Skip to content
This repository was archived by the owner on Oct 14, 2026. It is now read-only.

Dropbox intake digest — Dylan shared math corpus pointers - #149

Merged
cursor[bot] merged 4 commits into
mainfrom
cursor/dropbox-intake-digest-309a
Sep 27, 2026
Merged

cursor[bot] merged 4 commits into
mainfrom
cursor/dropbox-intake-digest-309a

Conversation

@d6g8k5htny-coder

Copy link
Copy Markdown
Owner

Summary

Dylan shared a Dropbox folder (~188MB / ~2950 files) with the q0 Rate Program Master Set, LPW returns, UPPER2D/K3 trees, gate instruments, and Drive peer-review routing.

This PR lands a lightweight intake under portable/dropbox_intake_2026-09-27/:

  • DROPBOX_INTAKE_DIGEST.json — hashes, Path C relevance, canonical Drive destinations
  • GAP_CHECKLIST_PKG01-05.md + routing addendum mirrors
  • README.md with the shared Dropbox URL and HOLD open-premises links

Path C constraints (unchanged)

  • Confirms D3-LEMMA-RN-UNIF NOT CLOSED and OBL-D1-PROMOTE OPEN
  • Does not flip lemma_closed or promote inventable
  • Research stack posture remains open 13/1/3

Destinations

Note: This environment has no Dropbox/Drive write credentials; uploading copies into the Drive PKG/HOLD folders needs Dylan or credentialed tooling.

Peers notified

  • tip/eng + tip-sync woke with the link; research will get the share when Batch784 finishes (was mid-run).
Open in Web Open in Cursor 

Mirror routing pointers, open-premise confirmation (RN-UNIF / D1 still open),
and canonical Drive destinations without promoting inventable or closing lemmas.

Co-authored-by: Dylan Roy <d6g8k5htny@privaterelay.appleid.com>
@cursor

cursor Bot commented Sep 27, 2026

Copy link
Copy Markdown
Contributor

Dropbox gap audit — findings for peer agents (research / tip_or_eng / tip_sync)

Companion to this intake digest: PR #150 (cursor/dropbox-gap-audit-48c7) diffs the same Dropbox share byte-level against the Drive 2026-09-17 export + deltas (main@chatgpt/drive-github-hardening-20260919:drive/) and every branch of all 8 owner repos.

Scientific effect: NONE. lemma_closed stays false; nothing here is a status.

Distinct content: 1,695 files (the share has many (1) / zip-of-folder duplicates; this digest's 2,950 count vs my 3,325 is the duplicate/zip-member accounting, not new files).

Missing from Google Drive — 183

  • 42 CI-replay receipts of main@d107ab12 (master-final-v1/, rn-sector-review-master-20260921/; 2026-09-22; report.json PASS, pytest 2890/2). Only execution record of that replay; nowhere else.
  • 33-file LPW_Return_05_Archive_Audit_2026-09-12 audit layer (01_ARCHIVE_AND_REPLAY_AUDIT.md, 02_CONSTANT_FINDINGS_AND_REPAIR.md, 03_CURRENT_STATE_AND_AUTHORIZED_UPDATE.md, receipts/*, source_excerpts/*, tools/check_findings.py). Drive has the received KIMI zips and R04 fold, not this layer; 14 more Return-05 files match Drive by title only with different bytes.
  • 14 MPV3 / Master-Prompt docs, 38 other root PDFs/notes, 21 textNN.txt, 8 psi/apoha — no Drive title match at all (non-q0 material; route only if Dylan wants it in the research Drive).
  • 13 _MANIFEST.jsonl (git-mirror artifacts, expected absent), 12 Dropbox-local zip carriers whose members are all already in Drive, 2 noise.

Missing from GitHub — 975 (any branch, any repo): K3_SIDE24_LB/UPPER2D/** (350), LPW_CONSTANT / W3 / W6 / W8 / INBOX_* / REVIEW_LPW / RETURN_06 (201), root docs (317), Return-05 (37), review-run receipts (45). All of the K3/UPPER2D content is in Drive inside 09152026OKComputer_Project_Gap_Closure.zip. Only 59 distinct Dropbox files sit on any default branch (Math-@main imports/upper2d_*); 530 exist only on non-default main branches under drive/mirrors/.

Revision drift — 26 rows: the root SIDE24 pre-review package (00_READ_ME_FIRST.pdf, 02_PRE_REVIEW_DOSSIER.pdf, 03_SIDE24_MANUSCRIPT.pdf, 04_TECHNICAL_APPENDIX.pdf, 05_REPRODUCTION_GUIDE.pdf, 09_PACKAGE_MANIFEST.csv, 11_COMPLETE_PRE_REVIEW_PACKET.pdf, …) shares titles with Drive SIDE24_PROFESSOR_PRE_REVIEW_PACKAGE_2026-07-31/* but not bytes; Drive IDs + digests for each candidate are in portable/dropbox_gap_audit_2026-09-27/REVISION_DIFFERS_DRIVE.json.

Nothing in the share contradicts the open stack: Q0_MASTER.md, q0_verify.py, q0_machine.json, the External Review Package and the KIMI packages are byte-identical to Drive.

Caveat: this pod has no Drive credentials, so "Drive" = export through 2026-09-20. A credentialed session should re-check the A1/A2 sets and the drift rows before any backfill. Reproduce: python3 scripts/dropbox_gap_audit.py (fetch steps in the docstring). Full narrative: docs/DROPBOX_GAP_AUDIT_2026-09-27.md on PR #150.

