Nothing enters the catalog without a usable name: the pending store (decision f amended)

An unnamed release (no readable file name) is written to pending_releases
(migration 6300, the HDD) with its keys, counts, article span, upload tool
and probe targets, instead of releases, release_texts, posters and a chunk.
The edge sends it with an empty payload and its tool; the batch store splits
a batch by readable. The pending ladder (probes cheapest first, signature
verdicts, the move into the catalog) comes next. The replay report shows the
pending releases.

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>

REBUILD.md

@@ -550,6 +550,13 @@ releases; the inspector matches them again), `nntp_shares`, `nodes`,
   `release_texts`. The category trigger reads the groups from
   `group_sets`; the judge's dirty trigger watches `poster_id` instead of
   `poster`.
+- `pending_releases` (migration 6300, the HDD): the unnamed releases
+  outside the catalog (decision f): `release_key` (primary), `identity_key`,
+  `group_id`, `name` (the random stem), `poster` text, the posting span,
+  `first_article`, `last_article`, counts, `partial`, `finished`, `tool`,
+  `candidates` jsonb (the probe targets by effort), `status` (0 waiting, 1
+  named, 2 discarded), `tries`, `created_at`, `probed_at`; indexed by
+  (group, tool, status, created_at) for the ladder.
 - `release_texts(release_id, kind smallint, text varchar(255))` (6200),
   primary key on all three, one trigram index (fastupdate): every string
   search matches against, appended by whoever learns one and never an
@@ -682,10 +689,12 @@ until the window is done (LOG.md 2026-10-26 (111) has the first local
 numbers, (112) the homelab run).
 
 Next: the replay on homelab over the 2016 window (the NVMe bytes per
-header for real), the ordered range leases with the carried window in
+header for real), the pending ladder and the signature learner (decision
-the edge-node and its snapshot (design 1 and 5), the prober writing
+f, amended 2026-10-26: the prober working `pending_releases` cheapest
-`release_texts` kind 2, random-named singles without a chunk (decision f:
+first, verdicts per tool and group, a named release moved into the
-2.3 M of the first replay's 4.0 M singles are "Bob"'s 32-hex posts). Gone from the list: block
+catalog with its segments read again), the ordered range leases with the
+carried window in the edge-node and its snapshot (design 1 and 5), the
+prober writing `release_texts` kind 2. Gone from the list: block
 leases, the planner and the block merger (design 1, amended).
 
 ## Decisions
@@ -697,11 +706,33 @@ c. Group list: editable, owner by priority number (above). **Accepted.**
 d. Reposts: separate releases in phase 1, merged in phase 2. **Accepted.**
 e. Headers through the coordinator's relay, with prefetch (part 1).
    **Accepted.**
-f. Random-named releases (no file with a readable name) keep no chunk in
+f. **Nothing enters the catalog without a usable name** (amended
-   the header pass: their file keys and probe candidates only, until
+   2026-10-26, the user's direction; before: random-named releases keep
-   phase 2 names them (bodies) and fetches their segments again. The full
+   no chunk but their rows). A release the window closed on whose files
-   history is ~3 TB with them, ~1.2–1.4 TB without (725 M of the 1.25 B
+   all have random names is no release yet: it goes to `pending_releases`
-   kept files; LOG.md 2026-10-09). **Accepted** 2026-10-09, not built yet.
+   (migration 6300, the HDD) with its keys, counts, article span, upload
+   tool and probe targets, no row in `releases`, `release_texts` or
+   `posters`, no chunk. The pending ladder works them **from least to most
+   effort**: (0) no fetch: a signature verdict (tool and group, poster)
+   learned from earlier probes, a spot or NZB that names the message-ids;
+   (1) a par2 main file's first article (it names the files); (2) the first
+   article of the first file (the yEnc name, an archive's listing in its
+   first 16 KiB); (3) a 7z's tail; then nothing more. A probe that yields
+   a usable name moves the release into the catalog, its segments read
+   again from the corpus (the range stored) or the server; one that
+   yields none counts against its signature, and a signature with
+   ≥100 probes and ≤0.5% named (voyager 2's garbage learner, LOG.md
+   2026-10-08 (5)) is **discarded** from then on, a sample of it still
+   probed to notice a change. The reason, in the 2016 replay: one poster's
+   hash-named pieces ("Bob", 11 aliases, 3.2 M releases in five small
+   groups) were 76% of the rows and three quarters of the NVMe bytes for
+   11% of the segments and nothing anyone can find; the full history has
+   725 M random-named files (LOG.md 2026-10-09). Built 2026-10-26: the
+   edge sends an unnamed release with an empty payload and its tool, the
+   batch store writes `pending_releases` (LOG.md (114)). Not built: the
+   ladder itself (the prober working `pending_releases` cheapest first,
+   the signature verdicts, the move into the catalog with the segments
+   read again), and the spot/NZB match on pending keys.
 g. The account's NNTP connections shared by priority (`nntp_shares`): the
    coordinator (headers) before the prober, each between its minimum and
    maximum, 5 kept for the spotter, the spot inspector, the plain

