821 replies to this topic

[Ascension64]

Is externalizing multi/dual-class combinations much of a hassle?

Multi-class combinations are extremely difficult to externalise. This is because they have specific CLASS.IDS identifiers. I think DUALCLAS.2DA already handles dual-class combinations.

ITEMANIM.2da specifies (in a single column) what character sequence plays when activating an item's special ability. Right now that sequence plays whenever activating any ability for some item (for items with multiple abilities), and does not allow for attack animations to be overridden. Perhaps this could be fine-tuned (by adding more columns) to allow variations between extension headers, and allow attack animations to be overridden.

So you want something like:
2DA V1.0
-1
SEQUENCE SEQUENCE1 SEQUENCE2 OVERRIDE
A 1 2 3 4
?

What do you mean about attack animations not being overriden? Doesn't an item use usually have no animation?

I thought as according to Nythrun, STATS.IDS is not expandable.

The way that CDerivedStats is structured, no. What I would need to do is to make a second struct that would store extra stats, duplicate the current, prev, and diff states, and have their pointers store in each CCreatureObject. Certainly not easy by any means.

An effect to grant a percentage increase to all damage dealt of a selected type would be nice. Ex: Aqua Mortis and the like from IWD2.
Opcode #250 could help us out (with expanded parameters to account for damage dealt through spells/items, modifier type, etc).

This is more complicated than a single effect, because the damage opcode (or the creation of the effect) would need to be adjusted (and everything else that codes a damage-style reduction in HP) to multiply damage - which would either need to check for an existing effect opcode on the character (plus stacking) or introduce an extra element in CCreatureObject that would store the damage multiplier. I had a thought that the difficulty multiplier could be used, but this is a global feature.

* Simulacra use male casting voices, this could be fixed if spellcasting uses 0x0237 (sex) instead of 0x0275 (gender)

Yep, I have had already written down Sim's request.

* Spell Immunity-type menus (.2das) could be linked with spells (.spls) in an external .2da table so that the menu will expand to replace the spell when selecting contingencies/triggers.

Could you explain this further? I don't understand.

* Flags for spells/item abilities to allow targeting of invisible characters

This should be possible.

, ignore the 1-spell/round limit

I think this is an animation limitation, in that certain actions must only be performed a certain number of times per round, governed by the RNDBASE.BMP files.

, and (for item abilities) to be affected by aura cleansing

I can have a look at this.

Here's one I'm keen on:
The engine rolls saving throws and calculates luck bonuses before the first feature block of the extended header is applied, rather than before the first feature block where the saving throw or damage reduction is calculated. This prevents us from applying effects that give bonuses to saving throws and luck bonuses on the fly (ie. in the same extended header). This is not the case with damage/magic resistances, where they are calculated when first looked for, so this is some engine inconsistency.

If the engine could take saving throw modifiers/luck modifiers into account when placed in the same feature block, but right before the order in which the rest of the effects are applied, then it would be possible to implement AD&D saving throws vs. damage types / wisdom modifiers to saving throws against charm-related effects.

I would have thought this is the desired behaviour, otherwise the target could save against some effects in the same extended header and not to others, whereas we would want one save for the entire extended header. Imagine a charm spell going 'off' because the target saved against the Charm opcode but not the Display Portrait Icon opcode...

I don't understand the question; I think having a component in ToBExConfig asking if you want that tweak or not (like the rest of the stuff) is fine with me (I doubt this would cause compatibility issues, since sane modders wouldn't pillage the BH/Assassin unusability bit for their kit).

Sure, I think it's doable, will need to find the code itself.

Yeah, I know but that would be better to include that kind of options in creating a character. As far as I remember that kind of mods add subrace via dialogue.
However, probably it's just me who would like to see it so...

I think it is probably much less work to just make a mod that operates via dialogue to 'configure' your sub-race at the start of the game.

