Guido Tack

This is the intended behaviour. Only variables declared without a right hand side are part of the default output. The rationale is that when using `--output-mode dzn` you can simply...

The problem is that, strictly speaking, assignments to variables shouldn't really be part of data files at all. The MiniZinc type checker has always been quite lenient about what it...

The problem is not so much the variable if-then-else, but the let expression (for which we currently can't compute proper bounds). I'll put this on our todo list.

We can change FlatZinc and still support older solvers by adding optional configuration parameters into the library. For example, we could add `opt bool: fzn_solver_supports_compact_sets;` and then the solver could...

The `fzn_if_then_else_var_bool` calls in question do not come from a top-level if-then-else, but from one in a reified position (in the `isA` predicate, line 124). The compiler successfully removes the...

Oh, that's an annoying one. I'll need a bit more time to fix it I'm afraid.

Thanks, looks good! Just a few things to fix: - To be included in the library, you would have to separate the actual decomposition from the modelling interface, by splitting...

MiniZinc currently compiles the product expression separately and tries to infer bounds for the result. Since the compiler can't see that the result is bounded, it runs into this overflow...

The next version (and the current develop branch) can generate JSON output, which should make it quite easy to process in Python.

That's right, index sets that are not 1-based are not supported in JSON input at the moment.