app/services/batch_store.ts

@@ -8,6 +8,7 @@ import { decodeChunk } from '#lib/chunks/index'
 import { SIDECAR } from '#services/sidecar_folder'
 import { identityKey, releaseKey } from '#lib/release/keys'
 import { sourceCode } from '#lib/release/source'
+import { uploadTool } from '#lib/rules/index'
 import { TEXT_KINDS, TEXT_MAX, textsToStore } from '#lib/release/text_kind'
 
 /**
@@ -30,6 +31,8 @@ export interface StoredBatch {
   /** Provisional (tail) releases the batch finished. */
   finalized: number
   candidates: number
+  /** Unnamed releases written to the pending store instead of the catalog (decision f). */
+  pending: number
 }
 
 /** A chunk's slot in its batch: the key and the upload the edge numbered it with (ProbeCandidate points at it). */
@@ -73,18 +76,29 @@ class BatchStore {
         newReleases: 0,
         finalized: 0,
         candidates: 0,
+        pending: 0,
       }
 
+      // Nothing enters the catalog without a usable name (decision f): an unnamed release (no readable file name)
+      // goes to the pending store with its probe targets, the named ones to the catalog.
+      const named = batch.releases.filter((r) => r.readable)
+      const unnamed = batch.releases.filter((r) => !r.readable)
       const written = await this.writeReleases(
         trx,
         { groupId: Number(range.group_id), blockId },
         batch.from,
-        batch.releases,
+        named,
         batch.candidates
       )
       result.newReleases = written.newReleases
       result.finalized = written.finalized
       result.candidates = written.candidates
+      result.pending = await this.writePending(
+        trx,
+        Number(range.group_id),
+        unnamed,
+        batch.candidates
+      )
 
       if (batch.counts.length > 0) {
         await trx.rawQuery(
@@ -153,6 +167,68 @@ class BatchStore {
     return Number(block.id)
   }
 
