I7: odd bug with Eric Eve's Conversation Framework v.9?

The maddening thing is that I’m up-to-date on the I7 built for Mac and I could have sworn I didn’t have this error with v.7 of Conversation Framework.

Basically, if I create an NPC

Jenna is a person.  Jenna is in the Command Lounge.

and comment out Conversation Framework, I get normal behavior:

However, if I comment the extension back in and recompile making no other changes, I get this behavior:

(Aresa is a crowd NPC using Mike Ciul’s extension and I’m also baffled as to why it’s redirecting to the crowd-named-Aresa when Jenna’s right there, but one thing at a time. I created Jenna as a regular NPC in case the Crowds and Conversation Framework extensions weren’t playing nicely together, but that doesn’t seem to be the major issue.) I haven’t put in additional code to handle what’s even said in the conversation because I’m concerned about the “That’s not a verb I recognize” message that’s coming along for the ride. It only appears when I enable Conversation Framework and I haven’t modded it–is there some obvious debugging thing that I am overlooking?

Advice much appreciated, and sorry for being so inept. :-/

Inform parses the command HELLO, JENNA as two distinct commands, i.e. as HELLO and JENNA, just as if you had typed HELLO THEN JENNA.
So first it reacts to the command HELLO by choosing a random interlocutor among the people present.
Then it tries to parse the command JENNA, which is a verb the game doesn’t know.
JENNA, HELLO, on the other hand, will do what you want.

There’s a way to fix this, IIRC, using “before parsing the player’s command” or something like that. Convert the command to indexed text, juggle it around, and hand it back to the parser.

ooh, I get so steamed about that comma. Thanks to Ron Newcomb’s Parser in I7, though, I might actually be able to do something about it one day…

By the way, I’m working on a conversation extension too. As of last week, I think it’s finally started to settle down, so hopefully there won’t be any other big changes to it. The documentation’s still a big jumble, though.


Okay, I can’t promise there won’t be any other big changes. I’ve been thinking about making the “speaking out loud” activity more like the “writing a paragraph about activity,” such that if none of your rules print anything, a “refusal to respond” will be triggered.

That’s true, but it’s straightforward only if you handle individual cases. The general solution is a bit more complex, mostly because disambiguation makes things considerably more messy. A bit of self-promotion: the Reversed Persuasion Correction extension turns all COMMAND, NPC commands around automatically.

Thanks, all–I’d completely forgotten about the comma behavior. I will give the Reversed Persuasion Correction extension a look!