Joel Andersson

I'm not sure what it would look like or even if it would make sense. I would not start with it.

I can reproduce this error with the BaseModelica branch of CasADi

> Hey @jaeandersson I also noticed my NLPs fail when I compile the FMUs in OpenModelica and I use the flag -d=-disableDirectionalDerivatives. Basically I get a SIGABRT (signal 6) error,...

This issue should be resolved in https://github.com/casadi/casadi/commit/4c2dbd3b5c3562191967acc3ffcb9620062dcd41. @icupeiro - feel free to check if it also resolved the OpenModelica issue or if that's unrelated. If unrelated, feel free to open...

> This issue should be resolved in [4c2dbd3](https://github.com/casadi/casadi/commit/4c2dbd3b5c3562191967acc3ffcb9620062dcd41). @icupeiro - feel free to check if it also resolved the OpenModelica issue or if that's unrelated. If unrelated, feel free to...

Is it in a header file? If so, the inline declaration can be used to avoid linker conflicts.

> Is it in a header file? If so, the inline declaration can be used to avoid linker conflicts. Actually, nevermind. Since they are templated functions, it shouldn't be necessary...

Hmm, this sounds like a tricky thing to support. Do we want `v[[False, True, False]]` to behave differently from `v[[0, 1, 0]]`? I'm not sure.

`casadi.DM` is using MATLAB-syntax, not `numpy.ndarray` syntax (which isn't even everything-is-a-matrix), so that's not a good argument. Anyway, logical indexing is working in MATLAB: ```matlab v = [1;2;3]; v([false;true;false]) %...

Same bug for trilsolve ``` g dep(1).sparsity(), g.work(arg[1], this->dep(1).nnz()), ```