+  /**
+   * The unnamed releases of a batch into pending_releases (migration 6300, the HDD): keys, counts, the upload
+   * tool and the probe targets, no chunk. The pending ladder (REBUILD.md decision f) probes them cheapest first,
+   * and a name moves one into the catalog; a provisional tail release sent again adds its counts up.
+   */
+  async writePending(
+    trx: TransactionClientContract,
+    groupId: number,
+    releases: ReleaseChunk[],
+    candidates: ProbeCandidate[]
+  ): Promise<number> {
+    if (releases.length === 0) return 0
+    const targets = new Map<string, { kind: string; messageId: string; file: string }[]>()
+    for (const c of candidates) {
+      const key = slot(c.releaseKey, c.upload ?? 0)
+      const list = targets.get(key) ?? []
+      list.push({ kind: c.kind, messageId: c.messageId, file: c.file })
+      targets.set(key, list)
+    }
+    let written = 0
+    for (let i = 0; i < releases.length; i += 1000) {
+      const rows = releases.slice(i, i + 1000).map((r) => {
+        const probes = targets.get(slot(r.key, r.upload)) ?? []
+        return {
+          release_key: releaseKey(r.key, groupId, r.firstArticle).toString('hex'),
+          identity_key: identityKey(r.key).toString('hex'),
+          name: r.name.slice(0, TEXT_MAX),
+          poster: r.poster || null,
+          first: r.postedAt,
+          last: r.postedLast ?? r.postedAt,
+          first_article: r.firstArticle,
+          last_article: r.lastArticle,
+          files: r.files,
+          segments: r.segments,
+          bytes: r.bytes,
+          partial: r.partial === true,
+          finished: r.finished !== false,
+          tool: r.tool ?? (probes[0] ? uploadTool(probes[0].messageId) : 'unknown'),
+          candidates: probes,
+        }
+      })
+      const { rows: done } = await trx.rawQuery(
+        `INSERT INTO pending_releases (release_key, identity_key, group_id, name, poster, posted_at, last_posted_at, first_article, last_article,
+                                       files, segments, bytes, partial, finished, tool, candidates)
+         SELECT decode(v.release_key, 'hex'), decode(v.identity_key, 'hex'), ?, v.name, v.poster, to_timestamp(v.first / 1000.0), to_timestamp(v.last / 1000.0),
+                v.first_article, v.last_article, v.files, v.segments, v.bytes, v.partial, v.finished, v.tool, v.candidates
+         FROM jsonb_to_recordset(?::jsonb) AS v(release_key text, identity_key text, name text, poster text, first bigint, last bigint, first_article bigint, last_article bigint,
+                                               files int, segments bigint, bytes bigint, partial boolean, finished boolean, tool text, candidates jsonb)
+         ON CONFLICT (release_key) DO UPDATE SET
+           files = pending_releases.files + excluded.files, segments = pending_releases.segments + excluded.segments, bytes = pending_releases.bytes + excluded.bytes,
+           last_posted_at = greatest(pending_releases.last_posted_at, excluded.last_posted_at), last_article = greatest(pending_releases.last_article, excluded.last_article),
+           partial = excluded.partial, finished = excluded.finished,
+           candidates = CASE WHEN jsonb_array_length(pending_releases.candidates) = 0 THEN excluded.candidates ELSE pending_releases.candidates END
+         WHERE NOT pending_releases.finished
+         RETURNING release_key`,
+        [groupId, JSON.stringify(rows)]
+      )
+      written += done.length
+    }
+    return written
+  }
+
   /**
    * The releases of a batch into the catalog, one chunk each, and their probe candidates. Every chunk is one
    * release: inserted, or (a provisional tail release sent again) finished.

commands/replay_corpus.ts

@@ -333,6 +333,11 @@ export default class ReplayCorpus extends BaseCommand {
       `SELECT source, count(*)::bigint AS n FROM releases WHERE group_id = ? GROUP BY source ORDER BY source`,
       [groupId]
     )
+    const { rows: pending } = await db.rawQuery(
+      `SELECT count(*)::bigint AS n, coalesce(sum(segments), 0)::bigint AS segments, count(*) FILTER (WHERE jsonb_array_length(candidates) = 0)::bigint AS unprobable
+       FROM pending_releases WHERE group_id = ?`,
+      [groupId]
+    )
     const { rows: headerRows } = await db.rawQuery(
       `SELECT coalesce(sum(headers), 0)::bigint AS headers FROM ranges WHERE id = ANY(?::bigint[])`,
       [ids]
@@ -343,12 +348,17 @@ export default class ReplayCorpus extends BaseCommand {
               sum(pg_total_relation_size(c.oid))::bigint AS bytes
        FROM pg_class c JOIN pg_namespace n ON n.oid = c.relnamespace
        WHERE n.nspname = 'public' AND c.relkind = 'r'
-         AND (c.relname IN ('releases', 'release_texts', 'posters', 'group_sets', 'probe_candidates', 'rule_counts', 'rule_samples', 'blocks')
+         AND (c.relname IN ('releases', 'release_texts', 'posters', 'group_sets', 'probe_candidates', 'rule_counts', 'rule_samples', 'blocks', 'pending_releases')
               OR c.relname LIKE 'release_chunks%')
        GROUP BY 1 ORDER BY 2 DESC`
     )
     // On CHUNK_TABLESPACE (the HDD) in the catalog: the chunks (migration 3200), inspections and candidates (6000).
-    const hdd = new Set(['release_chunks', 'probe_candidates', 'release_inspections'])
+    const hdd = new Set([
+      'release_chunks',
+      'probe_candidates',
+      'release_inspections',
+      'pending_releases',
+    ])
     let nvme = 0
     let other = 0
     const c = counts[0]
@@ -364,6 +374,9 @@ export default class ReplayCorpus extends BaseCommand {
     this.logger.info(
       `By source: ${sources.map((s: any) => `${s.source}: ${fmt(s.n)}`).join(', ')}.`
     )
+    this.logger.info(
+      `Pending (unnamed, outside the catalog): ${fmt(pending[0].n)} releases, ${fmt(pending[0].segments)} segments, ${fmt(pending[0].unprobable)} without a probe target.`
+    )
     for (const s of sizes) {
       const bytes = Number(s.bytes)
       if (hdd.has(s.name)) other += bytes