Some time ago I've asked in PM for no casting interruption if no damage (100% res) is taken.
Thinking more about it, could the concentration check be implemented, using creature's level as a base?

if ((level + 1d20) < damage) then interrupted=1
else interrupted=0

I did write this down, but I think it is tricky, because the interruption occurs if the character is 'hit' in the code, not when the character is damaged.

I've almost wrote a "no, not possible" post, but I decided to test it once again (I've been trying to get this to work using external .effs for years), and it sort of worked. I'm giving you credit for giving me the idea to delay the external damage-applying .eff. I would have to:
1.) externalize all damage feature blocks into .effs
2.) make sure that .eff fires away with a delay of 0 (set in the extended header)
3.) make sure the saving throw modifier is placed before the external .eff with a duration of 1 (0 doesn't work, bc it expires before the .eff is fired).
It's still not optimal, because many things can happen within that one second when the saving throw/luck roll is penalized. Furthermore, I would have to externalize every single opcode in the game, which is a pain.

I'm also giving you credit because this made me test opcode 133 [Luck Non-Cumulative] instead of 22 [Luck modifier]. Opcode 133 is applied right when the game reads through that feature block, as opposed to opcode 22 (which is applied when the game reaches the end of the feature block list). You are now my friend, GeN1e.

In summary: I'd like opcodes 22, and 33-37 to behave like opcode 133 in terms of when they are applied.

Would it work if you made a permanent stat change at the start of the feature block and reverted that change at the end of the feature block?

You could restore opcode #144 / #279, param 2 = 10 (i.e. ability to enable/disable the Find Traps button)

This can be tricky, because the button state depends on an array value in CDerivedStats, which I described earlier, needs 'tagging' with a CDerivedStatsEx or something-or-other onto the CCreatureObject. If I can get something like this sorted, it may well be doable.


So, it does look like extending STATS.IDS seems to a want that has come up a few times. I might investigate this first (after the engine bugs).

[Galactygon]

Is externalizing multi/dual-class combinations much of a hassle?

Multi-class combinations are extremely difficult to externalise. This is because they have specific CLASS.IDS identifiers. I think DUALCLAS.2DA already handles dual-class combinations.

I was thinking something in terms of externalizing the classes as combinations of different classes. ie class 7 (fighter-mage) would be composed of 2 (fighter) and 1 (mage), something like this:

MULTICLASS  CLASS1  CLASS2  CLASS3
7           2       1       0
8           1       3       0
9           1       4       0
...

So you want something like:
2DA V1.0
-1
SEQUENCE SEQUENCE1 SEQUENCE2 OVERRIDE
A 1 2 3 4
?

First column will have to be MYITEM.itm. I don't know what you mean by the last (override) column.

What do you mean about attack animations not being overriden? Doesn't an item use usually have no animation?

Vanilla ITEMANIM.2da looks like this (ironically, it does nothing to FIRESEED.itm) :

FIRESEED  8

It's going to give all extended headers of FIRESEED.itm (except the ranged attack header) the SEQ_SHOOT animation. Changing the number to 2 will give the SEQ_CAST animation to all extended headers except those labelled melee or ranged.

* Spell Immunity-type menus (.2das) could be linked with spells (.spls) in an external .2da table so that the menu will expand to replace the spell when selecting contingencies/triggers.

Could you explain this further? I don't understand.

Right now it's impossible for players to store Spell Immunity Divination in a trigger/contingency, because the selection spell immunity (SPWI510.spl) will be offered instead. Meanwhile, the AI is free to fire those spells when buffing. SCS goes around this by making the various spell immunity spells available to the player.

There can be a better way to do this: instead of having the contingencies offer SPWI510.spl, they will offer whatever spells are listed in SPWI510.2da. This will require a .2da that links SPWI510.spl with SPWI510.2da.

I would have thought this is the desired behaviour, otherwise the target could save against some effects in the same extended header and not to others, whereas we would want one save for the entire extended header. Imagine a charm spell going 'off' because the target saved against the Charm opcode but not the Display Portrait Icon opcode...

