Rebuild: re-uploads are releases of their own

A poster re-uploading under the same name made one release of all
uploads: in a sample, 24% of releases with 2+ files had a file name more
than once (4.3% of files extra copies), 70% of those copies posted a day
or more apart; their NZBs listed each copy.

- File assembly keeps a re-upload's copy apart: a part already in every
  open copy, posted more than 3 h from them, starts another copy; a
  missing part fills the nearest copy (a re-send of what was missing),
  a re-sent part near a copy that has it is ignored as before.
- A key's files in a batch split into uploads where posting pauses for
  more than 3 h, counted from the latest article posted so far (a file
  that takes hours to post doesn't split it); each upload is judged and
  chunked on its own (ReleaseChunk.upload, .postedLast; candidates carry
  their upload).
- The catalog matches a chunk to the release of its key whose articles
  the range's group owns and whose posting window is within 3 h, else
  starts one (key, key:2, …): releases get identity_key, group_id,
  last_posted_at. Keys are locked in order (advisory locks), so
  concurrent batches neither deadlock nor both start one; the merger
  writes in key order too. The same upload multi-posted to other groups
  is a release per group.
- The gap: 91% of uploads post within 3 h; 73% of re-uploaded copies
  come later (releases:upload-gaps). releases:duplicates measures the
  duplicate names.

Both implementations; check.ts compares the windows and uploads too
(identical on all captures; Rust tests mirror the TypeScript cases).

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

lib/edge/index.ts

@@ -1,4 +1,4 @@
-export { RangeIndexer, chunkOf, candidatesOf, releaseKeyOf, type EdgeBatch, type ReleaseChunk, type DroppedChunk, type ProbeCandidate, type CandidateKind, type RuleCount, type Fragment, type RangeIndexerOptions } from './range_indexer.js'
+export { RangeIndexer, chunkOf, candidatesOf, releaseKeyOf, uploadsOf, UPLOAD_GAP_MS, type EdgeBatch, type ReleaseChunk, type DroppedChunk, type ProbeCandidate, type CandidateKind, type RuleCount, type Fragment, type RangeIndexerOptions } from './range_indexer.js'
 export { encodeFrame, decodeFrames, isKeepalive, KEEPALIVE, type Frame } from './stream.js'
 export { toWire, fromWire, type WireBatch } from './wire.js'
 export type { Lease } from './lease.js'

lib/edge/range_indexer.ts

@@ -8,12 +8,16 @@ import { encodeChunk, type ChunkFile } from '../chunks/index.js'
 /** A release's files from one batch (REBUILD.md, chunk store): what the catalog adds up, and the encoded chunk. */
 export interface ReleaseChunk {
   key: string
+  /** Which upload of the key in this batch (0, 1, …; uploadsOf): the catalog keeps uploads apart. */
+  upload: number
   name: string
   source: string
   fileTotal: number | null
   poster: string
   groups: string[]
   postedAt: number | null
+  /** The latest article's date: with postedAt, the upload's posting window. */
+  postedLast: number | null
   files: number
   segments: number
   bytes: number
@@ -35,6 +39,8 @@ export type CandidateKind = 'archive' | 'par2' | 'lone' | '7z-tail'
 /** An article phase 2 should read (REBUILD.md, probe queue). */
 export interface ProbeCandidate {
   releaseKey: string
+  /** Its chunk's upload (ReleaseChunk.upload). */
+  upload: number
   kind: CandidateKind
   messageId: string
   file: string
@@ -106,7 +112,7 @@ const MAX_CANDIDATES = 4
  * waiting for parts.
  */
 export class RangeIndexer {
-  private aggregator = new FileAggregator<EdgeSegment>()
+  private aggregator = new FileAggregator<EdgeSegment>(UPLOAD_GAP_MS)
   private scanner = new SegmentScanner()
   private segments: EdgeSegment[] = []
   private rules: RuleSet
@@ -194,22 +200,24 @@ export class RangeIndexer {
     }
   }
 