database/migrations/1770000006300_pending_releases.ts

@@ -0,0 +1,60 @@
+import { BaseSchema } from '@adonisjs/lucid/schema'
+import env from '#start/env'
+
+/**
+ * pending_releases (REBUILD.md decision f, 2026-10-26): nothing enters the catalog without a usable name. A release
+ * the window closed on whose files all have random names is written here instead of releases: its keys (as the
+ * catalog would store them), its counts and article span, the upload tool from its message-id (with the group its
+ * signature), and its probe targets by effort (the edge's candidates: a par2, an archive start, a lone file, a 7z
+ * tail), no chunk. The pending ladder probes them cheapest first and moves a named one into the catalog, reading its
+ * segments again from the corpus or the server; a signature that names nothing is discarded. On CHUNK_TABLESPACE
+ * (the HDD): written once, read by the ladder, never by search.
+ */
+export default class extends BaseSchema {
+  async up() {
+    this.schema.raw(`
+      CREATE TABLE pending_releases (
+        -- lib/release/keys.ts: the identity, the group and the first article; the identity alone.
+        release_key bytea PRIMARY KEY,
+        identity_key bytea NOT NULL,
+        group_id integer NOT NULL,
+        -- The stem or post the edge keyed it by (a random name: no release name, but the inner-stem merge's handle).
+        name varchar(255) NOT NULL,
+        poster text,
+        posted_at timestamptz,
+        last_posted_at timestamptz,
+        first_article bigint NOT NULL,
+        last_article bigint NOT NULL,
+        files integer NOT NULL DEFAULT 0,
+        segments bigint NOT NULL DEFAULT 0,
+        bytes bigint NOT NULL DEFAULT 0,
+        -- A file misses parts.
+        partial boolean NOT NULL DEFAULT false,
+        -- False: a tail release still open; a later chunk of its key adds its counts up.
+        finished boolean NOT NULL DEFAULT true,
+        -- lib/rules uploadTool: ngPost, nyuu, hex32, random, other, …
+        tool varchar(16) NOT NULL,
+        -- [{kind, messageId, file}] by effort (lib/edge ProbeCandidate); empty: nothing to read, only a signature.
+        candidates jsonb NOT NULL DEFAULT '[]',
+        -- 0 waiting, 1 named (in the catalog), 2 discarded (its signature's verdict, or every probe gave nothing).
+        status smallint NOT NULL DEFAULT 0,
+        tries smallint NOT NULL DEFAULT 0,
+        created_at timestamptz NOT NULL DEFAULT now(),
+        probed_at timestamptz
+      )`)
+    // The ladder's order: a group's and tool's waiting releases, oldest first.
+    this.schema.raw(
+      `CREATE INDEX pending_releases_ladder_idx ON pending_releases (group_id, tool, status, created_at)`
+    )
+    const hdd = env.get('CHUNK_TABLESPACE')
+    if (hdd) {
+      this.schema.raw(`ALTER TABLE pending_releases SET TABLESPACE ${hdd}`)
+      this.schema.raw(`ALTER INDEX pending_releases_pkey SET TABLESPACE ${hdd}`)
+      this.schema.raw(`ALTER INDEX pending_releases_ladder_idx SET TABLESPACE ${hdd}`)
+    }
+  }
+
+  async down() {
+    this.schema.raw(`DROP TABLE IF EXISTS pending_releases`)
+  }
+}

