OpenMW 0.26.0

[Zini]

Not a huge amount has changed since 0.25.0. We (mostly) have Melee Combat now, but the rest from my original 0.25.0 posting is still valid. I'll quote myself:

* GetLOS
* God Mode

We need the former for a of things (stealth, combat AI and so on). And we need the later so we can implement player death without making testing a mess.

We are also accumulating more bugs again. Fixes definitely welcome.

I added the remaining Ai packages to the 0.26.0 roadmap. Shouldn't be too hard to implement.

With the packages almost completely implemented, it is also time we have a go at:

* Initialise MWMechanics::AiSequence from ESM::AIPackageList

And sorry, if I annoy people with that, but I would like to once again highlight:

* Implement NIF record NiFlipController


Editor progress is still slow unfortunately. With graffy temporary out of commission while he fixes his system, I am currently the only one that actively works on the editor code. All other people have either gone silent or have given their tasks back to me or graffy. That won't get us anywhere.


Anyway, we are still in the slow part of the year (damn you, summer!), so probably 4-6 weeks for 0.26.0, depending on how it goes.

[psi29a]

What are our current set of blockers for 'game completion' after finishing up combat? Is the main quest playable without hacks?

[Zini]

Haven't tested it myself, but according to this pull request it seems we are missing at least one more script function.

[TorbenC]

There is actually quite a large amount of research that needs to be done to fully understand how the AI packages are initialized automatically. It isn't as cut and dry as it may seem, especially with certain AI packages having what appears to be priority over other AI packages.

Certain AI packages seem to remove other AI packages from the queue when run, and others seem to assert themselves to the top of the queue if they are even in it. It's a touchy subject, I'm not sure of anything regarding the topic, these are purely speculations based on my observations when testing AI Wander in vanilla for research.

I'm sure I have all this stuff noted either in the aiwander.cpp/hpp, the thread I made for AIWander, or in a notes.txt file on my computer (which I haven't had access to for a little while and won't for around 15-20 days longer from the looks of it). I'll post it somewhere when I get a hold of it.

As for the rest of it, looking like this will be an awesome version, can't wait to be able to code again!

[Zini]

Really? I always assumed I works like this:

If there is only one package, use that.

If there is more than one package (or a package is added before another one has finished, with the exception of Wander), do something semi-random, which basically means you are screwed. If there is any regular pattern to it, then I haven't noticed it yet.

I know that the original CS allows plugging in multiple packages. But is this feature actually used anywhere and do we need to investigate the semi-randomness to not break content?

[TorbenC]

As I said, it is pure speculation, and your comment a while back:

As for AI-package eccentricities, I always worked with the assumption that if you do something with a NPC via script that involves AI/moving while an AI package is running, something will break.

In the AIWander thread, may very well be a good methodology, especially considering the unlikeliness that any mod actually relies on AITravel over-riding AIWander unless certain conditions are met (complete off the top of my head example, not really how it works).

But, it may be worth looking into. I know the interactions between AIWander and AITravel/AIFollow seemed very strange to me at the time. Damn I wish I had those notes with me >.>

But I do know there is indeed a pattern to it, that I am positive about. No matter how random it may look, the AI packages do follow a pattern of "priority" from the looks of it. Sadly I don't have Morrowind on this computer to test it with and gather some info, but I remember wondering about how I was going to implement it at the time.

[WeirdSexy]

It's looking like I'm going to be unavailable for video-making until the 20th of this month or so.

[Zini]

No problem. We won't have a release before early September anyway.

[cdoublejj]

I figured it would come down to research not long ago you guys were saying you needed more research. try not to strike me down but, proper OS 10 support would be nice some one was saying it would be hard to keep 32 bit support and it would be nice to have some more help on the OSX side of things.

which brings my next question, do you just real in all your help or do community members go and spread the words on other forums to try and recruit (or point them to new dev application section)?

just let me know if i'm way off topic and talking crazy voodoo nonsense.

[corristo]

If you could find someone who would be able to make universal 32/64-bit Qt build that would work on OS X 10.6-10.8 — it'd be awesome