Update README.md with installation instructions

[isaacsas]

@TorkelE I think we should add some installation instructions in the README that tell people not to install ModelingToolkit versions post 9.59.

[TorkelE]

I am not a fan, but given people are getting errors due to somehow mixing Catalyst with the latest MTK version we probably have to stick this in there until things clear out.

[ChrisRackauckas]

What is left for the Catalyst updates?

[TorkelE]

On the top of my head I think we got:

• Regression: Can no longer solve SteadyStateProblems after remake ModelingToolkit.jl#3483 (although can probably just mark as broken as it is not too major.
• This issue Cannot recreate circadian model JumpProcess example #1199 just popped up. There is also a new events related error in the Catalyst tests that was not there a few days ago which I will need to figure out.
• This Catalyst test also started failing suddenly just a few days ago: https://github.com/SciML/Catalyst.jl/blob/master/test/simulation_and_solving/simulate_jumps.jl#L179, which is likely also due to some upstream change.
• We need to figure out what to do about conservation law removal for NonlinearSystems.

I will try to look at this in the next few days. But realistically each of the unreported issues will take me a while to isolate and create reports on, and I was already planning to spend the Sunday catching up on my normal work as I skipped some of it will trying to push through to enable the Catalyst release. Right now it is a bit of an infinite work though, as every time I have figured out some issue another one has popped up.

[ChrisRackauckas]

SciML/ModelingToolkit.jl#3483 (although can probably just mark as broken as it is not too major.

That should be solved now. A new remake function was added for SteadyStateProblem that covers that case now.

This issue #1199 just popped up. There is also a new events related error in the Catalyst tests that was not there a few days ago which I will need to figure out.
This Catalyst test also started failing suddenly just a few days ago: https://github.com/SciML/Catalyst.jl/blob/master/test/simulation_and_solving/simulate_jumps.jl#L179, which is likely also due to some upstream change.

Try the latest patch when it comes out in 20 minutes. One change is we removed an old hack and that makes CI more correct SciML/ModelingToolkit.jl#3486 . If one of those don't go away from that update then let me know and I can dive in there.

We need to figure out what to do about conservation law removal for NonlinearSystems.

Have you setup the codegen to the form of SciML/ModelingToolkit.jl#3458 (comment) yet? That should be the solution. Now there are some cases where you will need SciML/ModelingToolkit.jl#3452, but we can keep those separate since you already don't handle those cases robustly anyways, so nothing regresses without it, it would just fix new cases (and in fact, that should be the final fix to make conservation laws fully correct, well the previous isn't incorrect but BrownBasic isn't entirely what the user would expect so it would look corrected because it's fixing an ambiguity... long story can elaborate more if you want)

[TorkelE]

Saw that SciML/ModelingToolkit.jl#3483 was closed basically as soon I posted. Haven't looked at the solution to SciML/ModelingToolkit.jl#3458 (comment), but will do and hope for now that the other stuff gets solved as the upstream fixes filters through

[isaacsas]

@ChrisRackauckas this test:

https://github.com/SciML/Catalyst.jl/blob/master/test/simulation_and_solving/simulate_jumps.jl#L179

is still failing. Some issue with initialization and variable jump problems (i.e. ODEProblem over JumpProblem when there are no ODEs).

[isaacsas]

Here is the error: https://github.com/SciML/Catalyst.jl/actions/runs/14032571458/job/39283295596?pr=1186#step:6:701