edge/src/indexer.rs

@@ -53,6 +53,11 @@ pub struct ReleaseChunk {
     pub edge: bool,
     /// A file misses parts.
     pub partial: bool,
+    /// The upload tool from the first article's message-id (names::upload_tool), for an unnamed release: with the
+    /// group its signature, what the pending ladder learns verdicts by. None for a named one.
+    pub tool: Option<&'static str>,
+    /// The files' names and segments (codec); empty for an unnamed release (decision f: nothing in the catalog
+    /// without a usable name; the pending store keeps its keys and probe targets until a probe names it).
     pub payload: Vec<u8>,
     /// The first files' subjects (at most SUBJECTS): edge-bench's --releases dump for the grouping sweeps.
     pub subjects: Vec<String>,
@@ -1154,17 +1159,26 @@ impl RangeIndexer {
                 let mut chunk = chunk_of(key, &list, upload);
                 chunk.finished = finished;
                 chunk.edge = edge;
+                let album_none = album.is_none();
                 if let Some(name) = album {
                     chunk.name = name;
                     chunk.source = "album";
                     chunk.file_total = None;
-                } else if let Some(total) = chunk.file_total.filter(|&t| t >= BATCH_COUNTER_MIN_TOTAL && (chunk.files as i64) < t) {
+                }
+                if !chunk.readable {
+                    // Unnamed: no chunk (its segments are read again once a probe names it), the tool for its signature.
+                    chunk.tool = Some(names::upload_tool(&list[0][0].message_id));
+                    chunk.payload = encode_chunk(&[]);
+                }
+                if chunk.readable {
+                    self.pending_files += chunk.files;
+                }
+                if let Some(total) = chunk.file_total.filter(|&t| album_none && t >= BATCH_COUNTER_MIN_TOTAL && (chunk.files as i64) < t) {
                     // A batch counter: no file total (the release is as complete as its files are).
                     if self.batch_totals.get(&(chunk.poster.clone(), total)).is_some_and(|(n, _)| *n >= BATCH_COUNTER_RELEASES) {
                         chunk.file_total = None;
                     }
                 }
-                self.pending_files += chunk.files;
                 self.batch.releases.push(chunk);
                 self.batch.candidates.extend(candidates);
             }
@@ -1267,6 +1281,7 @@ fn chunk_of(key: &str, files: &[Vec<Seg>], upload: usize) -> ReleaseChunk {
         finished: true,
         edge: false,
         partial: files.iter().any(|f| (f.len() as u32) < f[0].file.total_parts),
+        tool: None,
         payload: encode_chunk(&chunk_files),
         subjects: chunk_files.iter().take(SUBJECTS).map(|f| f.subject.to_string()).collect(),
     }
@@ -1411,6 +1426,23 @@ mod tests {
         assert!(ix.albums.is_empty());
     }
 
+    #[test]
+    fn an_unnamed_release_keeps_no_chunk_but_its_tool_and_probe_target() {
+        let mut priorities = FxHashMap::default();
+        priorities.insert(GROUP.to_string(), 1);
+        let mut ix = RangeIndexer::new(GROUP, 1, 10_000, 1, Rules::defaults(priorities), 5);
+        let lines: Vec<String> = (0..4).map(|i| line(100 + i, &format!(r#"[1/1] - "q3B9OrpKiCxUnzgtUGnp.part01.rar" yEnc ({}/4)"#, i + 1), i)).collect();
+        ix.add_text(&lines.join("\n"), 10_001);
+        ix.finish();
+        let batch = ix.take();
+        assert_eq!(batch.releases.len(), 1);
+        let r = &batch.releases[0];
+        assert!(!r.readable && r.tool == Some("other") && r.files == 1 && r.segments == 4);
+        assert_eq!(r.payload, encode_chunk(&[]));
+        assert_eq!(batch.candidates.iter().map(|c| c.kind).collect::<Vec<_>>(), vec!["archive"]);
+        assert_eq!(ix.pending(), 0);
+    }
+
     #[test]
     fn album_templates_and_names() {
         assert_eq!(album_template("1481439079025_014.jpg.data").as_deref(), Some("#_#.jpg.data"));

