Skip to content

EOSManager: support reviewed LALSimulation families (O4c) - #184

Open
oshaughnessy-junior wants to merge 16 commits into
oshaughn:rift_O4cfrom
oshaughnessy-junior:codex/rift_O4c-lalsim-multibranch
Open

oshaughnessy-junior wants to merge 16 commits into
oshaughn:rift_O4cfrom
oshaughnessy-junior:codex/rift_O4c-lalsim-multibranch

Conversation

@oshaughnessy-junior

Copy link
Copy Markdown

Summary

  • adapt EOSManager to both released and reviewed multi-family LALSimulation APIs
  • require explicit branch selection when a mass has multiple stable solutions while preserving the scalar legacy interface
  • keep native NMB tabular/PCA and legacy sequence inputs compatible
  • preserve Kedia-style four-parameter EOS constructors and document the downstream CLI contract

This is the O4c backstop. The corresponding O4d PR targets the junior production line separately.

Validation

  • 9 compatibility tests pass with installed LALSimulation 7.6.1
  • 9 compatibility tests pass with reviewed-development LALSimulation 6.2.1 build
  • Python compile check passes for EOSManager and the compatibility adapter
  • branch is a descendant of the current oshaughn/rift_O4c tip

@oshaughnessy-junior
oshaughnessy-junior deployed to private-review-dispatch-rift-upstream August 29, 2026 11:00 — with GitHub Actions Active
@oshaughnessy-junior
oshaughnessy-junior deployed to private-review-dispatch-rift-upstream September 2, 2026 01:20 — with GitHub Actions Active
@oshaughnessy-junior

Copy link
Copy Markdown
Author

Final O4c backstop real-interface repair and adversarial acceptance update

The junior head is now 4f01edef7040675a3ba689b65a8687ecef7c3091, targeting oshaughn/rift_O4c. It removes the same invented ChoiceDirtyPT / *PerBranch API and uses the exact reviewed LALSimulation SWIG interface at clean commit 974c0ef468b76e8298e67fd8baf71ed259cc5fee.

CIT acceptance at the exact commits:

  • Combined compatibility + real-build suite: 21 passed, 1 xfailed.
  • Output-bound enforcement, built-in multipart SLY, same-build legacy SLY, true two-/nine-column readers, genuine overlapping twin branches, minimal/extended fields, EOSManager Lambda/causality routing, pressure-floor construction, and cross-line helper compatibility all passed.
  • JUnit evidence: local artifact .codex-gate-artifacts/pr213-o4c-4f01edef-combined-gate.xml.

The sole isolated XFAIL is the same upstream reviewed-LALSimulation defect: the authoritative two-transition review fixture exits with SIGSEGV/return code 139 after the correct array conversion. The table/twin/extended/EOSManager acceptance test passes independently and is not masked by that XFAIL.

O4c and O4d now share the downstream-facing contract: legacy default, explicit multipart opt-in, finite family_log_pressure_min = ln(Pc / Pa), canonical multipart=, deprecated reviewed_multibranch= alias, and conflict rejection. NMB sequence and Kedia-style scalar compatibility tests remain green. Independent adversarial review found no remaining RIFT blocker other than the stated upstream two-transition limitation.

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Independent automated review completed at the recorded exact commit. Detailed findings were withheld from public output by the private-context egress policy and require private human declassification.

oshaughn added a commit that referenced this pull request Sep 4, 2026
…af-comment

CIP/EOS tree fits: drop the false leaf-size comment
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants