Last call for SCS v32 fixes

[DrAzTiK]

19 hours ago, Ulb said:

- Possible use of SI:Abj, SI:Div in SR (Draztik, 1/23)

Maybe no need to check it out , Bartimeus told me that we need to tweak the settings.ini of SR (or SRR) :

ai_spell_immunity               = 1 // set to 0 if you want the AI to never use the old AI-only Spell Immunity spells (and have them replaced with spells of similar function or power)

 

Also not a bug but, de you think it could be possible to nerf the thief detect illusion ability.? It is quite overpowed considering it reveal all kind of invisibility and illusions. (and bypass  Spell immunity divination).  I guess it is surely harcoded hélas.

 

[AnonymousHero]

Oh, another one (I'm doing a playthrough right now if you couldn't already tell ):

 

The thralls in the Planar Prison behave extremely erratically, wandering around (which I think might be a good thing overall in terms of difficulty), but they'll happily wander into the petrification trap where Haer'dalis and friends are being held prisoner. This makes for quite a weird experience where you'll just see random "Thrall - Petrified" messages in the message log. They should probably just stay put until they at least spot the player's party members.

Edited January 27, 2019 by AnonymousHero

[subtledoctor]

6 hours ago, DrAzTiK said:

Maybe no need to check it out , Bartimeus told me that we need to tweak the settings.ini of SR (or SRR) :

ai_spell_immunity               = 1 // set to 0 if you want the AI to never use the old AI-only Spell Immunity spells 

I don't understand why someone would use such an option. That means players could use immunity spells that the AI can't. 

6 hours ago, DrAzTiK said:

Also not a bug but, de you think it could be possible to nerf the thief detect illusion ability.? It is quite overpowed considering it reveal all kind of invisibility and illusions. (and bypass  Spell immunity divination).  I guess it is surely harcoded hélas.

This would be a good request for Bubb over at the Beamdog forums. He's doing engine-modification stuff, he might be able to do something like, change thieves' DI to give them personal invisibility detection instead of dispelling invisibility. 

[DrAzTiK]

49 minutes ago, subtledoctor said:

I don't understand why someone would use such an option. That means players could use immunity spells that the AI can't. 

 

With SRR (and SR also imo) player never get access to spell immunity spells. I think this option should be set to 0 by default.

[subtledoctor]

1 hour ago, DrAzTiK said:

With SRR (and SR also imo) player never get access to spell immunity spells. I think this option should be set to 0 by default.

Players get spells that are effectively the sane as the SI spells. And IIRC the SI spells are changed to match the players' versions. So the immunity spells used by players, and immunity spells used by the AI, are the same. If you tell the AI not to use those spells, it will be unequal. 

[Relay]

Unless I misunderstood Bartimaeus, that option is there because v32 mages would still cast individual spell immunities with SR installed. That option makes it so enemy mages will never use vanilla immunities and instead only cast SRR versions of those spells.

[DavidW]

On ‎1‎/‎23‎/‎2019 at 12:14 PM, DrAzTiK said:

it seems enemies mage ... rarely cast mislead/ project image/ simulacrum

Just one note on this. I don't currently use project image/simulacrcum, as I think I might have mentioned. I do use mislead a bit... but you won't see it called 'mislead'...

[subtledoctor]

2 hours ago, Relay said:

Unless I misunderstood Bartimaeus, that option is there because v32 mages would still cast individual spell immunities with SR installed. That option makes it so enemy mages will never use vanilla immunities and instead only cast SRR versions of those spells.

But... I don't think that's true.  I mean, let's take the example of SI:Divination.  With SRv4, Nondetection fulfills this role.  So, what SR does/should/will do is:

• change Nondetection to provide full immunity to divination attacks
• edit SI:Div to match the text and effects of Nondetection

The player will cast the "real" Nondetection and the AI will cast SI:Div, which is now identical to Nondetection.  Meaning, both will end up using the same spell in the same way.  Parity!  If you stop the AI casting SI:Div, then it can't do that.  Now, maaaybe Bartimaeus is changing scripts to have the AI cast Nondetection whenever it would normally cast SI:Div, BUT 1) I don't think I've seen Bartimaeus do much work with scripts, so I'm not confident he's actually doing that; 2) even if he is doing that, it's a lot of work to reach a nearly identical result; 3) that only creates more chances for things to go wrong; and finally, since we're in the SCS foprums right now, 4) even if Bartimaeus changes vanilla game scripts in his SRR project, there's no way he can change SCS scripts, which if I'm not mistaken direct AI mages to use SI:Div to block divination attacks.

If your point is to ask DavidW to edit SCS scripts to use the SR(R) variant spells instead of the SI spells, well, IMHO points 2) and 3) above militate against doing so.  The way SR treats SI spells, IIRC, was specifically designed with SCS in mind, with the goal of approximating parity between the player and the AI. 

Now, I'm not sure how perfectly SR achieves that.  I'd like to improve it for the SRv4 final release if it needs improving, but I need information to do that.  In particular, I'd like to know which SI spells SCS mages actually use.  I vaguely recall DavidW saying it only uses ~4 of the subspells, but I can't find the statement when I search.  SR just needs to make the variants converge; if we know how SCS uses them, we'll know what needs to be done to ensure parity for the player.

