test master

[isaacsas]

No description provided.

[isaacsas]

@ChrisRackauckas we're now getting test failures in ODE solvers at

Catalyst.jl/test/dsl/dsl_options.jl

Line 791 in af5b8c2

@test sol[S][end] ≈ sol.ps[p]/sol.ps[d]

Nothing should have changed in Catalyst in the last day to cause this suddenly (tests passed yesterday). Is this related to initialization changes or structural_simplify changes?

cc: @TorkelE this looks like a DAE example you made?

[isaacsas]

Ahh, I see the issue.

@TorkelE this isn't a consistent initialization. At t = 0 the equations should hold, and as such X and Y are fully deteremined by S(0), however, you are specifying values for them. Can you update any such examples in the tests? There seem to be multiple failures now due to such issues.

[ChrisRackauckas]

Yes, that would be from SciML/ModelingToolkit.jl#3032. It fixes @TorkelE 's issue of SciML/ModelingToolkit.jl#2619 which there were cases we let inconsistent initial conditions through before. Now that we plugged that, yes that would error now 😅

[TorkelE]

Glad to see that it got fixed. Yes, superfluous initial conditions should be removed now, I should be able to get to it in a day or so

[isaacsas]

Thanks!