Reland "Test raising one moment microphysics"#609
Conversation
This reverts commit b7cf547.
Codecov Report✅ All modified and coverable lines are covered by tests. 📢 Thoughts on this report? Let us know! |
|
@giordano / @kaiyuan-cheng I think i see the issue. #587 added The temporary fix is to set use the non-smooth branch by setting Breeze.jl/ext/BreezeCloudMicrophysicsExt/one_moment_microphysics.jl Lines 844 to 846 in b7cf547 cause this avoids the code path. I wonder if we can add the smoothness flag to the |
|
what does "non-smooth branch" mean? |
|
smoothing the kink @inline conv_q_icl_to_q_sno_no_supersat(
(; τ, q_threshold, k)::CMP.Acnv1M{FT},
q_icl::FT,
smooth_transition::Bool = false,
) where {FT} =
smooth_transition ?
CO.logistic_function_integral(q_icl, q_threshold, k) / τ :
max(0, q_icl - q_threshold) / τ |
v interesting |
|
Still failing with Julia v1.12 only: https://github.com/NumericalEarth/Breeze.jl/actions/runs/24053001251/job/70152934537?pr=609#step:10:632 Is there a recursive function somewhere? |
|
The stackoverflow error seems to be gone for the time being, but now we have a new exciting error: |
|
Exciting! I've been hitting this elsewhere. Is this |
Looks like it |
|
Breeze.jl/src/AtmosphereModels/set_atmosphere_model.jl Lines 169 to 173 in 1ccbb30 we could try using |
why do the grids differ? |
|
They don't, the issue is computing equalities between |
Do we need a custom equality operator for Reactant then? |
Reverts #608 and relands #595. This will need some work to fix the raising failure caused by #587 + #595, but opening the PR not to forget about it