Damage Engine 5.A.0.0

[Bribe]

That said, the native looks pretty rudimentary. I assume it would still return true for triggered damage of DAMAGE_TYPE_NORMAL.

I think it's based on the boolean isAttack in UnitDamageTarget. If it is true, BlzGetEventIsAttack is true.

 

[chopinski]

I think it's based on the boolean isAttack in UnitDamageTarget. If it is true, BlzGetEventIsAttack is true.

unfortunately i dont think that is the case. If i remember from my testings the boolean value in the damage native is pretty much useless and the EventIsAttack antive was not considering it. Could be wrong though.

 

[runi95]

I downloaded the Damage System map and tried using the various units in the map.
In most cases it displays a texttag with the damage amount on it as intended, but I noticed some inconsistencies:

1. Casting Frost Nova on an enemy unit shows no damage tag even though it does deal damage to the unit(s).
2. Attacking a unit with the Dryad only displays the Slow Poison damage over time, but not the initial attack damage dealt.
3. Sea Witch's Frost Arrow damage is displayed with auto-cast off, but turning it on removes the damage display.
4. No damage is shown when Firelord's Incinerate ability has auto-cast toggled on, but with auto-cast off it displays a texttag again.
5. Letting the Phoenix hit an enemy unit with its Phoenix Fire ability does not display any initial damage and only shows damage over time effect.
6. Any hero with Orb of Corruption, Orb of Frost or Orb of Venom will also lose their damage display (as well as Orb of Slow, but only when the 25% chance to slow hits)

 

[Uncle]

I downloaded the Damage System map and tried using the various units in the map.
In most cases it displays a texttag with the damage amount on it as intended, but I noticed some inconsistencies:

1. Casting Frost Nova on an enemy unit shows no damage tag even though it does deal damage to the unit(s).
2. Attacking a unit with the Dryad only displays the Slow Poison damage over time, but not the initial attack damage dealt.
3. Sea Witch's Frost Arrow damage is displayed with auto-cast off, but turning it on removes the damage display.
4. No damage is shown when Firelord's Incinerate ability has auto-cast toggled on, but with auto-cast off it displays a texttag again.
5. Letting the Phoenix hit an enemy unit with its Phoenix Fire ability does not display any initial damage and only shows damage over time effect.
6. Any hero with Orb of Corruption, Orb of Frost or Orb of Venom will also lose their damage display (as well as Orb of Slow, but only when the 25% chance to slow hits)

I'm pretty sure the Demo Triggers folder is meant to be used as an example that you can follow in your own map. It's not like you're intended to just copy and paste his demonstrations.

Anyway, I imagine this trigger would work for all cases:

• Damage As Text
  • Events
    • Game - DamageEvent becomes Equal to 1.00
  • Conditions
  • Actions
    • Set VariableSet DamageString = (String(DamageEventAmount))
    • Game - Display to (All players) the text: (String(DamageEventAmount))

You can do whatever you want with the DamageString variable, display it in Floating Text, use a 3rd party Text system like the one in the demo, etc.

 

[runi95]

I'm pretty sure the Demo Triggers folder is meant to be used as an example that you can follow in your own map. It's not like you're intended to just copy and paste his demonstrations.

Anyway, I imagine this trigger would work for all cases:

• Damage As Text
  • Events
    • Game - DamageEvent becomes Equal to 1.00
  • Conditions
  • Actions
    • Set VariableSet DamageString = (String(DamageEventAmount))
    • Game - Display to (All players) the text: (String(DamageEventAmount))

You can do whatever you want with the DamageString variable, display it in Floating Text, use a 3rd party Text system like the one in the demo, etc.

Yes, the demo triggers just being examples is a fair point, but it does not change the fact that I expect them to display correct information.

What if I wanted to display the damage output of all spells and attacks? Then both the demo trigger and the one you've supplied me with here fails to display the correct damage output for the Dryad. They will both output "4.0" when the correct output should be "14.0" (attack damage) + "4.0" (slow poison).
If you try to cast Frost Nova with the trigger you've supplied me with here then it does not produce an output at all, but Frost Nova does in fact deal damage in reality.

 

[Uncle]

Yes, the demo triggers just being examples is a fair point, but it does not change the fact that I expect them to display correct information.

What if I wanted to display the damage output of all spells and attacks? Then both the demo trigger and the one you've supplied me with here fails to display the correct damage output for the Dryad. They will both output "4.0" when the correct output should be "14.0" (attack damage) + "4.0" (slow poison).
If you try to cast Frost Nova with the trigger you've supplied me with here then it does not produce an output at all, but Frost Nova does in fact deal damage in reality.

Try messing around with different Events. The one I used was what I've personally never had any issues with.

There's always the possibility that there's an actual bug in the system but I find that unlikely.

 

[runi95]

Try messing around with different Events. The one I used was what I've personally never had any issues with.

