Repository navigation
Remove 1350 G20 ADDs through exact dependency-trace resynthesis - #5828
Merged
Mister-Meeseeks merged 1 commit intoOct 27, 2026
Merged
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
PR5807 optimizes contiguous scalar windows. This construction crosses only proved independent operations, including disjoint native events, and applies exact four-role synthesis to the projected dependency trace. The full emitted word removes75 ADDs per call, or1350 across18 calls. Its positive majorant falls2653582→2625858. The two-primitive-per-ADD source-context count falls952838305581→952838302881.
Simple commuting cancellation saved only two ADDs; the larger dependency projection makes68 strict replacements over five rounds. The final word is a fixed point for the stated closure and one-conflict commutation scans, not a global circuit minimum.
Every actual native/reset operation sees exactly the same complete RAW input. Complete685-axis forward and reverse endpoints, source frames, paid profiles, owned peak, exact four-role minima and exhaustive local truth tables pass. Controls reject a dependent native crossing and a changed query owner. Registered normal/-O, frozen-source hashes, committed bytes, package export and sealed Git-free replay pass.
The recursive paid profile and target inventory are unchanged. Source native proofs and the full caller remain dependencies. The global event-index/compiler artifact has not been rebuilt. No physical ROW/tape admission, new kappa or selected-frontier change is claimed. Hosted CI and independent acceptance remain separate.
Retain PR5807/5793, PR5708/5679, Dugongue, ai-nanachi and all G20/native/ordinary credits in inputs/NOTICE. New source-wide construction by Ali Romman with substantial OpenAI Codex assistance; no subagents.
Reproduce:
python3.12 -B scripts/research_package.py --package research/g20-trace-resynthesis.Canonical claim and handoff: #5804
Base:
de5ec2e7a6074068452f45b1a191b5c72832df4a. Verified head:fc24bd1fce81b3e40ba3fa544160d36da8d1cec2.Prepared for Ali Romman with substantial OpenAI Codex assistance. Retained tool and mathematical predecessor credits apply. No selected-frontier change.
Hosted status at this exact head: the planning job fails in the unchanged
test_research_current_assignment.pysuite (eight failures and two errors); scientific replay is skipped. Exact logs were checked. Existing PR5761 contains the shared test repair. The complete local and sealed-export checks above are the scientific evidence supplied here; hosted scientific success is not claimed.Follow-through: PR5834 installs the released local word in all 18 physical calls and rebuilds every global query, lifetime and cost record. That integration is a separate verification scope and does not add a second local-savings claim here.