Continue() : How is it processed in a BG1 AI script ?

[Gom Jabbar]

Sorry for such a basic question, but can you explain how Continue() is processed in a BG1 AI script ?

If the following script is applied to one of PCs in BG1 (standard install Version 1.1.4315 without any mod)

 

// Block1
IF
See([ANYONE])
THEN
RESPONSE #100
	DisplayString(Myself,1001)
	Continue()
END
// Block2
IF
True()
THEN
RESPONSE #100
	DisplayString(Myself,1002)
END
//Block3
IF
True()
THEN
RESPONSE #100
	DisplayString(Myself,1003)
END

I get String 1001 then String 1002 and finally String 1003 and that sequence loops as long as the script is active.

I thought the Block3 'd never been executed because the Block2 is always true. I expected Block2 to be parsed after executing Block1(because of continue()) and then the script restarts. But it seems that the Continue() action makes the script execute every next block with a valid trigger until the last one before restarting.

 

Can you confirm that Continue() works in that way for BG1 ? Do I get the same behavior if that AI script runs in BG2 or BG1:TotSC ?

[Echon]

I get the same result but as you said, I would never expect it to reach the third block. I do not know why it it behaves this way.

[Gom Jabbar]

Thanks for your prompt reply. Did you test it on BG1, BG1:TotSC or BG2 ?

 

Of course, it is a just an example to instance this issue and I get the same behaviour with more complex triggers and scripts.

[Avenger]

I'm shocked.

[Gom Jabbar]

I'm shocked.

All the more strange since the game uses Continue() in its own scripts for NPC !

 

I noticed this as I was trying and making a thief a little smarter for fun. I’m not very experienced in scripting but making an advanced AI for BG1 sounds like a challenge because of many restrictions (unmanageable ModaleStales, a lot of missing functions, matters to handle inventory, etc ).

If the Continue() bug is confirmed, it’s a severe limitation on the scripts for BG1

[Echon]

BG:TotSC, of course. If not, my finding would not really be useful.

[Avenger]

BG1 works as reported. (displays all 3 if the first block has continue, displays only the first, if the continue was removed).

BG2 works as expected. (displays only 1001/1002 with continue in first block)

 

Doh, this means, i have to add some external compatibility flag to simulate this difference.

[Guest Guest_Marupal_*]

Doh, this means, i have to add some external compatibility flag to simulate this difference.

Ouch! That seems like an awful lot of work just to duplicate a bug. It seems more than likely that this was something that slipped by the BG1 team because none of the scripts in BG1 fell prey to any ill effects from this flaw. I haven't seen any mention of problems from this with the BGT group. I'd be willing to go through the scripts in the original game files to see if it even makes a difference. I'd be willing to bet a few others would as well to save you the effort of simulating errors.

[Avenger]

Yeah, it would be good to know if any of the official BG1 scripts depend on this bug. Probably not, and i hope no one 'uses' it. On 'use' i mean even involuntary dependence. I know the Dragonlance TC was dependent on some quirks due to sloppy coding which was made obvious by gemrb

[erik]

Non-expanded IWD1 also works sanely / BG2-like.