The engine isn't consistent here; I can modify magic resistance (opcode 166) and have that affect the subsequent feature blocks. The engine is not going to roll a new magic resistance check. I am merely speculating on why opcodes 22, 33-37 kick in after the last feature block of the extended header is executed.

In summary: I'd like opcodes 22, and 33-37 to behave like opcode 133 in terms of when they are applied.

Would it work if you made a permanent stat change at the start of the feature block and reverted that change at the end of the feature block?

Just made test, and none of them (opcodes 22, 33-37) kick in until after the entire extended header is processed.

-Galactygon

[Ascension64]

About 2 weapon wielding,
if adding prof (that act as last 2 stars of 2 weapon style) would be impossible to implement, than
additional bit in item, that readed by engine grants wielder effects equal to 1st star of 2 weapon style would be excellent!
BUT if follow D&D closely, than we must remove this 1st star from 2 weapon style, so if someone actially has 2 stars in it, effects would be equal to only 2nd&3rd stars from unmodded game.

We'll see, there are so many requests at the moment...

Another request is vision range

Sorry, because I'm on low bandwidth and MB limit (away from home), I am avoiding viewing imsages. Can you describe what it is you want to change with vision range?

I was thinking something in terms of externalizing the classes as combinations of different classes. ie class 7 (fighter-mage) would be composed of 2 (fighter) and 1 (mage), something like this:

I think this is a no can do. The reason, the engine makes so much use of

switch (class) { //by CLASS.IDS
case FIGHTER:
...
case FIGHTER_MAGE:
case CLERIC_MAGE:
...etc.

that new class combinations would have to require all kinds of re-writes. The engine does not treat multi-classes as class A + B, but a completely different class altogether. However, dual classes are treated as A + B, so it makes it much more customisable.

Right now it's impossible for players to store Spell Immunity Divination in a trigger/contingency, because the selection spell immunity (SPWI510.spl) will be offered instead. Meanwhile, the AI is free to fire those spells when buffing. SCS goes around this by making the various spell immunity spells available to the player.

There can be a better way to do this: instead of having the contingencies offer SPWI510.spl, they will offer whatever spells are listed in SPWI510.2da. This will require a .2da that links SPWI510.spl with SPWI510.2da.

Would you want this by effect opcode, such that effect opcode 234 (Create contingency) would check for available spells with opcode 214 (Select spell) and adjust accordingly? Anyway, wouldn't this only remove a minor nuisance, because when the contingency triggers, you get to select what spell school to protect against, which allows better fine-tuning by the player?

The engine isn't consistent here; I can modify magic resistance (opcode 166) and have that affect the subsequent feature blocks. The engine is not going to roll a new magic resistance check. I am merely speculating on why opcodes 22, 33-37 kick in after the last feature block of the extended header is executed.

Probably something to research, I imagine the engine sometimes checks the m_BaseStats while other times it checks against the DerivedStats variations.

Edited by Ascension64, 14 January 2011 - 12:46 AM.

[Jarno Mikkola]

Another request is vision range

Sorry, because I'm on low bandwidth and MB limit (away from home), I am avoiding viewing imsages. Can you describe what it is you want to change with vision range?

He wishes the fog of war(the gray not_revealed_area) to be not so much everywhere around his, but more on his sides and back.
Prepare for some really terrifyingly bad visualization:

Like this:  Instead of:
  _____        _____
 '     '      '     '
( X---->)    (   X-->)  // the "-->" is visual area range in front of the character, which is the X.
 '_____'      '_____'

Edited by Jarno Mikkola, 14 January 2011 - 01:13 AM.

[the bigg]

that new class combinations would have to require all kinds of re-writes. The engine does not treat multi-classes as class A + B, but a completely different class altogether. However, dual classes are treated as A + B, so it makes it much more customisable.

