better documentation for behavior of "edge" events scheduled for exactly the end time of a simulation run

[filchristou]

When I run the docs example https://juliadynamics.github.io/ConcurrentSim.jl/stable/#Example I get exactly the output shown.

But then if I run run again I get further output:

julia> run(sim, 2)
slow 2.0
fast 2.0

Is this the intended functionality ?
I would say that "slow 2.0" and "fast 2.0" should either be printed the first time I call run(sim, 2) or not be printed at all. In the docs you say

The simulation will then stop when the internal clock reaches 10 but will not process any events scheduled for time 10. This is similar to a new environment where the clock is 0 but (obviously) no events have yet been processed.

So I guess, "slow 2.0" and "fast 2.0" shouldn't be printed even when I call run(sim, 2) again ?

Overall, it's pretty confusing that the edge events are called only the second time I call run and I have the impression it shouldn't be this way ?

[Krastanov]

This does seem to have been done on purpose for consistency with what happens if you start with run(sim, 0)

julia> run(sim, 0)
fast 0.0
slow 0.0

I am on the fence about whether it is a good decision. On one hand, it is nice that calling run(sim, 0) executes the immediate events. It is cleaner than having to run run(sim, epsilon). On the other hand, the inconsistency you point out is annoying.

Do you know what similar tools in python and C# do in this situation?

Either way, for the moment I am reluctant to change this behavior. The inconsistency is annoying, but it is a natural consequence of the behavior at run(sim, 0). It definitely needs to be better documented though. Any PRs on this are appreciated.

[Krastanov]

And changing it would be a breaking change, which makes me even more reluctant to modify things

[filchristou]

Hm..I see.

Do you know what similar tools in python and C# do in this situation?

I don't know.