Issue 3576

General discussion regarding the OpenMW project.
For technical support, please use the Support subforum.
Post Reply
User avatar
raevol
Posts: 2501
Joined: 07 Aug 2011, 01:12
Location: Caldera

Issue 3576

Post by raevol » 07 Dec 2016, 10:03

Just a note, as I am going through the changelog:

https://bugs.openmw.org/issues/3576

It looks like the fix for this deviates from Vanilla behavior? Was this discussed anywhere? From what I understand this fix should not have been accepted, as we are trying to replicate original engine behavior.

User avatar
Atahualpa
Posts: 791
Joined: 09 Feb 2016, 20:03

Re: Issue 3576

Post by Atahualpa » 07 Dec 2016, 10:18

It seems that Allofich had some sort of monologue on GitHub: Make AI response to spell hits more like original MW #1083. Although I agree with his decision, there could have been a more public discussion.

Here is the relevant part:
Allofich wrote:[...]
Well, some time has passed and I thought about this again, and as often happens I've changed my mind about the implementation.

I had decided to bring the behavior largely in line with that of original Morrowind, but I don't like that with this PR you can cast a hostile spell at someone and the guards don't react if it was absorbed/reflected. This is especially because, as far as I know, there are no 100% absorption or reflection abilities in vanilla Morrowind, so you are always threatening someone when you cast a hostile spell at them, and it should be considered an assault. This also would be more consistent with physical attacks, where even if you miss it is considered an assault.

Allofich
Posts: 104
Joined: 28 May 2016, 12:50

Re: Issue 3576

Post by Allofich » 07 Dec 2016, 12:41

Yeah, looking back, maybe I shouldn't have done that. The main goal of that change was not that part, it was counting each effect in a spell or enchantment as a separate "hit" for when followers decide to go hostile, and this is what the bug report title was referring to, but I added that in on top after I encountered that behavior. It seemed like a good idea at the time, and I know of some other places where we are deviating from vanilla, but I'm open to changing it to be the same as vanilla.

The problem is that when you get into the fine details of how these things work, it's pretty common to run across something where it's not clear whether what is happening is the intended behavior or a bug, or just something that wasn't specifically coded for because it was too much of an edge case.

Actually, I think it might be a good idea to have a discussion about how closely we follow vanilla behavior. I have a few particular bugtracker issues in mind, mainly AI combat related, where I think this may be holding things up. To just go ahead and say my view, I think we should basically just do what vanilla does on everything except what looks like buggy behavior (there could be differences of opinion here), and have alternate behaviors as configurable options, which we are already doing somewhat. (And yeah, I know this contradicts how I handled this issue, but as I said, I'm willing to change it)

User avatar
Atahualpa
Posts: 791
Joined: 09 Feb 2016, 20:03

Re: Issue 3576

Post by Atahualpa » 07 Dec 2016, 19:03

You could always create a thread about that topic (buggy vanilla behaviour in general) and name the issues you have in mind. This way, we could collect edge cases in vanilla behaviour and decide what we want to implement -- and how. I'd be really interested in your thoughts as a developer, who messes around with the actual source code.

User avatar
raevol
Posts: 2501
Joined: 07 Aug 2011, 01:12
Location: Caldera

Re: Issue 3576

Post by raevol » 08 Dec 2016, 11:13

Allofich wrote:Yeah, looking back, maybe I shouldn't have done that.
Hey I don't think you took offense, but just in case, I apologize if my first post came across a little confrontational. That's not what I meant by it at all, I was just writing quickly because I was in "summarize for the changelog" mode. Thank you for replying here, and I agree this is a good discussion for us to have!

Allofich
Posts: 104
Joined: 28 May 2016, 12:50

Re: Issue 3576

Post by Allofich » 08 Dec 2016, 15:48

I didn't take offense, but I thought you had a point about this being different from vanilla, and Atahualpa had one that there could have been discussion instead of me just making the decision myself.

I'm thinking now that this should probably be changed to vanilla behavior. The vanilla behavior isn't a bug, and it's conceivable someone made or will make a mod that expects things to work that way, and that alone is probably enough reason to do it the same way. In general I'm thinking now we should stick very close to vanilla, possibly with configuration options for alternate behaviors. It should help keep development work moving in the same direction, let us feel good about our changes and get bug tracker issues closed.
You could always create a thread about that topic (buggy vanilla behaviour in general) and name the issues you have in mind.
Yeah, I might do that.

User avatar
raevol
Posts: 2501
Joined: 07 Aug 2011, 01:12
Location: Caldera

Re: Issue 3576

Post by raevol » 08 Dec 2016, 23:27

Allofich, you're a hero. :)

Allofich
Posts: 104
Joined: 28 May 2016, 12:50

Re: Issue 3576

Post by Allofich » 10 Dec 2016, 08:36

I made the topic here viewtopic.php?f=6&t=3985.

Post Reply

Who is online

Users browsing this forum: No registered users and 3 guests