How would a Fighter -> Bard choose its unusabilities? A Fighter -> Mage uses the F/M bit once it regains its old class.

because when the contingency triggers, you get to select what spell school to protect against, which allows better fine-tuning by the player?

I'm sure it doesn't work.

[GeN1e]

The engine does not treat multi-classes as class A + B, but a completely different class altogether. However, dual classes are treated as A + B, so it makes it much more customisable.

Are they? To my knowledge, once dualclassed PC regains old abilities, the only difference between dual and multi is a CRE flag, and usability flags are identical between them.

because when the contingency triggers, you get to select what spell school to protect against, which allows better fine-tuning by the player?

You'd have to actually cast it then, after selecting, which defeats the purpose of putting it into contingency in the first place.

Edited by GeN1e, 14 January 2011 - 11:13 AM.

[Jarno Mikkola]

because when the contingency triggers, you get to select what spell school to protect against, which allows better fine-tuning by the player?

You'd have to actually cast it then, after selecting, which defeats the purpose of putting it into contingency in the first place.

This is exactly the need for the fix, that the Contingency etc. spells would read what characters spell book has, so that the Spell Immunity spell, if it's memorized, is actually any one of the casted spell it allows to cast. This would allow to instant cast them. Just like the DavidW's SCSII AI can instant cast them(because they are memorized differently).

There is also the Nahal's Reckless Dweomer.

Edited by Jarno Mikkola, 14 January 2011 - 11:45 AM.

[Galactygon]

Would you want this by effect opcode, such that effect opcode 234 (Create contingency) would check for available spells with opcode 214 (Select spell) and adjust accordingly?

I think it's safer to do this manually, and it's more customizable (some modder might want subspells a,b,c to be storeable but not subspells d,e,f). Opcodes 234,257 would look in some sort of a 2da, which links spells and 2da spell menus.

I'm not sure if storing Reckless Dweomer is possible, since you need to store extra information (maybe the LOCAL variable slot could be used?). You will have to force SPWI124.spl onto the player in addition to whatever the player picked in the menu to make sure that the spells in the menu fire off in a wildsurge. And if SPWI510.spl is included in the spell menu, we're dealing menus nested in menus.

The engine isn't consistent here; I can modify magic resistance (opcode 166) and have that affect the subsequent feature blocks. The engine is not going to roll a new magic resistance check. I am merely speculating on why opcodes 22, 33-37 kick in after the last feature block of the extended header is executed.

Probably something to research, I imagine the engine sometimes checks the m_BaseStats while other times it checks against the DerivedStats variations.

You should be able to reproduce these by placing a feature block (with a duration of 0) before any saving-throw allowing, damage-inflicting feature block.

-Galactygon

Edited by Galactygon, 14 January 2011 - 02:05 PM.

[Ascension64]

The engine does not treat multi-classes as class A + B, but a completely different class altogether. However, dual classes are treated as A + B, so it makes it much more customisable.

Are they? To my knowledge, once dualclassed PC regains old abilities, the only difference between dual and multi is a CRE flag, and usability flags are identical between them.

I made more of a conjecture here, I haven't fully looked at dual-classing, but certainly multi-class is separate.

You'd have to actually cast it then, after selecting, which defeats the purpose of putting it into contingency in the first place.

Correct if I am wrong, but if the contingency triggers, Spell Immunity is automatically cast, you just need to pick the spell school, so this does not appear to defeat the purpose of contingency, because you aren't manually casting Spell Immunity.

There is also the Nahal's Reckless Dweomer

What's that about?

Edited by Ascension64, 14 January 2011 - 09:04 PM.

[GeN1e]

Correct if I am wrong, but if the contingency triggers, Spell Immunity is automatically cast, you just need to pick the spell school, so this does not appear to defeat the purpose of contingency, because you aren't manually casting Spell Immunity.

