test: several test fixes

[AayushSabharwal]

Checklist

• Appropriate tests were added
• Any code changes were done in a way that does not break public API
• All documentation related to code changes were updated
• The new code follows the
  contributor guidelines, in particular the SciML Style Guide and
  COLPRAC.
• Any new documentation only uses public API

Additional context

Add any other context about the problem here.

[github-actions]

Benchmark Results (Julia vlts)

Time benchmarks

	master	0d3a55f...	master / 0d3a55f...
ODEProblem	6.62 ± 0.46 ms	5.88 ± 0.27 ms	1.12 ± 0.093
init	0.095 ± 0.027 ms	0.0694 ± 0.026 ms	1.37 ± 0.65
large_parameter_init/ODEProblem	18.3 ± 1.4 ms	15.7 ± 1.5 ms	1.17 ± 0.14
large_parameter_init/init	0.1 ± 0.04 ms	0.091 ± 0.035 ms	1.1 ± 0.61
mtkcompile	11.3 ± 0.64 ms	9.3 ± 0.41 ms	1.22 ± 0.088
sparse_analytical_jacobian/ODEProblem	23.8 ± 1.1 ms	20.4 ± 1.3 ms	1.16 ± 0.092
sparse_analytical_jacobian/f_iip	0.07 ± 0 μs	0.07 ± 0.001 μs	1 ± 0.014
sparse_analytical_jacobian/f_oop	0.384 ± 0.012 ms	0.387 ± 0.011 ms	0.992 ± 0.043
time_to_load	6.27 ± 0.04 s	5.97 ± 0.095 s	1.05 ± 0.018

Memory benchmarks

	master	0d3a55f...	master / 0d3a55f...
ODEProblem	0.0382 M allocs: 2.21 MB	0.0382 M allocs: 2.21 MB	1
init	0.428 k allocs: 0.0725 MB	0.428 k allocs: 0.0725 MB	1
large_parameter_init/ODEProblem	0.159 M allocs: 5.99 MB	0.159 M allocs: 5.99 MB	1
large_parameter_init/init	0.616 k allocs: 0.175 MB	0.616 k allocs: 0.175 MB	1
mtkcompile	0.059 M allocs: 3.4 MB	0.059 M allocs: 3.38 MB	1
sparse_analytical_jacobian/ODEProblem	0.117 M allocs: 5.69 MB	0.117 M allocs: 5.71 MB	0.996
sparse_analytical_jacobian/f_iip	0 allocs: 0 B	0 allocs: 0 B
sparse_analytical_jacobian/f_oop	0.634 k allocs: 19.6 kB	0.634 k allocs: 19.6 kB	1
time_to_load	0.153 k allocs: 14.5 kB	0.153 k allocs: 14.5 kB	1

[github-actions]

Benchmark Results (Julia v1)

Time benchmarks

	master	0d3a55f...	master / 0d3a55f...
ODEProblem	6.7 ± 0.42 ms	6.42 ± 0.48 ms	1.04 ± 0.1
init	0.074 ± 0.019 ms	0.0811 ± 0.02 ms	0.913 ± 0.32
large_parameter_init/ODEProblem	19.3 ± 2 ms	18.8 ± 2.2 ms	1.03 ± 0.16
large_parameter_init/init	0.0876 ± 0.021 ms	0.0897 ± 0.027 ms	0.977 ± 0.38
mtkcompile	9.67 ± 0.43 ms	9.49 ± 0.71 ms	1.02 ± 0.089
sparse_analytical_jacobian/ODEProblem	22.7 ± 1.4 ms	23.3 ± 2 ms	0.976 ± 0.1
sparse_analytical_jacobian/f_iip	0.09 ± 0.009 μs	0.09 ± 0.01 μs	1 ± 0.15
sparse_analytical_jacobian/f_oop	0.176 ± 0.013 ms	0.155 ± 0.013 ms	1.14 ± 0.13
time_to_load	6.14 ± 0.07 s	6.14 ± 0.047 s	1 ± 0.014

Memory benchmarks

	master	0d3a55f...	master / 0d3a55f...
ODEProblem	0.0364 M allocs: 1.65 MB	0.0364 M allocs: 1.65 MB	1
init	0.426 k allocs: 0.0545 MB	0.426 k allocs: 0.0545 MB	1
large_parameter_init/ODEProblem	0.152 M allocs: 5.17 MB	0.152 M allocs: 5.15 MB	1.01
large_parameter_init/init	0.809 k allocs: 0.154 MB	0.809 k allocs: 0.154 MB	1
mtkcompile	0.0556 M allocs: 2.59 MB	0.0556 M allocs: 2.61 MB	0.994
sparse_analytical_jacobian/ODEProblem	0.117 M allocs: 4.76 MB	0.117 M allocs: 4.76 MB	0.999
sparse_analytical_jacobian/f_iip	0 allocs: 0 B	0 allocs: 0 B
sparse_analytical_jacobian/f_oop	0.848 k allocs: 27 kB	0.848 k allocs: 27 kB	1
time_to_load	0.145 k allocs: 11 kB	0.145 k allocs: 11 kB	1

[ChrisRackauckas-Claude]

The reason no Tests / Sublibrary CI checks are showing up on this PR is the env: block added in 0d3a55f, not path filtering.

A job that calls a reusable workflow with uses: may not carry an env: map, and the failure mode is not "the env is ignored" — the whole workflow file becomes invalid and GitHub creates zero jobs:

$ actionlint .github/workflows/Tests.yml .github/workflows/SublibraryCI.yml
.github/workflows/Tests.yml:30:5: when a reusable workflow is called with "uses", "env" is not available.
  only following keys are allowed: "name", "uses", "with", "secrets", "needs", "if", and "permissions" in job "tests" [syntax-check]
.github/workflows/SublibraryCI.yml:27:5: ... in job "sublibrary-ci" [syntax-check]

Both runs exist but are listed by file path with no jobs — https://github.com/SciML/ModelingToolkit.jl/actions/runs/30901431064 reports "This run likely failed because of a workflow file issue."

For the record on the path-filtering theory: sublibrary test-only changes do select the sublibrary. compute_affected_sublibraries.jl puts any lib/<pkg>/** change into the direct set (only reverse-dep propagation is gated on src//Project.toml), and the root grouped-tests.yml matrix is not diff-filtered at all. Piping lib/ModelingToolkitBase/test/symbolic_events.jl into --projects-matrix yields the full 25-cell MTKBase expansion, and #4843 (which touched only lib/ModelingToolkitBase/test/bvproblem.jl) ran both Tests and Sublibrary CI.

Two PRs:

• ci: revert the env: blocks that invalidated Tests.yml and SublibraryCI.yml #4873 — reverts the two env: blocks against this branch, so CI runs again. actionlint .github/workflows/*.yml exits 0.
• feat: add auto-precompile input to the reusable test workflows .github#125 — adds an auto-precompile input to tests.yml (job-level, so it covers julia-buildpkg as well as julia-runtest) and forwards it from grouped-tests.yml and all four sublibrary-project-tests.yml shards. Once that merges and v1 is retagged, this becomes auto-precompile: false in with:. v1 only moves on a vX.Y.Z tag push, so that step needs a release cut.

Both are drafts pending @ChrisRackauckas review.