ForceFreeStates - REFACTOR - Single-source the auto chunk-count target

[d-burg]

Release note

• Audience: developers
• Numerical impact: none (harness @ a47f409) — the resolved chunk targets are unchanged; diiid_n1_riccati is bit-identical head vs base
• Migration: none

The nchunks = 0 auto chunk-count target and the crossing floor were each written out in three places: the authoritative formula inside balance_integration_chunks plus hand-copied mirrors in runtests_parallel_integration.jl and runtests_decomposition_invariance.jl. They are now single functions — ForceFreeStates.auto_chunk_target(msing) and ForceFreeStates.min_crossing_chunks(msing) — used by the solver and both test files.

Regression report

regress --cases diiid_n1_riccati --refs 30eb52073,a47f409f0 --force (base = #362's head feature/thread-invariance, so the comparison isolates exactly this refactor; raw shas, both sides fresh):

Regression Report: diiid_n1_riccati
Ref 1: 30eb52073  @ 30eb52073 (2026-08-28)
Ref 2: a47f409f0  @ a47f409f0 (2026-08-28)
Quantity                       30eb52073      a47f409f0      Diff     Status
delta prime (BVP diagonal)     [5 elem]       [5 elem]       0.0e+00  OK
delta prime (raw side-major)   [10 elem]      [10 elem]      0.0e+00  OK
edge coil response delta_coil  [10 elem]      [10 elem]      0.0e+00  OK
total energy Re(et[1])         8.038607e-01   8.038607e-01   0.0e+00  OK
plasma energy Re(ep[1])        -1.344628e+00  -1.344628e+00  0.0e+00  OK
vacuum energy Re(ev[1])        2.148488e+00   2.148488e+00   0.0e+00  OK
total energy (all)             [35 elem]      [35 elem]      0.0e+00  OK
# singular surfaces            5              5              0.0e+00  OK
singular psi locations         [5 elem]       [5 elem]       0.0e+00  OK
singular q values              [5 elem]       [5 elem]       0.0e+00  OK
ODE steps (saved)              51             51             0.0e+00  OK
ODE steps (total)              1667           1667           0.0e+00  OK
mpert                          35             35             0.0e+00  OK
npert                          1              1              0.0e+00  OK
q0                             1.204212e+00   1.204212e+00   0.0e+00  OK
q95                            4.781723e+00   4.781723e+00   0.0e+00  OK
beta_n                         1.372511e+00   1.372511e+00   0.0e+00  OK
Runtime (s)                    107.8s         107.8s                  --
Summary: 17 unchanged

Validation

• runtests_parallel_integration.jl at a47f409f0, julia -t 4, fresh worktree: 95/95
• runtests_decomposition_invariance.jl same conditions: 11/11
• The test workflow does not trigger while the base is feature/thread-invariance (it filters on base develop/main); it will run once the base is retargeted after Test - TEST - Decomposition invariance of the unified Riccati Delta-prime path #362 merges — the local runs above cover the gap

Notes for reviewers

• One deliberate mirror survives, converted from plumbing into a pin: runtests_parallel_integration.jl still asserts auto_chunk_target(msing) == max(2msing+3, 8(msing+1)+msing). That keeps the tripwire — a change to the src formula fails a test loudly and forces a conscious update — while the operational values everywhere come from the single source. Without the pin, a formula change would propagate silently into every consumer and no test would notice.
• The min_bvp_intervals conditioning rationale (8 sub-chunks per segment, STRIDE's 33 intervals, cond ~ 10²⁴) moved verbatim from the inline comment into auto_chunk_target's docstring.
• Stacked on Test - TEST - Decomposition invariance of the unified Riccati Delta-prime path #362 (feature/thread-invariance), since runtests_decomposition_invariance.jl only exists there; retarget to develop after Test - TEST - Decomposition invariance of the unified Riccati Delta-prime path #362 merges.

---

Draft until #362 lands and Daniel names reviewer/assignee.

🤖 Generated with Claude Code

[github-actions]

This pull request is missing an assignee and a reviewer.

If you are not ready to name them, mark this pull request as a draft.
docs/development/contributors.md suggests lead developers to ask.
Merging is not blocked here, but no pull request may be merged without human review.