Fix the three QA failures on master (explicit import, piracy whitelist, check_reexports)#1112
Merged
Merged
Conversation
- ExplicitImports: `HomotopyPolyAlgorithm` is used in `src/poly_algs.jl` but was only reaching the module through a reexport; import it explicitly from NonlinearSolveBase (introduced by SciML#1105). - Aqua piracy: `initialization_alg(::SciMLBase.HomotopyProblem, autodiff)` in `src/default.jl` is an intentional dispatch on the continuation problem type, alongside the `AbstractNonlinearProblem` method that is already whitelisted; add `HomotopyProblem` to `treat_as_own` (also SciML#1105). - check_reexports: SciMLTesting 2.4.0 enabled this check by default. NonlinearSolve is a facade that deliberately reexports the whole solver stack, so pass a `reexports_allow` list built from the sublibraries' public API names and module names, unioned with the existing external-reexport set already computed for the docs check. Verified locally: `GROUP=QA Pkg.test()` now passes 19/19 (was 16 passed, 2 failed, 1 errored). Co-Authored-By: Chris Rackauckas <accounts@chrisrackauckas.com>
ChrisRackauckas
marked this pull request as ready for review
July 25, 2026 13:48
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Important
This PR should be ignored until reviewed by @ChrisRackauckas.
Fixes the
tests / QAjob on master, which has been failing with16 passed, 2 failed, 1 errored. Three failures, two unrelated causes — none of them a defect in shipped behavior:no_implicit_imports—HomotopyPolyAlgorithmatsrc/poly_algs.jl:137using NonlinearSolveBase: ...listPiracy—initialization_alg(::SciMLBase.HomotopyProblem, autodiff)atsrc/default.jl:61AbstractNonlinearProblemmethod was already whitelisted, this one wasn'tSciMLBase.HomotopyProblemto the existingpiracies.treat_as_ownNo unapproved public reexportscheck_reexportsdefaulting totrue. First failing run wasb92a835e(2026-07-21), the first to resolve 2.4.0reexports_allow— NonlinearSolve is a facade that deliberately reexports the whole stackFor the third one I built the allow-list the same way the file already builds its docs-check ignore list: the public API names of each sublibrary (
NonlinearSolveBase,NonlinearSolveFirstOrder,NonlinearSolveSpectralMethods,NonlinearSolveQuasiNewton,SimpleNonlinearSolve,BracketingNonlinearSolve) plus the module names, unioned with the existing external-reexport set (ADTypes/SciMLBase/LineSearch/LinearSolve). That keeps the check meaningful — an accidental public reexport from some other package would still fail — while accepting the intentional facade surface.Verification
GROUP=QA julia --project -e 'using Pkg; Pkg.test()'locally on Julia 1.12.4: 19/19 pass (from 16 passed / 2 failed / 1 errored on the same checkout without these changes).Not fixed here (master is still red for two other reasons)
test/Core/homotopy_alloc_tests__item1.jl. Root cause is Julia 1.12 removing thereshape(a::Array, dims) === afast path, surfacing through ForwardDiff'sreshape_jacobian; the real fix is upstream in ForwardDiff and the interim is a tolerance decision, so both are flagged for you there rather than done here.Broyden(true_jacobian, bad_broyden)but problem 8 for the same algorithm still flips; commented on that issue, since the mechanism that fits is a test skip and needs your sign-off.🤖 Generated with Claude Code
https://claude.ai/code/session_01B9GHoo4pFa3DU9yhyTkevL