Fix master CI: import OrdinaryDiffEq sub-packages explicitly#62
Merged
ChrisRackauckas merged 4 commits intoMay 11, 2026
Merged
Conversation
After OrdinaryDiffEq v7's split into sub-packages, `TRBDF2` is no longer in scope from a bare `using OrdinaryDiffEq`. Import OrdinaryDiffEqSDIRK explicitly so the tests resolve it, and add the sub-package to the test/Project.toml deps. Co-Authored-By: Chris Rackauckas <accounts@chrisrackauckas.com>
… eachindex(sol)
Under SciMLBase v3 / RecursiveArrayTools v4, `eachindex(::ODESolution)` returns
a `CartesianIndices{2}`, so generators like
`[exact_solution.(mesh_points, sol.t[i]) for i in eachindex(sol)]` index
`sol.t::Vector` with a `CartesianIndex{2}` and throw a `BoundsError`. The
intended iteration is over time steps, so iterate over `eachindex(sol.t)`.
Also replace `using PrecompileTools` in `src/precompilation.jl` with an
explicit `using PrecompileTools: @setup_workload, @compile_workload` so the
`check_no_implicit_imports` test passes.
Co-Authored-By: Chris Rackauckas <accounts@chrisrackauckas.com>
Contributor
Author
Iteration 2 updatePushed
CI status after this commit (Julia 1 + lts)Both targeted issues resolved. Test summary went from
These were latent on The clean fix is to regenerate the reference PNGs on a GitHub-Actions runner so byte-identical reproducibility is guaranteed, but doing that locally (sandbox renderer differs from CI's) would just shift the failure. Out of scope for this PR's targeted fixes; recommend a follow-up PR that runs |
…rences The reaction-diffusion test's analytical solution involves exp(-16π² t) and exp(-25π² t) terms, which decay to O(1e-3) by t = 0.05. Default solver tolerances (abstol = 1e-6, reltol = 1e-3) leave residuals comparable to the signal itself, so the rtol = 0.1 comparison against the closed-form exact_u no longer holds under OrdinaryDiffEq v7's TRBDF2 step-size heuristics. Tighten the solver tolerances (abstol = 1e-10, reltol = 1e-8) so the comparison is meaningful (norm error drops from ~0.26 to ~1.6e-5 relative). Apply the same tighter tolerances in the matching docs example so the displayed figure can be reproduced from the documented code. Regenerate the three surface PNG references (porous_surface, robin_diffusion_surface, dirichlet_source_surface) against CairoMakie v0.15 output. Renderer/font changes since the references were committed dropped PSNR below 25 (22.14 / 22.69 / 21.76) without any change to the underlying solutions; the regenerated PNGs match the CairoMakie v0.15 renderer used on current CI. Co-Authored-By: Chris Rackauckas <accounts@chrisrackauckas.com>
Co-authored-by: Christopher Rackauckas <accounts@chrisrackauckas.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
After OrdinaryDiffEq v7's split into sub-packages,
TRBDF2is no longer in scope from a bareusing OrdinaryDiffEq. ImportOrdinaryDiffEqSDIRKexplicitly so the tests indirichlet_source.jl,fisher.jl,heat.jl,porous_medium.jl, androbin_diffusion.jlresolve it.Master CI on this repo has been red since the OrdinaryDiffEq v7 ecosystem-bump merge. This PR resolves the 5x
UndefVarError: TRBDF2 not definedfailures.Note: after the import is fixed, two additional pre-existing master failures remain (NOT introduced by this PR):
explicit_imports.jltest fails becausesrc/precompilation.jldoesusing PrecompileToolsrather thanusing PrecompileTools: @setup_workload, @compile_workload.BoundsErrorat[exact_solution.(mesh_points, sol.t[i]) for i in eachindex(sol)]becauseeachindex(::ODESolution)returnsCartesianIndices{2}under SciMLBase v3 / RecursiveArrayTools v4 instead ofBase.OneToas before. Test code needs to switch toeachindex(sol.t)(or similar).Both are out of scope for this PR per task constraints. Filing as a separate concern.
Ignore until reviewed by @ChrisRackauckas.