David Widmann
David Widmann
I didn't focus on the `return` component too much actually when copying the example. I was more thinking that perhaps the meaning of `:=` is less clear than a more...
I don't think it's a good idea but I want to bring it up anyways: What about the opposite behaviour and making everything that is defined with `=` instead of...
> if I may pull you in here? Could you see something like that happen (StatsBase depending on a light ArraysOfArrays)? I would prefer if this would be solved language-wide...
> Still, JuliaLang/julia#32310 will not bring efficient vectors of similar matrices (e.g. to represent multiple samples of matrix-variate distributions) and so on, from what I understand. It will support them...
> Interestingly, the cost of using Requires is quite low in this case. As expected, the timings show that Requires increases loading times in particular if the conditional dependencies are...
@quinnj Bump 🙂 I would be very interested in the additions to the API (and the improvements, of course).
This should be possible with `GeneralDomainCallback`. You can e.g. define the residual function such that the residuals for unconstrained states are always 0, that residuals of negative states are 0...
Thank you!
Can be avoided by setting `save_everystep=false` (done in https://github.com/SciML/DiffEqCallbacks.jl/pull/111).