Francesco Casella
Francesco Casella
I moved the discussion to #4479 to avoid spillover.
Given how the discussion in #4479 goes, maybe I should re-open this ticket. We do have a regression in OpenModelica, due to #4042, which changed from MSL 4.0.0 to 4.1.0....
I'd like to hear opinions outside the set of Dymola developers, if possible 😃
I thought a bit more about this issue, and I think I can articulate a more compelling argument in favour of changing the defaults of this model than "it doesn't...
> That is not correct reasoning according to IEEE-754; it depends on **how much you add**, in this case the addition is insignificant and doesn't impact the result. IEEE-754 is...
> I think the problem is identified, "solved". [#4478 (comment)](https://github.com/modelica/ModelicaStandardLibrary/issues/4478#issuecomment-2413355601) That's what I thought, until I went through a [more thorough analysis](https://github.com/modelica/ModelicaStandardLibrary/issues/4478#issuecomment-2420676004) later on. I hope I am entitled to...
> As far as I previously understood the bug in the OpenModelica compiler is that `if time
@dzimmer I agree with your comments in #4479 _for the long term_, regarding how we could handle infinity better than we do now, both in Modelica 3.7 and MSL 4.2.0,...
I wish some of the summoned reviewers (@AHaumer, @christiankral, @kristinmajetta, @christophclauss) of #4481 also agrees, so we can move forward with the 4.1.0 release.
Maybe after all #4481 is not necessary, see discussion in #4503