about JuliaReach/MathematicalSystems.jl#164: i agree that it would be nice to return a better error message for the examples that you sent
but given the slow progress of this PR i would rather postpone it to a follow up issue
i will open the new issue
related to MathematicalSystems, can we handle from the library a case with input constraints but no state constraints? these appear quite often
but
ConstrainedAffineControlContinuousSystem restricts to both input and state
i think that there are 2 options (at least):
1) create a new type, eg.
StateConstrainedAffineControlContinuousSystem
ah yes, well but we dont have types yet
but i would not add them
just use
nothing
otherwise you have 8 combinations per system type
and the naming becomes ridiculous
it's an option to use nothing, but then the type parameter is Nothing
so you can dispatch on that
having to write
ConstrainedAffineConstrainedControlConstrainedNoisyContinuousSystem does not sound like something you want
haha true, but the idea is that
@system macro should handle this for you..
well, ok. i prefer this than adding a new type, finally. what i thought was more like using a think layer for dispatch
i mean, to use MathematicalSets.AbstractSet and defiene a universal set in that library
storing an unnecessary
nothing is not really bad
ah i see what you mean
then ill make a PR for the "nothing" cases soon..
as it was the case for the "scalar" issues, we can handle this at the level of the
@system macro or by creating new systems constructors
because it may be simpler :smile:
And also, generic algorithms do not have to consider the missing information as a special case becuase LazySets.Universe behaves as a set ;)
well https://github.com/JuliaReach/MathematicalSystems.jl/releases/tag/v0.10.0 should have been tagged v0.9.1 .. i was not too cautious about that :/
ill have this in mind for future releases
But now that we are forced to put upper bounds on Compat, doing a minor version bump creates a CompatHelper PR for each packaged depending on it ^^