You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
This repository was archived by the owner on Aug 12, 2026. It is now read-only.
Approve CDT++ v1.0.0 as ready for the final publish-and-archive operation.
This is the release-readiness umbrella. It starts after #96 and must close before #97. Issue #97 alone owns creating the final tag and release, verifying the DOI and artifacts, closing the last tracker, and archiving the repository.
Current release contract
Preserve the supported 2+1-dimensional spherical CDT implementation as an independent C++23 reference for causal-triangulations.
Preserve the manifested initial-state workflow: initialize --output writes an initial-triangulation.off/.off.meta pair, and cdt --input PATH validates that exact pair before starting a new Metropolis-Hastings run with a new transition seed.
Treat that manifested pair as the archival input boundary for a future causal-triangulations native importer. The Rust side should validate the pair and reconstruct its own invariant-bearing representation; CDT++ does not add a second converter or claim that its CGAL payload is generic mesh OFF.
Require one clean-checkout build/test contract, enforced release gates, a validated generated documentation site, synchronized metadata, and a complete archival handoff.
Add no scientific capability beyond the completed supported 2+1-dimensional move set.
Exact combinatorial fields, named numerical tolerances, explicit deterministic transitions, one bounded end-to-end reference run, and complete provenance are versioned and reproducible.
Randomized CGAL results and heuristic estimators are documented without presenting random topology counts as exact scientific oracles.
The Restore and validate multithreaded C++ support #88 concurrency boundary, deterministic single-thread behavior, and existing scaling evidence are documented without overstating trajectory reproducibility.
The local Python harness reproduces stored comparison summaries offline and contains no ML or hosted experiment requirement.
The Qt workflow regenerates the committed hero artifact from repository-owned inputs while the default build remains headless.
initialize --output and cdt --input complete the documented four-step initial-state workflow: generate the pair, validate and load it, start a new seeded CDT transition run, and prove the cross-executable path in CTest.
cdt --input accepts only a manifested initial-triangulation, rejects construction-option conflicts plus checkpoint/final artifacts, validates payload and causal metadata, reconstructs the manifold, and records the source seed, initialization stream, placement fingerprint, and topology fingerprint in evolved provenance.
The docs distinguish initial-state loading from checkpoint resume, show how to generate many independently seeded starting states, and place the future causal-triangulations importer direction in reference/README.md rather than presenting it as a current top-level feature.
Release notes, README, API documentation, CITATION.cff, REFERENCES.md, version metadata, limitations, security/support boundary, successor link, and archival instructions agree.
No 3+1-dimensional work, broad architecture rewrite, comprehensive new stochastic study, unrelated performance program, or other out-of-scope feature is included.
Lifecycle
Close #90 only as the readiness approval for #97. Do not tag, publish, or archive from this issue.
Summary
Approve CDT++ v1.0.0 as ready for the final publish-and-archive operation.
This is the release-readiness umbrella. It starts after #96 and must close before #97. Issue #97 alone owns creating the final tag and release, verifying the DOI and artifacts, closing the last tracker, and archiving the repository.
Current release contract
causal-triangulations.initialize --outputwrites aninitial-triangulation.off/.off.metapair, andcdt --input PATHvalidates that exact pair before starting a new Metropolis-Hastings run with a new transition seed.causal-triangulationsnative importer. The Rust side should validate the pair and reconstruct its own invariant-bearing representation; CDT++ does not add a second converter or claim that its CGAL payload is generic mesh OFF.Remaining readiness path
Follow the native GitHub dependency graph:
initializetocdt --inputhandoff, integration test, and user/reproducibility/reference documentation required here.Acceptance criteria
initialize --outputandcdt --inputcomplete the documented four-step initial-state workflow: generate the pair, validate and load it, start a new seeded CDT transition run, and prove the cross-executable path in CTest.cdt --inputaccepts only a manifestedinitial-triangulation, rejects construction-option conflicts plus checkpoint/final artifacts, validates payload and causal metadata, reconstructs the manifold, and records the source seed, initialization stream, placement fingerprint, and topology fingerprint in evolved provenance.causal-triangulationsimporter direction inreference/README.mdrather than presenting it as a current top-level feature.CITATION.cff,REFERENCES.md, version metadata, limitations, security/support boundary, successor link, and archival instructions agree.Lifecycle
Close #90 only as the readiness approval for #97. Do not tag, publish, or archive from this issue.