edge/src/wire.rs

@@ -27,6 +27,8 @@ struct WireRelease<'a> {
     finished: bool,
     edge: bool,
     partial: bool,
+    #[serde(skip_serializing_if = "Option::is_none")]
+    tool: Option<&'a str>,
     payload: String,
     #[serde(skip_serializing_if = "Option::is_none")]
     rule: Option<&'a str>,
@@ -87,6 +89,7 @@ fn release<'a>(r: &'a ReleaseChunk, rule: Option<(&'a str, &'a str)>) -> WireRel
         finished: r.finished,
         edge: r.edge,
         partial: r.partial,
+        tool: r.tool,
         payload: base64::engine::general_purpose::STANDARD.encode(&r.payload),
         rule: rule.map(|(rule, _)| rule),
         rule_key: rule.map(|(_, key)| key),

lib/edge/range_indexer.ts

@@ -35,6 +35,11 @@ export interface ReleaseChunk {
   edge?: boolean
   /** A file of the release misses parts (the window closed on it): not complete, whatever the counter says. */
   partial?: boolean
+  /**
+   * Voyager 3, an unnamed release (readable false): the upload tool from its first article's message-id, its
+   * signature with the group; the payload is empty (no chunk until a probe names it: pending_releases).
+   */
+  tool?: string
   payload: Buffer
 }
 

tests/functional/batch_store.spec.ts

@@ -181,7 +181,13 @@ test.group('BatchStore', (group) => {
     )
     const first = ix.take()
     const stored = await BatchStore.store(rangeId, nodeId, first)
-    assert.deepEqual(stored, { releases: 1, newReleases: 1, finalized: 0, candidates: 2 })
+    assert.deepEqual(stored, {
+      releases: 1,
+      newReleases: 1,
+      finalized: 0,
+      candidates: 2,
+      pending: 0,
+    })
 
     const release = await db.from('releases').first()
     assert.include(release, {
@@ -487,6 +493,51 @@ test.group('BatchStore', (group) => {
     assert.notEqual(rows[0].release_key, rows[1].release_key)
   })
 
+  test('an unnamed release goes to the pending store with its probe targets, not into the catalog', async ({
+    assert,
+  }) => {
+    const { rangeId, nodeId } = await rangeIn(GROUP, 1)
+    const batch = batchOf(GROUP, 1, [
+      ...upload(10, 'q3B9OrpKiCxUnzgtUGnp.part01.rar', 4, '[1/2]', '2026-10-05T10:00:00Z'),
+      ...upload(20, 'q3B9OrpKiCxUnzgtUGnp.par2', 1, '[2/2]', '2026-10-05T10:01:00Z'),
+    ])
+    assert.lengthOf(batch.releases, 1)
+    assert.isFalse(batch.releases[0].readable)
+    const stored = await BatchStore.store(rangeId, nodeId, batch)
+    assert.include(stored, { releases: 1, newReleases: 0, candidates: 0, pending: 1 })
+    assert.equal(Number((await db.from('releases').count('* as n').first())!.n), 0)
+    assert.equal(Number((await db.from('probe_candidates').count('* as n').first())!.n), 0)
+    assert.equal(Number((await db.from('release_chunks').count('* as n').first())!.n), 0)
+    const pending = await db.from('pending_releases').firstOrFail()
+    assert.include(pending, {
+      files: 2,
+      segments: '5',
+      partial: false,
+      finished: true,
+      tool: 'other',
+      status: 0,
+      poster: 'p <p@x>',
+    })
+    assert.lengthOf(pending.release_key, 20)
+    assert.deepEqual(
+      pending.candidates.map((c: { kind: string; file: string }) => [c.kind, c.file]).sort(),
+      [
+        ['archive', 'q3B9OrpKiCxUnzgtUGnp.part01.rar'],
+        ['par2', 'q3B9OrpKiCxUnzgtUGnp.par2'],
+      ]
+    )
+    // Sent again as a later chunk of a tail release: its counts add up while it is open, not once finished.
+    const again = await BatchStore.store(rangeId, nodeId, {
+      ...batch,
+      from: batch.next,
+      next: batch.next + 1000,
+      releases: [{ ...batch.releases[0], files: 1, segments: 5, bytes: 100 }],
+      candidates: [],
+    })
+    assert.include(again, { pending: 0 })
+    assert.include(await db.from('pending_releases').firstOrFail(), { files: 2, segments: '5' })
+  })
+
   test("an album's members stay findable by their own names (release_texts kind member)", async ({
     assert,
   }) => {