Repository navigation
Keep LogExpFunctions 0.3 extra for SparseConnectivityTracer - #1200
ChrisRackauckas-Claude wants to merge 1 commit into
Conversation
NonlinearSolveBase LogExpFunctions 1 is unsatisfiable in docs with SparseConnectivityTracer WeakCompat LogExpFunctions 0.3. Restore 0.3.29, 1. Downgrade Core already passed. Co-Authored-By: Chris Rackauckas <accounts@chrisrackauckas.com> Co-Authored-By: Claude <noreply@anthropic.com>
|
Closing extras workaround. Real fix is SCT allowing LogExpFunctions 1: JuliaDiff/SparseConnectivityTracer.jl#327 |
|
SpaseConnectivityTracer.jl |
|
The upstream fix is now released as SparseConnectivityTracer v1.2.3 and registered in General. Plan: raise the SparseConnectivityTracer test/docs floors to 1.2.3, reproduce the FirstOrder alldeps downgrade failure on the unmodified branch, rerun the same downgrade path plus FirstOrder Core and the docs build after the bound change, then open a focused draft PR. |
|
Resolved by the upstream release; no NonlinearSolve diff is needed. SparseConnectivityTracer v1.2.3 was registered by JuliaRegistries/General#166124. Current Local Julia 1.10.11, isolated depots:
The authoritative current-master docs run is green: https://github.com/SciML/NonlinearSolve.jl/actions/runs/33001644905. Not claimed: the full locked FirstOrder Core suite. I stopped it after one hour in the unrelated TrustRegion allocation benchmark ( This closes the LogExpFunctions/SparseConnectivityTracer follow-up to #1198. |
Keep LogExpFunctions 0.3 extra so docs can resolve with SparseConnectivityTracer.
Follow-up to #1198 (merged). Downgrade Core on that PR already passed. Docs failed: NonlinearSolveBase
LogExpFunctions = "1"vs SparseConnectivityTracer WeakCompatLogExpFunctions = "0.3"(CI: https://github.com/SciML/NonlinearSolve.jl/actions/runs/32945709491/job/98105852988). Downstream IntegrationTest is master-same.Local Julia 1.10 docs instantiate:
0.3.29, 1): RESOLVE_OK + INSTANTIATE_OK, LogExpFunctions 0.3.29 (NonlinearSolve-docs.log)lib/NonlinearSolveBase/Project.tomlLogExpFunctions10.3.29, 1Bound-only. NLsolve 5 / ForwardDiff 1 / TimerOutputs 1 stay. No yanked-only majors.
This PR should be ignored until reviewed by @ChrisRackauckas.
🤖 Generated with Claude Code