Robert Schwarz

It seems that in [CSIP](https://github.com/SCIP-Interfaces/CSIP/blob/master/src/csip.c#L268), we were just passing a pointer to an array as the "last" argument to `SCIPexprCreate`. I did not know that this is valid. I guess...

Thanks for the hint. I understand that the issue stems from a limitation of Julia's `ccall`, so I would be surprised to find a solution in a Julia package, but...

OK, great, so it does support variadic arguments of mixed types properly :open_mouth:

I have never used SCIP-SDP before. I'm not sure whether there is support of building a shared library (like `libscipopt.so`) with it, currently. If this is attempted, I suggest not...

I should not be mentoring this year, either. But I wanted to bring it up anyway, since @Ayush-iitkgp seemed to be interested in working on SCIP.jl.

Not sure, the SCIP team did not show much interest in Julia so far, and they invest their time in the [Python interface](https://github.com/SCIP-Interfaces/PySCIPOpt) instead.

See also discussion in https://github.com/JuliaOpt/CPLEX.jl/pull/207 for a candidate interface that could be used as a template.

My personal motivation would be to expose constraint handler plugins first (in full generality) with a thin convenience wrapper (à la lazy constraint callback), in the coming weeks. Then heuristics,...

For constraint handlers, there are 3-4 [required callbacks](https://scip.zib.de/doc-6.0.1/html/CONS.php#CONS_FUNDAMENTALCALLBACKS): - `check`: given a solution candidate, check constraint violation and return boolean - `enfo`(rce): given a solution candidate, iff constraint is violated,...

> If I understand correctly, after you register a constraint handler for X-type constraints, you can add arbitrarily many X-type constraints, right? That is correct, although it is also optional....