Allow setting continuous states in Event Mode if reinit = false

[t-sommer]

fixes #1950

[KarlWernersson]

did some changes

[t-sommer]

Should setting (all) variables with <<causality>> = <<parameter>> really be allowed? I suggest to change

This function can be called for variables with <<causality>> = <<input>>, <<causality>> = <<parameter>>, <<variability>> = <<tunable>>, and <<state,continuous-time states>> with <<reinit>> = `false`.

to

This function can be called for variables with <<causality>> = <<input>>, <<variability>> = <<tunable>>, and <<state,continuous-time states>> with <<reinit>> = `false`.

[chrbertsch]

Torsten S: I want to change the second comma do "and" .
Masoud: I will look into this regarding clocks.

[masoud-najafi]

May be I'm missing something here. Acccording to the definition of "reinit", if reinit=false, you are not allowed to reinitialize the state variable. So you cannot call fmi3SetXX. Is it a typo or I'm missing something?

[t-sommer]

reinit:

If true, state may be reinitialized by the FMU in Event Mode.
If false, state will not be reinitialized.

I think this is correct, since an importer may only set a continuous state in Event Mode, if the FMU does not recalculate it.

[masoud-najafi]

OK, I got your point. Then let's first clarify the reinit attribute. Is it just for information purposes? What's useful for?
An FMU may go to the Event mode for several reasons. For a particular event, the FMU may reinitialize the state with reinit==true). For other events, the "importer" may reinitialize the same state. Let's take the BouncingBall example. The FMU is supposed to change the height and the velocity at zero-crossing event instants, so reinit for height and velocity is true. The importer may also be able to change these state variables at other events (other that zero-crossing events), for example if t>10, the ball change the velocity due to a sudden external wind.
By saying "reinit=true", should the importer be forbidden from updating the state? I'm not really in favor of this.

[KarlWernersson]

@masoud-najafi I see your point, i don't have a strong requirement that you are forbidden to set continuous time state with reinit = true, its a larger chance that you screw things up though, but this is still an advanced feature so i am ok with letting all states be set. from the master.

[chrbertsch]

FMI Design Webmeeting:

Torsten: are clocks covered by this, or is this unrelated?
Karl: yes, it is unrelated.
Klaus: it is fine, but an explicit bullet list would be more clear.
Torsten: this would be geeneral issue at several places.
Torsten: this PR is about making things consistent. Masoud's suggestion goes beyond this. @masoud-najafi : if you want to proceed with this, please open a new PR
Klaus, Karl: this would not be a bugfix.