There's always the possibility that there's an actual bug in the system but I find that unlikely.

I sort of figured it out. I tried all the events and none of them worked with Frost Nova or Dryad, but I was using the "Lua Damage Engine 2.0.w3x" since the map I wanted to use this system for is a LUA map.

After trying all the events I switched to the other map that's included in the same downloaded .zip. And everything works as expected in the "Damage Engine 5.A.0.0.w3x" version of the map (even the demo / example triggers). Therefore I suspect there's a bug in the LUA version of the map as I expect the events to behave the same in both maps.

 

[runi95]

For anyone else wanting to use the LUA version of this library I've found out that the library errors in 2 places during events like Frost Nova, Frost Arrow, Slow Poison, Incinerate (and many more).

local function failsafeClear()
    setArmor(true) -- 1st error occurs inside this call
    canKick = true
    kicking, totem = nil, nil
    runEvent(_DAMAGED) -- 2nd error occurs inside this call
    eventsRun = true
    finish()
end

1st error if _USE_ARMOR_MOD config is set to true:

local function setArmor(reset)
    if _USE_ARMOR_MOD then
        if reset then -- We already know reset = true from failsafeClear()
            pierce = current.armorPierced -- 1st error occurs here as "current" is nil
            -- ... more code that never gets to run
        end
    end
end

2nd error always occurs no matter what config you have:

local function damageOrAfter() return current.damageType == DAMAGE_TYPE_UNKNOWN end -- 2nd error occurs here as "current" is nil
breakCheck[_DAMAGED]    = damageOrAfter
local function runEvent(head)
    local check = breakCheck[head] or defaultCheck
    if dreaming or check() then -- "dreaming" = false here so it proceeds to "check()"
        return
    end
    -- ... more code that never gets to run
end

The easiest soultion seems to be modifying the failsafeClear function to this:

local function failsafeClear()
    if current ~= nil then
        setArmor(true)
    end
    canKick = true
    kicking, totem = nil, nil
    if current ~= nil then
        runEvent(_DAMAGED)
    end
    eventsRun = true
    finish()
end

The JASS version of the library already checks for udg_DamageEventDamageT != 0 before running DAMAGE.run() within the failsafeClear method and it skips the issue with USE_ARMOR_MOD because it only references non-nullable global variables.

 

[Bribe]

Thanks for the detailed feedback, that's great. Lua Damage Engine has a lot of issues beyond that, and if I ever code in Lua again I'll certainly look into it, but I am waiting for Sumneko to make some needed QoL changes to LLS. I also can't use TypeScript2Lua because it does not do metatables natively, which resources like Hook and GlobalVariableRemapper need.

 

[runi95]

That's fair. I've actually been using this damage engine for several years now in a TypeScript project, but it was based off of an older version of this library (originally I rewrote the GUI version to TypeScript) and coming back to modding I wanted to see if there was a newer version I could use. Since it's a Lua map that's the one I tried to use, but it turned out to have issues. Since hooks and global variable remapping aren't useful for me they're not really a concern and fixing these nil references + adding a basic event table I can reference seems to make it work well for me. At least the little bit I've tested it so far.

 

[window789456]

For the Damage Engine folder from the version 5.A.0.0.w3x, I copy pasted it to my map without those Demo Triggers, and then these errors popped up, how do I fix them?

 

[Asuna Tichondrius]

Hello, Can it work on 1.30.4 version?

 

[Spacebuns]

Is the variable DamageEventPrevAmt is deprecated or still in use? I actually don't know what it does now.

 

[ThompZon]

Hello, Can it work on 1.30.4 version?

You'd need to use the "legacy edition" of the damage engine found at: Damage Engine 3A.0.0.0 and 3.8.0.0
Damage Engine has since version 5.0 required the new natives from WC3 version: 1.31 IIRC.

 

[ThompZon]

Is the variable DamageEventPrevAmt is deprecated or still in use? I actually don't know what it does now.

Under the "features" in the description, it say "DamageEventPrevAmt: the amount of damage prior to being modified by the game engine".
I don't know if I've really used it, but I can see many good usecases for this value.

In other words: This is a "read only" value that might be useful for "boost/reduce base value of the attack".

From what I can see in the vjass code, it seems to work like the following:

Spoiler: Example

If I am correct:

You have a unit dealing 100 damage, with bonus 25% of target's missing hp and a 25% chance to critical hit for 2x damage.

You don't want this low-health bonus to also critial hit and don't really want to think about the order of triggers that much because it can get confusing and also change over time if more triggers are added.