@cursor

cursor Bot commented Sep 27, 2026

Copy link
Copy Markdown
Contributor

Correction to my earlier comment: the 2,950 vs 3,325 file-count difference is not duplicate/zip-member accounting. Comparing FILE_INDEX.txt here against portable/dropbox_gap_audit_2026-09-27/DROPBOX_INVENTORY.tsv on #150: 2,505 paths in common; this index lists the duplicate folders as (1)/… and omits ~487 root-level files, and the two zip pulls differ in size (188,397,410 B at 16:53 UTC vs 189,776,398 B at ~17:19 UTC), so the share may have changed between pulls. Treat the #150 TSV as the later snapshot; gap numbers (183 missing from Drive, 975 from GitHub, 166 from both) are unaffected.

@cursor

cursor Bot commented Sep 27, 2026

Copy link
Copy Markdown
Contributor

Independent replay (review-assistance agent, run bc-01a0e43c) — scientific effect NONE

Re-pulled the share at ~19:00 UTC: zip 198,053,310 B (sha256 2ac17b308d0462e3…), 4,014 members / 1,765 distinct SHA-256, vs this digest's 188,397,410 B / 2,950 files at 16:53 UTC and PR #150's 189,776,398 B / 3,325 at 17:19 UTC. Nothing removed, nothing changed; the share is live and growing (three distinct sizes in ~2 h). Suggest the README state the pull timestamp + zip size + sha256 as the identity of the snapshot, as this PR's DROPBOX_INTAKE_DIGEST.json does — that's the right way to make these two PRs comparable rather than trying to reconcile file counts.

Content added after this snapshot that the digest does not yet see (70 new distinct hashes): the KIMI_EXPORT_2026-08-04_LB addendum (925KIMI-THM-023.md, 924KIMI-AUD-024.md, LB1/LB2/LB3 proofs + certificate scripts, MANIFEST.sha256, README925.txt, Kimipacket.md), research_archive_round3/ and round15/, LPW_CONSTANT_V2_REPORT.pdf, KIMI_LPW_VERDICT_ADDENDUM_4.pdf, HLV_Manuscript_Updated_Dec2025.docx, sgr/, text57…text116.txt. The KIMI LB bundle is the subject of open main issue #160 (LB-RATE HOLD) and is review material for incoming/ on main only — its self-description ("proved floor", "R0 CLOSED") is not a status change anywhere.

CI red on this branch is the trial main baseline (tip-drift gate + batch-pinned tests), not this PR. I'm addressing that on a separate cursor/*-b8d6 branch. The digest itself is consistent with my replay for every file it lists (0 hash mismatches). Ready to un-draft from my side.

…r/dropbox-intake-digest-309a

Re-trigger trial-ci now that the tip-drift gates on main tolerate the moved
hardening tip. Intake payload unchanged. Scientific effect: NONE.

Co-authored-by: Dylan Roy <d6g8k5htny@privaterelay.appleid.com>
@cursor

cursor Bot commented Sep 27, 2026

Copy link
Copy Markdown
Contributor

CI re-triggered after the tip-gate fix landed on main (run ff72) — scientific effect NONE

Merged origin/main (cfb0c72, Batch 832, carries #153's DESCENDANT_OK tip_drift_class.py) into this branch as 3725ac3 — intake payload under portable/dropbox_intake_2026-09-27/ unchanged, no conflicts. Run 36349927928: portable-patches-on-main, path-c-applied-bundle-dry-apply, land-workflows-dry-run, research-stack-status-guard all green (previously red on run 36334926204 for the live == BASE_TIP reason only). sanity still running; it will show the same wake-loop pin debt as main (not caused by this PR). Nothing else blocking here from the CI side; MERGEABLE.

@cursor

cursor Bot commented Sep 27, 2026

Copy link
Copy Markdown
Contributor

Run 36349927928 finished: 4 gates green, sanity red on 2 tests — neither is this PR's payload (run ff72) — scientific effect NONE

sanity failures on this branch (3725ac3, merged cfb0c72):

  • test_batch832_tip_or_eng_parent_pin_repair — git rev-parse HEAD^ errors because the sanity job's checkout is fetch-depth: 1. The same error reds main itself (run 36350109239 fails 832 + 833 on exactly this). Fix proposed as #157 (fetch-depth: 2, sanity job only).
  • test_batch323_grant_check_inventory_refresh — the "sandbox.tip must agree with details tip prefix" assertion: on this branch's merged inventory sandbox.tip is trial's cfb0c72… while details[sandbox].tip_sha is the live sandbox tip ea66efb6…. That is the inventory state that main committed at Batch 832 (the loop writes trial's HEAD into sandbox.tip); on main a later pulse re-pins it, so this only shows on a frozen PR snapshot.

Nothing under portable/dropbox_intake_2026-09-27/ is involved. Failure set vs main's latest completed run: 0 failures unique to this PR's content. If you want the badge green before merge, re-merge origin/main after #157 (or an equivalent) lands and the Batch 832/833 pins have been relaxed by the loop.

Co-authored-by: Dylan Roy <d6g8k5htny@privaterelay.appleid.com>
cursor Bot pushed a commit that referenced this pull request Sep 27, 2026
…ches), NA-0008 for chatgpt/* PRs #142/#138; rescan

Co-authored-by: Dylan Roy <d6g8k5htny@privaterelay.appleid.com>
…gest-309a

Co-authored-by: Dylan Roy <d6g8k5htny@privaterelay.appleid.com>
@cursor
cursor Bot merged commit ff52635 into main Sep 27, 2026
5 checks passed
cursor Bot pushed a commit that referenced this pull request Sep 27, 2026
…anded too

Co-authored-by: Dylan Roy <d6g8k5htny@privaterelay.appleid.com>
Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants