ecosystem dates (BL-0048, BL-0049)

- scraper/metadata/release-dates.js fills x-dates.release from MangaUpdates
  publisher releases. MangaUpdates also lists English back-catalog re-releases,
  so a date is kept only inside the AniList original run, when no scanlation or
  source upload is more than 4 days earlier, and when it fits the release order
  (no lower chapter dated after a higher one). Seasons that restart chapter
  numbers are skipped; dates this job wrote and no longer accepts are withdrawn.
- The scraper refreshes ongoing series nightly before its update checks, only
  while no download runs, and bumps the publish revision so the reader reconciles.
- tools/release-dates.js for a whole library (dry run default). Production dry
  run: 728 chapters dated (Kagurabachi, JJK, Chainsaw Man, One Piece from 928).
- Peer sync: chapters this library already has adopt a peer's earlier ecosystem
  date (job result ecosystemLowered); the peer listing now carries dates.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

docs/library-format.md

@@ -229,8 +229,22 @@ An AniList refresh with a different ID is stopped for review before chapter publ
 - `source`: when the source site uploaded this version (WeebCentral's full chapter list,
   MangaDex `publishAt`). `batch: true` means it arrived with 3 or more other uploads
   less than 10 minutes apart, which is a site import, not a release.
-- `release`: the official public release (`precision` `day`, `month` or `year`). Not
-  filled yet; a metadata job will add it.
+- `release`: the official public release (`precision` `day`, `month` or `year`). It is
+  filled from MangaUpdates releases by publishers such as Viz, MANGA Plus or Webtoon
+  (`scraper/metadata/release-dates.js`). Those lists include English re-releases of old
+  chapters, so a date is kept only when all of these hold:
+  - it lies within the work's original run on AniList;
+  - no scanlation or source upload shows the chapter more than 4 days earlier;
+  - it fits the release order: walking down from the newest chapter, a lower chapter
+    can't be dated more than 2 days after a higher one.
+
+  Series whose chapter numbers restart in a later book are skipped.
+  - The scraper refreshes ongoing series nightly before its update checks, but only
+    while no download is running.
+  - `node tools/release-dates.js <library>` covers a whole library: a dry run by
+    default, then `--apply --writers-stopped`.
+  - A date that this job wrote and no longer accepts is removed. Release dates from any
+    other source stay.
 - `ecosystem`: when the chapter first entered any Manga Circus library. Peer transfers
   carry it, and the earliest known date always wins. A peer's claim can't be earlier than
   the source upload or later than now.

docs/peering.md

@@ -118,7 +118,10 @@ date at which any version of that chapter number entered a Manga Circus library
 to the sender. Older chapters fall back to their `added` date. The receiver stores the
 date as `x-dates.ecosystem` (`via: "peer <alias>"`) and only lets it lower an existing
 date, never raise it. See `docs/library-format.md`. Older peers omit `dates`, and the
-receiver then uses its own arrival time.
+receiver then uses its own arrival time. A sync also lowers the ecosystem date of
+chapters this library already has, including ones it scraped itself, when the peer
+knows an earlier date for the same book and chapter number. The job reports these as
+`ecosystemLowered`. The reader shows the new date after its next full reconciliation.
 | `GET /peer/v1/blobs/{id}` | One image, optional `Range: bytes=N-` and `If-Match`; 206 on resume |
 
 Unknown/unshared IDs return 404. Revoked serving returns 403; bad/missing grants