[JASS] Fastest way to check for invulnerability?

[Quantum_Menace]

Am I missing something here? Of all the unittype variables none of them correspond to "invulnerable".

What's the fastest way to determine if a unit is invulnerable? Checking for a long list of buffs and abilities that make it invulnerable hardly seems efficient, even if it only bothers to check units determined to be spell immune (apparently invulnerable units are always magic immune).

 

[spiwn]

Just wondering how many invulnerability spell will you have in your map, in order for the long list to be ... long.

 

[donut3.5]

GetUnitAbilityLevel(yourUnit, 'Avul') == 1

This is the condition for it having the invulnerable ability, but I'm not sure if it detects whether it was made invulnerable via triggers.
--donut3.5--

 

[spiwn]

For having the ability?
Is it activated?
Check for buffs. When making one unit invulnerable via triggers add some buff.
To make it easier make a function that enlists all the buffs for invulnerability that returns false or true and call it whenever you want to see.
function inv takes unit a returns boolean
if(UnitHasBuffBJ(a, 'Bvul') == true) then
return false
else
call DoNothing()
endif
if(UnitHasBuffBJ(a, 'BHds') == true) then
return false
else
call DoNothing()
endif
if(UnitHasBuffBJ(a, 'BOvd') == true) then
return false
else
call DoNothing()
endif
return true
endfunction

 

[Quantum_Menace]

I'm currently doing that with a list of "and"'s, and am not aware of any triggers causing invulnerability. There are several buffs (Divine Shield, potion of invulnerability, Cyclone, Tornado Spin, Big Bad Voodoo). These may not be in the map, but there may be other custom buffs instead because I'm not the only one working on the map.

Even that's not ideal because some other abilities have some sort of transient invulnerability effect (Mirror Image, Impale, Sleep). I was wondering if there was any method that considered all cases.

I suppose I could check if a dummy is able to be ordered to use a channel spell on the unit that targets spell immunes but not invulnerables, but I'm not sure how slow that would be even if it only checks units found to be spell immune.

 

[donut3.5]

UnitHasBuffBJ? Ew. Use GetUnitAbilityLevel, buffs count as abilities too.
You could also try damage the unit for one point and see if its health changes.
--donut3.5--

 

[Quantum_Menace]

Yes, I'm using GetUnitAbilityLevel.

I would have expected the damage suggestion not to work because damage events always seem to be set off only when the elapsed time is a multiple of .125 seconds, so I looked up the armor detection system which uses this method and found that it has a "IsUnitInvulnerable" function based on it.

Thanks!

Also, is there any difference in speed or reliability between "is GetWidgetLife(u) < 0.405" and "IsUnitType(u, UNIT_TYPE_DEAD)" ?

 

[donut3.5]

I just use > 1, since a unit will pretty much never have between .999 and .406 health.
--donut3.5--

 

[Troll-Brain]

I just use > 1, since a unit will pretty much never have between .999 and .406 health.
--donut3.5--

that's may be false, if you damage it with a trigger (equation)

 

[donut3.5]

True, but it should be extremely rare, and considering health regeneration, the bug will hardly be noticeable.
--donut3.5--

 

[Troll-Brain]

True, but it should be extremely rare, and considering health regeneration, the bug will hardly be noticeable.
--donut3.5--

According to you for the rare, but not for the regeneration, even if the unit has a very fast life regeneration, if his health is less or equal to 0.405; then she dead.
And if you already use >1 why not >0.405 ?

 

[donut3.5]

Because I never bothered to remember the exact value before And 1 is easier to remember/type.
--donut3.5--

 

[Troll-Brain]

Because I never bothered to remember the exact value before And 1 is easier to remember/type.
--donut3.5--

use many times and you won't never forget it any more ^^

 

[PurplePoot]

Troll-Brain, there is no practical application I have ever found for using .405 instead of 0.

 

[Captain Griffen]

...except that you are less likely to get random bugs relating to units being dead, but not according to your triggers.

 

[Dr Super Good]

Damage events fire as soon as the damage is inflected but before the actual damage is delt. Thus you can get the damage inflected but the hitpoints will be as if it was not in conditions or actions run by the event.

It also seems that units are only dead with 0 HP. If the HP are less than 1 then it automatically rounds to 0 or something like that and thus a unit dies and its hp are set to 0. A unit can not have negetive HP unless you exceede the maximum number capacity of the HP object in which case you are really stupid and your maps badly made.

However in games like supreme commander where it rounds properly, it is actually possiable to get a unit with 0 HP out of however many and it will still be alive.

 

[Troll-Brain]

Troll-Brain, there is no practical application I have ever found for using .405 instead of 0.

oh you can ignore it, but an unit with less or equal to 0.405 hp is dead, it's a fact

 

[spiwn]

Then is a unit with 0.406 alive?

 

[PurplePoot]

oh you can ignore it, but an unit with less or equal to 0.405 hp is dead, it's a fact

It's also a fact that a dead unit has 0 hp, making that entire theory entirely worthless unless you do some really weird stuff.

 

[Troll-Brain]

Then is a unit with 0.406 alive?

yes

It's also a fact that a dead unit has 0 hp, making that entire theory entirely worthless unless you do some really weird stuff.

ofc, if you don't use any damage system with GetEventDamage for example, then you don't need to worry about

 

[PurplePoot]

ofc, if you don't use any damage system with GetEventDamage for example, then you don't need to worry about

Which raises the question as to why all the wc3cers insist on GetWidgetLife(u) <= .405 for dead checks >>

It also doesn't matter unless you're checking to see if they are still alive, as I said, very specific situations.

 

[Dr Super Good]

Another simple way to check if a unit is alive is to check if HP are not equal to 0. Seems to work.

 

[donut3.5]

function IsUnitDeadBJ takes unit whichUnit returns boolean
    return GetUnitState(whichUnit, UNIT_STATE_LIFE) <= 0
endfunction

This is how Blizzard checks whether a unit is dead, and they use 0.
--donut3.5--

 

[HINDYhat]

Blizzard's Jass also fails in numerous ways, donut.

Also, GetWidgetLife is faster than GetUnitState

 

[Troll-Brain]

if you use the gui function ok, but if you use the native one, what is wrong with use 0.405 instead of 0.
More hard ?

And according to HINDYhat it's faster and just try you will see that it's not a random number ...

 

[PurplePoot]

if you use the gui function ok, but if you use the native one, what is wrong with use 0.405 instead of 0.
More hard ?

It's longer, and yields no advantages 99.999% of the time.

And according to HINDYhat it's faster

Did you even read HINDY's post?

and just try you will see that it's not a random number ...

It isn't random, but as I said, it's still useless 99.999% of the time.

 

[donut3.5]

Now for the big money question:
Which is more insignificant, the bits of memory preserved by the 2 fewer numbers (0. vs. .405) or the .000001% usefulness?
--donut3.5--

 

[PurplePoot]

Seeing as the .0001% usefulness only occurs in cases on the code side, not the game side, it's definitely the former, at least in 99.999% of cases

 

[Troll-Brain]

Well well, end of flood for me use 0 if you want ...

 

[Quantum_Menace]

Meh, at this rate I'll just avoid this argument and keep using UNIT_TYPE_DEAD unless someone has a reason not to.

 

[Dr Super Good]

Quantum_Menace I think it is slower.