Concordance of new features

[Avenger]

about opcode 132. I guess you meant Holy Might (0x84). It uses a state bit, just like many other effects.

It iwd1/2 they even have extended state flags (iwd1), and spell states (iwd2).

The spell states are the most closest to your spell ID idea, because there are some 96 bits (3 dwords).

In GemRB we support both the extended state and the spell states (128 bits altogether, though there are overlaps in hardcoded effects).

It is a pity you didn't design the engine, having an own field for this spell state value is definitely useful.

 

If Ascension will implement it, i will definitely follow suit

Do you think it is enough to maximize the value at 96? It would mean, i can use the existing spell state bitfield which i could set easily (storing more values in stats is not possible, so i would have to look up the 0x2c value in the effect queue), there is already an implemented trigger (spellstate) inherited from iwd2 and the spellstate methods are already written.

[Avenger]

I made a mistake in the previous post, we got 192 bits (6 dwords) for spell states.

I hope that's enough?

[Ascension64]

Question

// 0x93 Spell:Learn
int fx_learn_spell (Scriptable* /*Owner*/, Actor* target, Effect* fx)
{
if (0) printf( "fx_learn_spell (%2d): Resource:%s Mode: %d\n", fx->Opcode, fx->Resource, fx->Parameter2 );
//parameter1 is unused, gemrb lets you to make it not give XP
//probably we should also let this via a game flag if we want
//full compatibility with bg1
target->LearnSpell(fx->Resource, fx->Parameter2^LS_ADDXP);
return FX_NOT_APPLIED;
}

I couldn't find the definition of LS_ADDXP anywhere. Which bit flag is it? I will implement the same in TobEx.

I've been requested to add a few extra flags, of the following:

-enforce mage school restrictions

-exclude sorcerer class

They will use whatever lowest bit flags are available in param2.

[lynx]

it's 1. We have the following bits in use:

LS_ADDXP = 1 #give xp for learning it

LS_LEARN = 2 #give message when learned it

LS_STATS = 4 #check stats (alignment, etc)

LS_MEMO = 8 #memorize it instantly (add innate)

[Ascension64]

Hmm, the default behaviour of the opcode is to check stats (which is the die roll and the int modifier). It looks like GemRB implements the opposite; check stats only if the flag is set. Is this intended?

	if (flags & LS_STATS) {
	// chance to learn roll
	if (LuckyRoll(1,100,0) > core->GetIntelligenceBonus(0, GetStat(IE_INT))) {
		return LSR_FAILED;
	}
}

 

Also, does GemRB account for change in difficulty with spell school?

if (creTarget.m_BaseStats.kit[1] != KIT_TRUECLASS) {
if (g_pChitin->pGame->GetMageSchool(creTarget.m_BaseStats.kit[1]) == resSpell.GetSpellSchoolPrimary()) {
	nLearnSpellMod += 15;
} else {
	nLearnSpellMod -= 15;
}
}

[lynx]

It's opt-in, but the relevant callers should already use that flag.

 

Didn't know about the spell school bonus - cool, will add.

[Ascension64]

Umm, the only other issue is how is this compatible with PS:T and IWD2?

 

#147 (0x093) Spell: Learn Spell [147]

Parameter #1: Irrelevant

Parameter #2: Type

Description:

The targeted creature(s) learn the spell specified by the resource key, in the style specified by the 'Type' field.

 

Known values for 'Type' are:

0 Mage

1 Cleric

2 Innate

[lynx]

if their opcodes are really different, we can easily override them, which we already do for some others. I don't trust that iesdp description though.

[Avenger]

IESDP is correct about Original PST and BG1. Param2 is the spell type.

BG2 and IWD2 doesn't seem to care, they find out the spell type from the resref.

 

Anyway, this means, we cannot use Param2 (but we can use Param1) for the custom stuff.

[lynx]

added the specialist learning modifier and changed the effect to use parameter1. No spell(book) type forcing yet.

[Ascension64]

added the specialist learning modifier and changed the effect to use parameter1. No spell(book) type forcing yet.

An issue with this: in BG2:ToB, all the scrolls with Learn Spell have some value existing in param1 (which I believe is unused). Pretty much all of them are nParam1 = 1, but SCRL77.ITM has nParam1 = 5 (which is 0x4 & 0x1). Shall we try and avoid these bits?

[lynx]

I'm not sure if the original has param 3 and 4, but those are surely unused here, so gemrb could use them. I do wonder if param 1 is truly unused - then it doesn't matter anyway and the current state is fine.

[Avenger]

Param3/4 is currently not scriptable in GemRB's guiscript and v1 effects don't have these, so items/spells will be more complicated. I think using apply effect from scrolls isn't working either, because it won't be detected as a scroll anymore.

 

Currently, the best option i see is to use the bits of the high word of param1.

It is rather ugly, but all the other options are worse, as i see it.

[Ascension64]

It is rather ugly, but all the other options are worse, as i see it.

If you like it, I like it. Deal?

[Avenger]

It is rather ugly, but all the other options are worse, as i see it.

If you like it, I like it. Deal?

Ok