Refresh IntervalNonlinearProblem under v7 stack

[ChrisRackauckas-Claude]

Ignore until reviewed by @ChrisRackauckas.

Summary

Single-folder refresh of benchmarks/IntervalNonlinearProblem/ under the v7 stack.

• Stale-marker removal: dropped the # CI retrigger: v7 stack outputs (2026-05-15) comment that batch-1 (Retrigger v7 outputs for 5 cancelled folders (batch 1) #1558) left at the bottom of Project.toml. No new retrigger markers introduced.
• .jmd sources: scanned simpleintervalrootfind.jmd and suite.jmd for the v7 migration patterns (Bool autodiff/verbose, precs=, sol[i], ensemble prob_func arity, SymbolicUtils 4.26->4.30.x). None present -- no edits needed.
• Manifest.toml: refreshed via Pkg.update() on Julia 1.11. Bumps:
  • Adapt 4.5.2 -> 4.6.0
  • ArrayInterface 7.24.0 -> 7.25.0
  • ConcreteStructs 0.2.3 -> 0.2.4
  • DifferentiationInterface 0.7.17 -> 0.7.18
  • JLLWrappers 1.7.1 -> 1.8.0
  • NonlinearSolveBase 2.25.0 -> 2.26.0
  • RuntimeGeneratedFunctions 0.5.18 -> 0.5.19
  • SafeTestsets (+0.1.0)
  • SciMLBase 3.7.1 -> 3.13.0
  • SciMLLogging 1.9.1 -> 2.0.0
  • SciMLOperators 1.17.0 -> 1.21.0
  • SymbolicIndexingInterface 0.3.46 -> 0.3.48
  • Git_LFS_jll 3.7.0+0 -> 3.7.1+0
• Smoke test: SciMLBenchmarks.weave_file(..., (:script,)) succeeded for both simpleintervalrootfind.jmd and suite.jmd.

Test plan

• CI weaves both .jmd files end-to-end on the v7 stack.

🤖 Generated with Claude Code

[ChrisRackauckas]

CI fix pushed in be41aea.

Root cause (from job 80471540571, simpleintervalrootfind.jmd):

ERROR: UndefVarError: `Bisection` not defined in `Main.var"##WeaveSandBox#232"`
Hint: It looks like two or more modules export different bindings with this name, resulting in ambiguity.

Chunk 1 loads Roots and chunk 2 loads BracketingNonlinearSolve + SimpleNonlinearSolve in the same Weave sandbox module. All three export Bisection, so the unqualified Bisection() call became an ambiguous binding and errored. suite.jmd already carried the workaround (using BracketingNonlinearSolve: Bisection) and weaved cleanly in the same run; only simpleintervalrootfind.jmd was missing it.

Change: added using BracketingNonlinearSolve: Bisection to the second code chunk of benchmarks/IntervalNonlinearProblem/simpleintervalrootfind.jmd (one line), matching suite.jmd.

Verified locally by running the full chunk-1 + chunk-2 code sequence (with reduced N) in the benchmark's vendored Project/Manifest environment on Julia 1.11.8 — ITP(), Bisection(), and SimpleNewtonRaphson() paths all solve correctly.

The push retriggered the Benchmarks workflow: https://github.com/SciML/SciMLBenchmarks.jl/actions/runs/27267411961