Then you can in your critical hit trigger have this if(random(0,1) > 0.25) { DamageEventAmount = DamageEventAmount + DamageEventPrevAmt } (adds the "base value" of the attack to the current damage. Possible to be more sophisticated, but this is an example)
Execute trigger be DamageEventAmount = DamageEventAmount + 0.25 * (Unit State Max Health for (udg_DamageEventTarget) - Unit State Current Health for (udg_DamageEventTarget))

Because the crit-trigger use DamageEventPrevAmt instead of the DamageEventAmount directly, it don't care if the "execute trigger" is run before or after.

If source deals 100 damage, target has 123 missing hp, and source crits: (100 + 0.25 * 123 + 100) = 230.75
If crit used DamageEventAmount directly with the same values it'd deal either: (100 + 0.25 * 123) * 2 = 261.5 OR (100 * 2) + 0.25 * 123 = 230.75 depending on how you setup the event trigger's "value" defining the order of operation.

 

[Spacebuns]

The DamageFilter variables don't seem to work even though I use the same trigger structure in the sample map provided.

 

[Spacebuns]

Btw, does the numerical convention still work like this? For example
"PreDamageEvent = 1.00" runs before "PreDamageEvent = 1.50" runs before "PreDamageEvent = 2.00" runs before "ArmorDamageEvent = 1.00" runs before "ArmorDamageEvent = 2.00" and so on.

 

[Kakerate]

how does Armor Pierce work in this? To test, I created a knight and removed the proc rate from his ability.

call UnitDamageTargetBJ(knight, u, 100., ATTACK_TYPE_MELEE, DAMAGE_TYPE_NORMAL)

with 1000 Armor Pen deals 73 and some change damage. This doesn't seem right. Even weirder, when you change the maps' Armor Constant from 0.06, the damage dealt here changes? Changing to .02 resulted in 89 damage dealt. Setting Armor Constant to negative values actually increases the damage they take, (154 @ -6%). What is intriguing me is that there is no way to grab this Constant afaik, but changing it does appear to be doing something (more precisely, in the opposite direction of what you'd expect in some tests).

(removed a ton here, I think my issue was calculating eHP based off the % rather than just interpreting stuff like -20% Reduction as taking 20% more. Also, seems WC3's negative armor "cap" is -20, at 170% taken)

=============
my suggestions;



Negative Armor (for custom Armor Constants) is problematic, if my understanding is correct. Iirc, The formula for negative armor 2-0.94^( -armor ) is independent of Armor Constant. This means in maps with lower than 6%, that negative Armor is still as effective as in base WC3, and vice versa. Functionality that would allow users to correct this would be nice (in my map, however, I pretty much just factored for this so "fixing" it isn't necessary).

If a cap on the penalty from negative Armor exists (or even if it doesn't), functionality that allows you to have either behavior would be nice as well.

 

[Bribe]

Too bad I'm not Blizzard.

 

[Kakerate]

it is a bit unfortunate that "special" damage can't be handled by the system currently. ATTACK_TYPE_NORMAL and DAMAGE_TYPE_MAGIC gave a damageType reduced by %SpellResist and Armor. This damage doesn't apply Armor Pierce. Can rework some effects to account for this, but it is a bummer nonetheless. I get approaching from WC3s angle of attack types being relegated to the damage table, but for custom maps that don't care about it, you are essentially using attack/damage types for deciding what resists to apply. A flag to apply ArmorPierce regardless of damageType would probably suffice.

 

[Bribe]

it is a bit unfortunate that "special" damage can't be handled by the system currently. ATTACK_TYPE_NORMAL and DAMAGE_TYPE_MAGIC gave a damageType reduced by %SpellResist and Armor. This damage doesn't apply Armor Pierce. Can rework some effects to account for this, but it is a bummer nonetheless. I get approaching from WC3s angle of attack types being relegated to the damage table, but for custom maps that don't care about it, you are essentially using attack/damage types for deciding what resists to apply. A flag to apply ArmorPierce regardless of damageType would probably suffice.

armor doesn't apply at all unless you're using the physical damage type. To get around this, @SpasMaster uses physical damage types for every spell, and it allows WarCraft 3's armor to act as magic reduction.

Another solution is to code more robust magic reduction or amplification, perhaps with a UX mod achieved using frames so the user knows visually how much magic volatility they have.

 

[Daffa]

Is Pure damage truly unstoppable except with Invulnerability? Does Banish or Spell Immunity interact with Pure?

 

[Bribe]

Is Pure damage truly unstoppable except with Invulnerability? Does Banish or Spell Immunity interact with Pure?

DAMAGE_TYPE_UNIVERSAL can hit spell immune targets, so yes invulnerability is the only way to stop it.

I can't remember if Banish works to amplify spell damage from UNIVERSAL. If the attack type is MAGIC it should amplify in principle.

 

[Marchombre]

Hello Mr Bribe,

I'm getting confused by the behavior of IsDamageAttack.