Spell Immunity is a double - first the zero-speed SPWI510 opens the menu, then selected SPWI59[0-7] is being cast. Placing SI into contingency would instantly open the menu, but after it triggers PC would still have to perform the casting after making their choice.
Hence the reason to read the SI's selection menu from Contingency's, so that it's not SPWI510 is getting stored - which is not too helpful as described above, but the real immunity.



I have just recalled the thing - is it feasible for dispel opcode (58) to make an individual check for each resource? Atm it either kills the entire batch of 10+ buffs, or leaves them all intact.

Edited by GeN1e, 14 January 2011 - 09:21 PM.

[Ascension64]

Spell Immunity is a double - first the zero-speed SPWI510 opens the menu, then selected SPWI59[0-7] is being cast. Placing SI into contingency would instantly open the menu, but after it triggers PC would still have to perform the casting after making their choice.
Hence the reason to read the SI's selection menu from Contingency's, so that it's not SPWI510 is getting stored - which is not too helpful as described above, but the real immunity.

I see, you need the casting to be speed 0 like contingency should be.

I have just recalled the thing - is it feasible for dispel opcode (58) to make an individual check for each resource? Atm it either kills the entire batch of 10+ buffs, or leaves them all intact.

For the same reason that individually checking saves for the effects in a single extended header is a bad idea, imagine the case where the Charm opcode gets dispelled but not the Display Portrait Icon.

[Ascension64]

"I'm female and created a simulacrum, but it has a male voice."

Strangely, I cannot reproduce this. Both the voice and the casting sounds are female for a female character.

[Jarno Mikkola]

There is also the Nahal's Reckless Dweomer

What's that about?

It's a spell alike the Spell Immunity, it lets you choose a spell to cast, any spell from your spell book AND it doesn't need to be memorized, with a timer of 5. The only downside is that you are almost guaranteed to be effected by a wild surge... and you have to be a Wildmage to be able to cast the spell, well, there's custom kits that allow that too, but a no normal mage can.

[DavidWallace]

* Flags for spells/item abilities to allow targeting of invisible characters

This should be possible.

Let me second the request for this.

[GeN1e]

I see, you need the casting to be speed 0 like contingency should be.

Even if the consequent casting was speed zero, it would still be a casting - potentially interrupting actions currently being performed, as well as resetting aura cleansing. I did give it some thought, but afaict there's no way to get rid of it via softcode methods.

For the same reason that individually checking saves for the effects in a single extended header is a bad idea, imagine the case where the Charm opcode gets dispelled but not the Display Portrait Icon.

Sorry, I've been not quite clear. I meant not each effect, but each group of effects sharing the same source (0x94). Similar to how 229 and 230 (remove one school/sectype) work.

[Ascension64]

Sorry, I've been not quite clear. I meant not each effect, but each group of effects sharing the same source (0x94). Similar to how 229 and 230 (remove one school/sectype) work.

Same issue. The effects of a single extended header will come from the same source.

[Igneous]

Even if the consequent casting was speed zero, it would still be a casting - potentially interrupting actions currently being performed, as well as resetting aura cleansing.

There's another serious issue with displaying the Spell Immunity selection menu at Contingency firing time. Even if such a thing were possible to implement without resetting the aura or interrupting actions, the player would be forced to have the affected character currently selected when the Contingency fires, otherwise access to the SI selection menu?and, thus, the entire spell?is effectively lost.

[GeN1e]

Same issue. The effects of a single extended header will come from the same source.

Mmm, but that's what I've been saying, no? The same source string means it came from, pardon the pun, the same source, or using in-game terms - same spell/item.

Say, I cast Armor, Blur and Haste. The CRE will now have a set of effects with the source SPWI102, another bunch with SPWI201, and finally SPWI305. Vanilla dispel will roll a check, and if successful, will dispel all three of them. What I'd like to see, is a dispel check performed individually for each viable group, three in this case, so that I can lose Blur and Haste effects, but Armor's.