David Widmann

Maybe one could take some inspiration from how Test resets the task-local RNG: https://github.com/JuliaLang/julia/blob/1438b1578941d0f1cc8f8c958cf3bd2927fd482c/stdlib/Test/src/Test.jl#L1694-L1720

The more I thought about this issue, the more I got the feeling that this is an upstream issue in Malt. The behaviour seems completely surprising, in particular given that...

This is an approach for Distributed (maybe it can be done more elegantly or efficiently?): ```julia julia> using Distributed julia> addprocs(1) 1-element Vector{Int64}: 2 julia> const in = RemoteChannel(() ->...

Should we go with that? The reproducibility issue came up in our triage call today, and we thought we should inform users in the documentation about this problem if it...

The problem is ```julia (jl_PRakDJ) pkg> why Test Roots → Accessors → Test (jl_PRakDJ) pkg> why ChainRulesCore Roots → ChainRulesCore ```

I made a PR to Roots that would fix the ChainRulesCore issue: https://github.com/JuliaMath/Roots.jl/pull/445

> aqua ambiguity test fails until julia 1.9 It is somewhat expected that this highly depends on the Julia versions as some of the ambiguities might disappear due to upstream...

Is it clear why tests are failing? MTK v8.76.0 + SymbolicUtils v1.5.1 works fine for us, so maybe it's some other dependency that's causing tests to fail. v1.7.0 should be...

I started a CI run on the latest v8 release in https://github.com/SciML/ModelingToolkit.jl/pull/2976, to check if the test errors show up there as well.

In #2976 (latest release of MTK v8 + SymbolicUtils 1.5.1) the same errors are thrown, so it seems the update of some other dependency might have broken something. I guess...