-  /** Judges and encodes the releases completed in this batch. */
+  /** Judges and encodes the releases completed in this batch: each upload of a key on its own. */
   private judgeReleases() {
     const byRelease = this.completed
     this.completed = new Map()
-    for (const [key, list] of byRelease) {
-      const verdict = this.rules.judge(
-        key,
-        list.map((f) => ({ name: f.segments[0].file.name, totalParts: f.segments[0].totalParts, groups: f.segments[0].file.groups, firstMessageId: f.segments[0].messageId }))
-      )
-      if (verdict.rule === null) {
-        this.batch.releases.push(chunkOf(key, list))
-        this.batch.candidates.push(...candidatesOf(key, list))
-        continue
-      }
-      this.count(verdict.rule, verdict.key, list.reduce((n, f) => n + f.segments.length, 0), list.length)
-      if (verdict.sample) this.batch.samples.push({ ...chunkOf(key, list), rule: verdict.rule, ruleKey: verdict.key })
+    for (const [key, all] of byRelease) {
+      uploadsOf(all).forEach((list, upload) => {
+        const verdict = this.rules.judge(
+          key,
+          list.map((f) => ({ name: f.segments[0].file.name, totalParts: f.segments[0].totalParts, groups: f.segments[0].file.groups, firstMessageId: f.segments[0].messageId }))
+        )
+        if (verdict.rule === null) {
+          this.batch.releases.push(chunkOf(key, list, upload))
+          this.batch.candidates.push(...candidatesOf(key, list, upload))
+          return
+        }
+        this.count(verdict.rule, verdict.key, list.reduce((n, f) => n + f.segments.length, 0), list.length)
+        if (verdict.sample) this.batch.samples.push({ ...chunkOf(key, list, upload), rule: verdict.rule, ruleKey: verdict.key })
+      })
     }
   }
 
@@ -284,6 +292,51 @@ export function releaseKeyOf(s: EdgeSegment): string {
   return s.file.release?.key ?? hash('sha1', `file\n${s.fileKey}`)
 }
 
+/**
+ * A pause of more than this in posting starts another upload of the same
+ * release key (a re-upload): 91% of uploads post within 3 h, 73% of
+ * re-uploaded copies come later (LOG.md 2026-10-09 (10)).
+ */
+export const UPLOAD_GAP_MS = 3 * 60 * 60 * 1000
+
+/**
+ * A key's files split into uploads where posting pauses for more than
+ * UPLOAD_GAP_MS: files in order of their first article's date; a file
+ * starts another upload when it begins more than UPLOAD_GAP_MS after the
+ * latest article posted so far (so a file that takes hours to post doesn't
+ * split its upload). Earliest upload first, each in the files' own order;
+ * files without a date go with the first.
+ */
+export function uploadsOf<F extends AggregatedFile<EdgeSegment>>(files: F[]): F[][] {
+  if (files.length < 2) return [files]
+  const first = files.map((f) => postedOf(f.segments))
+  const last = files.map((f) => latestOf(f.segments))
+  const dated = files.map((_, i) => i).filter((i) => first[i] !== null).sort((a, b) => first[a]! - first[b]! || a - b)
+  if (dated.length < 2) return [files]
+  const upload = new Array<number>(files.length).fill(0)
+  let current = 0
+  let end = last[dated[0]]!
+  for (let k = 1; k < dated.length; k++) {
+    const i = dated[k]
+    if (first[i]! - end > UPLOAD_GAP_MS) {
+      current++
+      end = last[i]!
+    } else if (last[i]! > end) {
+      end = last[i]!
+    }
+    upload[i] = current
+  }
+  if (current === 0) return [files]
+  return Array.from({ length: current + 1 }, (_, u) => files.filter((_, i) => upload[i] === u))
+}
+
+/** The latest article date of a file's segments, or null. */
+function latestOf(segments: EdgeSegment[]): number | null {
+  let max: number | null = null
+  for (const s of segments) if (Number.isFinite(s.postedAt) && (max === null || s.postedAt > max)) max = s.postedAt
+  return max
+}
+
 function postedOf(segments: EdgeSegment[]): number | null {
   let min: number | null = null
   for (const s of segments) {
@@ -293,7 +346,7 @@ function postedOf(segments: EdgeSegment[]): number | null {
   return min
 }
 