[subtledoctor]

More relevant to the "last call for fixes:"

Tome & Blood has a component that converts the Sequencer and Contingency spells into innate abilities.  (It uses the stock white icons when the sequencer/trigger is "empty," and when the sequencer/trigger is "filled" and ready to use, it displays a blue icon that is I'm-sure-not-at-all-suspiciously-similar to the ones posted by @Bartimaeusearlier in this thread )

Obviously if that mod component and SCS innate sequencers are installed simultaneously, it will be problematic.  To avoid the conflict, you can detect the Tome & Blood component by checking whether "qdtnb_metamagic.qd" EXISTS_IN_GAME.  I'm working on expanding the TnB version to give several alternative options for how sequencers work (innate ability acquired at level x, innate learned from scrolls, and a spell version that will work with multiclass sorcerers), but I'll make sure to use the same marker file in the future.  And when SCS 32 is released I'll add a similar check to TnB on the chance that somebody does something wildly inappropriate like install SCS first. 

Edited January 28, 2019 by subtledoctor

[DrAzTiK]

2 hours ago, DavidW said:

Just one note on this. I don't currently use project image/simulacrcum, as I think I might have mentioned. I do use mislead a bit... but you won't see it called 'mislead'...

I have seen tolgerias using a strange mislead spell : the copy casted some spells but in the standard mislead, the copy is not supposed to cast spells.

 

some bugs with ascension :

1-In the first challenge of pocket plane Body does not react , she have no script.

2-Illasera in the tethyr forest : she use her teleport around script but does not go toward player 12 3 4 5 or 6. So she stay out of sight of the party and does not react until we come to her. (at least if all her allies are dead)

Edited January 28, 2019 by DrAzTiK

[Bartimaeus]

59 minutes ago, subtledoctor said:

More relevant to the "last call for fixes:"

Tome & Blood has a component that converts the Sequencer and Contingency spells into innate abilities.  (It uses the stock white icons when the sequencer/trigger is "empty," and when the sequencer/trigger is "filled" and ready to use, it displays a blue icon that is I'm-sure-not-at-all-suspiciously-similar to the ones posted by @Bartimaeusearlier in this thread )

Hah. I've never used Tome & Blood, so I had no idea (EE-only, isn't it?). Took like 60 seconds each in Photoshop to recolor them...and several more minutes to convert them back and fix their palette.

[AL|EN]

10 hours ago, subtledoctor said:

Tome & Blood has a component that converts the Sequencer and Contingency spells into innate abilities.  (It uses the stock white icons when the sequencer/trigger is "empty," and when the sequencer/trigger is "filled" and ready to use, it displays a blue icon that is I'm-sure-not-at-all-suspiciously-similar to the ones posted by @Bartimaeusearlier in this thread )

Obviously if that mod component and SCS innate sequencers are installed simultaneously, it will be problematic.  To avoid the conflict, you can detect the Tome & Blood component by checking whether "qdtnb_metamagic.qd" EXISTS_IN_GAME.  I'm working on expanding the TnB version to give several alternative options for how sequencers work (innate ability acquired at level x, innate learned from scrolls, and a spell version that will work with multiclass sorcerers), but I'll make sure to use the same marker file in the future.  And when SCS 32 is released I'll add a similar check to TnB on the chance that somebody does something wildly inappropriate like install SCS first. 

Then please use standard weidu methods for defining such conflicts like FORBID_COMPONENT, not "file detected" things which cannot be read and interpret via 3rd party tools.

Edited January 28, 2019 by AL|EN

[subtledoctor]

5 hours ago, AL|EN said:

Then please use standard weidu methods for defining such conflicts like FORBID_COMPONENT, not "file detected" things which cannot be read and interpret via 3rd party tools.

Are we going to have an argument about marker files all over again? They have been used since time immemorial and they work fine. As I said, the Tome & Blood component is going to soon grow to ~5 different components, all of which will conflict with the SCS component. SCS v33 might not be released for years, so it's not like the FORBID clause can just be updated later. 

SCS will generally be installed after TnB, so it's out of my hands to address the conflict - I did ask Wisp for something like FORBID_LATER_COMPONENT and I was shot down. If you can provide a constructive suggestion for how to actually deal with this, I'm all ears. But I think it's unreasonable to increase the risk that players will run into a conflict, for the sake of your unreleased alpha-stage app. 

EDIT - as I said, the forthcoming TnB components will use FORBID_COMPONENT for the errant instance of it being installed after SCS. So your app should be able to pick up on it, anyway. 

(EDIT 2 - man, the new forum text formatting is awful.)

EDIT 3 - @Bartimaeus ha ha, no worries. @DreamSlaveOne made the blue icons for me a couple years ago. That's one of my favorite mods that I've made... I'm not jealous of SCS doing the same (thus the in in my post). I think switching the trigger icons from white to blue is an effective and visually elegant adjustment. 

Edited January 28, 2019 by subtledoctor

[Bartimaeus]

Huh. I thought the colors would be switched from what you described. Blue would be empty (because it doesn't contain hostile effects in of itself), while white is full (because it could contain either hostile or friendly effects). I guess it's a little arbitrary either way you look at it, though.

Edited January 28, 2019 by Bartimaeus

[DavidW]

10 hours ago, DrAzTiK said:

I have seen tolgerias using a strange mislead spell : the copy casted some spells but in the standard mislead, the copy is not supposed to cast spells.

In vanilla BG2, AI-cast Mislead spells use illusory spells: they have visual effects only and don't actually affect anyone, but they can fool the inattentive player into thinking they're real. SCS just does the same thing.