[matt-pharr]

Release note

• Audience: users
• Numerical impact: none (harness @ 7310db1)
• Migration: none

When the Euler-Lagrange ODE solver gives up partway through an interval (for example when its step size underflows on a stiff kinetic run), GPEC now stops with an error naming the ψ it reached. Before, it carried on with a truncated solution and later failed with an unrelated MethodError in the edge dW scan.

Regression report

Regression Report: diiid_n1
Ref 1: feature/solve-fixed-kinetic  @ 0d7b8fe53 (2026-10-23)
Ref 2: bugfix/report-el-abort  @ 7310db1d1 (2026-10-23)
       env: julia 1.12.6, x86_64-linux-gnu, manifest aa94b997 (pinned), 8 threads/4 BLAS, jit build
Quantity                                      feature/solve-fixed-kinetic  bugfix/report-el-abort  Diff       Status
total energy Re(et[1])                        8.006123e-01                 8.006123e-01            0.0e+00    OK
plasma energy Re(ep[1])                       -1.372261e+00                -1.372261e+00           0.0e+00    OK
vacuum energy Re(ev[1])                       2.172874e+00                 2.172874e+00            0.0e+00    OK
PE plasma energy                              3.320677e+00                 3.320677e+00            0.0e+00    OK
NTV torque FGAR [N·m]                         5.483495e-01                 5.483495e-01            0.0e+00    OK
...
Summary: 53 unchanged

Regression Report: diiid_n1_riccati
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
total energy Re(et[1])         8.031971e-01                 8.031971e-01            0.0e+00  OK
...
Summary: 17 unchanged

Regression Report: solovev_n1
total energy Re(et[1])        1.090560e-02                 1.090560e-02            0.0e+00    OK
...
Summary: 22 unchanged

Regression Report: solovev_kinetic_calculated
total energy Re(et[1])    1.827879e+00                 1.827879e+00            0.0e+00  OK
total energy Im(et[1])    -1.326825e+00                -1.326825e+00           0.0e+00  OK
...
Summary: 17 unchanged

Notes for reviewers

• Stacked on ForceFreeStates - FEATURE - Accept the fixed kinetic source in solve(EulerLagrangeProblem) #550 (feature/solve-fixed-kinetic). The harness compares against that branch.
• solve_el is the single entry point for every Euler-Lagrange ODE solve: the forward integrate_el_region! and the five Riccati propagator solves. It now checks SciMLBase.successful_retcode(sol). No ForceFreeStates callback calls terminate!, so an unsuccessful return code always means a real failure.
• findmax_dW_edge! is unchanged: with the root cause caught, its findfirst can no longer return nothing.
• Tests:
  • a du/dψ = u² blow-up through solve_el (direct, deterministic), plus a decaying ODE as the control;
  • the original crash, the coarse Solovev fixture at kinetic_factor = 0.1 through solve, now raises the new message (it aborts at ψ ≈ 0.404).
• The Riccati Δ′ case (diiid_n1_riccati) is in the harness to show that none of the currently passing propagator solves were silently ending early.

🤖 Generated with Claude Code

[logan-nc]

Read in full. Ready to approve from my read; nothing to add. successful_retcode is the right gate given no FFS callback calls terminate!, the message names a real key (eulerlagrange_tolerance), and the blow-up/decay pair tests the gate directly rather than through a full solve.

[logan-nc]

Approving at 7310db1 on the comment review above. Stacked on #550.