add radius threshold to box selections #1607
Conversation
rts/Sim/Objects/SolidObject.h
Outdated
| @@ -383,6 +383,7 @@ class CSolidObject: public CWorldObject { | |||
| LocalModel localModel; | |||
| CollisionVolume collisionVolume; | |||
| CollisionVolume selectionVolume; | |||
| float selectionRadius = 0.0f; | |||
There was a problem hiding this comment.
Isn't there some neat convenience function to check CollisionVolume's distance to a plane? It would be good not to converge on a worse solution (for example only being able to do a sphere would fail hard for elongated units).
There was a problem hiding this comment.
Unfortunately only a distance to point exists
|
It's tempting to propose you use |
|
Using the existing selection volume (without an opt-out) would be pretty bad, see the middle section (with sc2 protoss screenshots) of #671 (comment) |
Co-authored-by: sprunk <[email protected]>
Co-authored-by: sprunk <[email protected]>
Perhaps you're right. |
rts/Sim/Units/Unit.cpp
Outdated
| @@ -307,6 +307,8 @@ void CUnit::PreInit(const UnitLoadParams& params) | |||
|
|
|||
| harvestStorage = unitDef->harvestStorage; | |||
|
|
|||
| selectionRadius = unitDef->selectionRadius; | |||
There was a problem hiding this comment.
Did you copy this as a separate variable, so it could be changed on a per-unit basis?
Otherwise u->unitDef->selectionRadius might work equally well in places you now use u->selectionRadius.
Not a big deal, just curious.
There was a problem hiding this comment.
I think I remember seeing some gadget that would scale unit sizes, maybe from sprung? Although I do not know how well supported that is with other engine things like collision volumes.
|
LGTM overall 👍 |
|
This needs a proper box/sphere intersection test in order to not get false positives in the drag selection. |
Adds a unitdef for a spherical threshold for box selecting units. A default value of 0 leaves the current behavior unchanged.
#671