Context:

---

I'm trying to make an ability than damage enemies and is considered as an attack (which allow to trigger on attack hit effects).
My function calls UnitDamageTarget to damage the targets with the following parameter.

call UnitDamageTarget(druid, u, dmg, true, false, ATTACK_TYPE_NORMAL, DAMAGE_TYPE_ENHANCED, WEAPON_TYPE_CLAW_HEAVY_SLICE)

To my understanding, UnitDamageTarget has the following parameters:

• 1/ unit source - who does the damage
• 2/ unit target - who receive the damage
• 3/ real amount - how much damage
• 4/ boolean attack - is the damage from an attack ?
• 5/ boolean ranged - is the damage from a ranged attack/spell ?
• 6/ attacktype attackType - which attack type
• 7/ damagetype damageType - which damage type
• 8/ weapongtype weaponType - which weapon type

With number 4/ what's interesting here.
My function was triggered, and properly called with parameter 4/ set to true. Damage of type DAMAGE_TYPE_ENHANCED is correctly applied.

However, after testing, I find:

• IsDamageCode = true. Expected as I'm calling from Jass function.
• IsDamageAttack = false. This is breaking my brain right now.
• IsDamageSpell = true. Edit: Expected as ATTACK_TYPE_NORMAL is spell in GUI. Tried with ATTACK_TYPE_HERO to make sure, IsDamageSpell = false but IsDamageAttack = false still.

Maybe it's just that my brain stopped working at the late hour
Am I missing something obvious ?
Isn't the 4th parameter what I think it is ?

---

PS: here is the trigger used for testing.

• Events
  • Game - ArmorDamageEvent becomes Equal to 1.00
• Conditions
  • DamageEventAmount Greater than 0.00
• Actions
  • Game - Display to (All players) for 3.00 seconds the text: ((Name of DamageEventSource) + ( is damaging + (Name of DamageEventTarget)))
  • If (All conditions are True) then do (Then Actions) else do (Else Actions)
    • If - Conditions
      • IsDamageCode to True
  • Then - Actions
    • Game - Display to (All players) for 3.00 seconds the text: CodeDamage
  • Else - Actions
    • Game - Display to (All players) for 3.00 seconds the text: Not Code Damage
  • If (All conditions are True) then do (Then Actions) else do (Else Actions)
    • If - Conditions
      • IsDamageAttack to True
  • Then - Actions
    • Game - Display to (All players) for 3.00 seconds the text: Attack Damage
  • Else - Actions
    • Game - Display to (All players) for 3.00 seconds the text: Not Attack Damage
  • If (All conditions are True) then do (Then Actions) else do (Else Actions)
    • If - Conditions
      • IsDamageSpell to True
  • Then - Actions
    • Game - Display to (All players) for 3.00 seconds the text: Spell Damage
  • Else - Actions
    • Game - Display to (All players) for 3.00 seconds the text: Not Spell Damage
  • Game - Display to (All players) for 3.00 seconds the text: 'Damage Type: + ((String(DamageEventDamageT)) + ( (Enhanced is: + ((String(DAMAGE_TYPE_ENHANCED)) + ) ))))

 

[Spacebuns]

DAMAGE_TYPE_UNIVERSAL can hit spell immune targets, so yes invulnerability is the only way to stop it.

I can't remember if Banish works to amplify spell damage from UNIVERSAL. If the attack type is MAGIC it should amplify in principle.

If a unit is in Ethereal form (Banished) it doesn't seem to take damage of Universal type.

 

[Bribe]

If a unit is in Ethereal form (Banished) it doesn't seem to take damage of Universal type.

ATTACK_TYPE_NORMAL (Spells in GUI) and ATTACK_TYPE_MAGIC are the only attack types that can hit ethereal.

 

[Daffa]

DAMAGE_TYPE_UNIVERSAL can hit spell immune targets, so yes invulnerability is the only way to stop it.

I can't remember if Banish works to amplify spell damage from UNIVERSAL. If the attack type is MAGIC it should amplify in principle.

This should be sufficient for my need. Thank you

 

[Kakerate]

I'm getting confused by the behavior of IsDamageAttack.

Never even thought to test damaging a unit with true for isAttack, (something about my WC3 instinct makes me not surprised this doesn't work), but the way you fix is by setting a flag before dealing the damage, then modifying the condition for onHit effects to be;
if ( FLAG or BlzGetEventIsAttack() ) then

 

[Marchombre]

Never even thought to test damaging a unit with true for isAttack, (something about my WC3 instinct makes me not surprised this doesn't work), but the way you fix is by setting a flag before dealing the damage, then modifying the condition for onHit effects to be;
if ( FLAG or BlzGetEventIsAttack() ) then

Thanks for the reply.
I forgot to edit my answer, but found that setting udg_NextDamageIsAttack = true before calling the function will do the trick, doing effectively what you're describing inside the damage engine.

I'm still confused on what the 4th argument of UnitDamageTarget is doing, but at least I can do what i wanted