Avoiding "push resistance" stucked units (develop version)

[sanguinariojoe]

It is supermegahacky, but it works!

https://youtu.be/0ZxID60eJaU

(Tested just on maintenance)

[gajop]

Sorry I'm not knowledgeable enough in the pathfinding system and no time to learn it right now,
so I won't be able to merge this without some smarter people coming in and giving this a review.

Is this an urgent issue for you?

[sanguinariojoe]

Is this an urgent issue for you?

We are having many problems with blocked units in s:44. However, I cannot say if it a good risk-benefit balance, since I don't know how many other games are using push resistance.

[sanguinariojoe]

Hey!

We (Spring-1944) can really enjoy these changes. Can someone please take a look, merge, and cherry-pick in maintenance?

I have been testing for a while, and I think the solution is robust and efficient.

[ashdnazg]

Math fix needs bumping path estimator version.
The other hack seems risky, I'll think about it.

[abma]

am i right with:

push resistence basicly is a hack to simulate small units can't push large units?

if so, shouldn't push resistence beeing removed and replaced by more realistic physics which takes mass and friction of objects into account?

can't stuck units beeing detected by the timeout of move commands and then moved by a lua gadged?

please correct me if i'm wrong! thank you!

[sanguinariojoe]

push resistence basicly is a hack to simulate small units can't push large units?

Nope. Push resistance completely disables the units pushing system, no matter the mass/size of the units. And actually that is the behavior we want in S44

can't stuck units beeing detected by the timeout of move commands and then moved by a lua gadged?

I tried this approach in the past with quite poor results. I really think this approach is the appropriate one, i.e. if 2 units are trying to get the same spot, just one of them may take it and the other one shall "keep moving"