-export function chunkOf(key: string, files: AggregatedFile<EdgeSegment>[]): ReleaseChunk {
+export function chunkOf(key: string, files: AggregatedFile<EdgeSegment>[], upload = 0): ReleaseChunk {
   const first = files[0].segments[0].file
   const chunkFiles: ChunkFile[] = files.map((f) => {
     const m = f.segments[0].file
@@ -304,6 +357,7 @@ export function chunkOf(key: string, files: AggregatedFile<EdgeSegment>[]): Rele
   let firstArticle = Number.POSITIVE_INFINITY
   let lastArticle = 0
   let postedAt: number | null = null
+  let postedLast: number | null = null
   for (const f of files) {
     for (const s of f.segments) {
       segments++
@@ -311,18 +365,21 @@ export function chunkOf(key: string, files: AggregatedFile<EdgeSegment>[]): Rele
       const n = s.articleNumber
       if (n < firstArticle) firstArticle = n
       if (n > lastArticle) lastArticle = n
+      if (Number.isFinite(s.postedAt) && (postedLast === null || s.postedAt > postedLast)) postedLast = s.postedAt
     }
     const t = postedOf(f.segments)
     if (t !== null && (postedAt === null || t < postedAt)) postedAt = t
   }
   return {
     key,
+    upload,
     name: first.release?.name ?? first.name,
     source: first.release?.source ?? 'file',
     fileTotal: files.find((f) => f.segments[0].file.release?.fileTotal)?.segments[0].file.release?.fileTotal ?? null,
     poster: first.poster,
     groups: first.groups,
     postedAt,
+    postedLast,
     files: files.length,
     segments,
     bytes,
@@ -339,7 +396,7 @@ export function chunkOf(key: string, files: AggregatedFile<EdgeSegment>[]): Rele
  * volume, of its par2 index, of a lone random-named file, and a 7z set's
  * last volume (its listing is in the tail). At most a few per release.
  */
-export function candidatesOf(key: string, files: AggregatedFile<EdgeSegment>[]): ProbeCandidate[] {
+export function candidatesOf(key: string, files: AggregatedFile<EdgeSegment>[], upload = 0): ProbeCandidate[] {
   const out: ProbeCandidate[] = []
   let tail: { volume: number; file: AggregatedFile<EdgeSegment> } | null = null
   for (const f of files) {
@@ -349,12 +406,12 @@ export function candidatesOf(key: string, files: AggregatedFile<EdgeSegment>[]):
       if (!tail || volume > tail.volume) tail = { volume, file: f }
       continue
     }
-    if (isArchiveStart(name)) out.push({ releaseKey: key, kind: 'archive', messageId: f.segments[0].messageId, file: name })
-    else if (/\.par2$/i.test(name) && !isLaterVolume(name)) out.push({ releaseKey: key, kind: 'par2', messageId: f.segments[0].messageId, file: name })
+    if (isArchiveStart(name)) out.push({ releaseKey: key, upload, kind: 'archive', messageId: f.segments[0].messageId, file: name })
+    else if (/\.par2$/i.test(name) && !isLaterVolume(name)) out.push({ releaseKey: key, upload, kind: 'par2', messageId: f.segments[0].messageId, file: name })
   }
-  if (tail) out.push({ releaseKey: key, kind: '7z-tail', messageId: tail.file.segments.at(-1)!.messageId, file: tail.file.segments[0].file.name })
+  if (tail) out.push({ releaseKey: key, upload, kind: '7z-tail', messageId: tail.file.segments.at(-1)!.messageId, file: tail.file.segments[0].file.name })
   if (out.length === 0 && files.length === 1 && isRandomFile(files[0].segments[0].file.name)) {
-    out.push({ releaseKey: key, kind: 'lone', messageId: files[0].segments[0].messageId, file: files[0].segments[0].file.name })
+    out.push({ releaseKey: key, upload, kind: 'lone', messageId: files[0].segments[0].messageId, file: files[0].segments[0].file.name })
   }
   return out.slice(0, MAX_CANDIDATES)
 }