[d-burg]

Release note

• Audience: users
• Numerical impact: none; a warning, documentation and tests (harness @ 62540fe)
• Migration: none

A run that sets chi_perp_Dc no longer warns about a chi_e fallback it does not use, and setting the key with D_c_type = "none" now warns that it has no effect. The docs now say that the deck key is one value for every surface and that it also moves the critical resonant field.

Regression report

regress --refs 728a7e7d5,62540feda --force (all cases) on feynman (SLURM, 8 threads, Julia 1.11.9, manifest pinned): every case is unchanged, all 25; gal_resistive_pe is N/A on both refs, as on develop. Later commits touch only docs/src/stability.md.

On the Mac (Julia 1.11.6) at 62540feda: runtests_slayer_inputs.jl 94/94, runtests_slayer_runner.jl 171/171.

Notes for reviewers

Follow-up to #616, from an adversarial review of it that finished after it merged.

• False warning. _slayer_chi_inputs counted χ⊥,e as used by the critical Δ even with chi_perp_Dc set. With P_perp_model = "tau_E", a non-none D_c_type, chi_perp_Dc set and no chi_e in the file, the run warned about a fallback that is never used.
• No-effect warning. chi_perp_Dc with D_c_type = "none" was silently ignored.
• Docs. SLAYERControl and stability.md now say the deck value applies at every surface, over any chi_e profile (the ψ-callable form exists only on build_slayer_inputs), that it is positive and finite, and that it reaches the critical resonant field through the critical Δ. The motivation in stability.md now treats the value as a modeling choice and names the chi_e/chi_i closure that reaches the same split, and where that fails.
• Which χ⊥ the sources mean. stability.md said the critical Δ takes the electron channel "in every case". The sources are single-fluid: Fitzpatrick, Phys. Plasmas 32, 062509 (2025), defines the width by electron temperature but with one χ⊥, and Connor et al. (2015) Eqs. 59 and 61 diffuse the total pressure with it. The docs now say the electron channel is GPEC's reading. The code is unchanged.
• Tests. The forwarding through _tearing_params was untested: removing the chi_perp_Dc=chi.chi_perp_Dc line left every test green. The new test fails with that line removed (checked by hand). The ψ-callable form is now covered too, and the three warning paths.

Not in this PR: a per-surface chi_perp_Dc deck table, like omega_E_kHz.

For the CHANGELOG: #616 merged with a ! in its title and Migration: none. Nothing in it breaks an existing deck, so it should read as a plain feature.

---

🤖 Generated with Claude Code

[logan-nc]

@d-burg @matt-pharr We're taking this over to carry it to the merge, since #602 landed and this now conflicts with develop. I'm merging develop in and pushing shortly so CI starts, then doing a full review and re-running the harness. Please hold off pushing to this branch until I post the follow-up.

[logan-nc]

Pushed 6c0a31ad3: develop merged in, with the change ported to where #602 moved it.

• The fallback and no-effect warnings now live in Tearing._slayer_chi_inputs and act on SLAYERModel.
• The chi_perp_Dc notes moved to the SLAYERModel docstring.
• The docs paragraph is kept on top of develop's SLAYERModel wording.
• The tests use the new API: _tearing_params(model, equil, surfaces, loaded), with kinetic data attached for that check.

CI is running. Targeted tests (slayer_inputs, slayer_runner, critical_resonant_field) are running on feynman, and a full review follows.

[logan-nc]

Review follow-up, pushed as be3709294.

• Same gap, one level down: _slayer_profiles (SLAYER/KineticView.jl) still loaded the file's chi_e whenever there was a critical Δ, and errors on any non-positive value. So a run with chi_perp_Dc set and an interpretive chi_e ≤ 0 stopped on a profile it never uses, which is the case the docs give for this key. It now skips chi_e there, the same way the warning does, and a test covers both sides.
• Test fix: the ported _tearing_params test now builds its kinetic data from arrays.

Targeted tests are re-running. The full harness follows the #617 run on alfven, and then the stamp gets updated.

@d-burg One question I couldn't settle: the docs paragraph cites Fitzpatrick, Phys. Plasmas 32, 062509 (2025), Eqs. 230–234, and Connor et al., PPCF 57, 065001 (2015), Eqs. 59 and 61. Neither paper is in docs/resources, so I couldn't check the equation numbers. Can you confirm them?

[logan-nc]

Targeted tests pass at be3709294: runtests_slayer_inputs 94/94, runtests_slayer_runner 156/156 and runtests_critical_resonant_field 71/71. The full harness (alfven job 860) is running; the stamp follows.

[Copilot]

🟡 Changes recommended

An active compatibility issue drops a valid chi_e profile and passes nothing to a Real-typed parameter, causing previously runnable inputs to fail and requiring a fix with regression coverage.

1 open issue

• Avoid passing nothing for unused chi_perp_e · New

What changed in this PR

This follow-up to #616 corrects SLAYER diffusivity warnings and profile handling, adds forwarding tests, and clarifies how chi_perp_Dc affects critical-Δ calculations.

Changes:

• Suppresses unused chi_e fallback warnings and warns when chi_perp_Dc has no effect.
• Tests callable inputs, parameter forwarding, and warning paths.
• Documents the uniform deck override and its physical interpretation.

The PR reports unchanged regression results. Keep the regression-harness report tied to the reviewed commit.

Third-party human approval is required before merging into develop.

File	Description
test/​runtests_slayer_runner.jl	Tests warnings and unused electron-profile handling.
test/​runtests_slayer_inputs.jl	Tests callable overrides and parameter forwarding.
src/​Tearing/​SLAYERTearing.jl	Corrects fallback and no-effect warnings.
src/​InnerLayer/​SLAYER/​SLAYER.jl	Clarifies override scope and critical-field effects.
src/​InnerLayer/​SLAYER/​KineticView.jl	Skips electron profiles when unused.
docs/​src/​stability.md	Explains diffusivity modeling choices and override behavior.

_{Review effort: Balanced}

[logan-nc]

@matt-pharr @jhalpern30 This needs a review from one of you, since we pushed the port and the chi_e fix ourselves. The diff against develop is small: +61/−11 across docs, two lines of warning logic in SLAYERTearing.jl, one condition in KineticView.jl, and tests. Targeted tests pass. The harness (alfven job 860) is running and the stamp goes in when it finishes.

[logan-nc]

Harness record, run after merge: regress --refs e51f4b0ee,be3709294 (all 25 cases) on alfven, 16 threads, Julia 1.11.7, manifest pinned. Every tracked quantity is unchanged. gal_resistive_pe reports N/A on both refs, develop included, so it is a stale case definition and not this PR.