faatal

My favorite feature so far is the fact I can switch the editor to release mode, which produces optimized assembly code, but I can still run the profiler on it. That means I get the same timings I would get with a full build, which takes my PC about 6 minutes to make, so I don't. I've been spending a lot of time optimizing and cleanup up lately, so a very nice feature.

I just finished this one, which is a big win:

Optimized terrain control textures to not update with every mesh update.

Those peaked at 2ms spikes on my PC, on several frames, when new chunks load, adding to

I am testing Unity 2020.2 currently. Only needed 3 code changes. Seems fine and is supposed to have some performance improvements. Feels a bit faster. 2020.3 (LTS) is supposed to be out soon, so A20 should end up being on that.

Exp will have another update, so I just merged it to 19.4.

The counter is probably correct, since the AwardKill func was calling AddScoreServer with 1 zombie for a vulture. It was the QuestEventManager that ignored vultures.

I found the code. Added as a zombie kill. Thx

That has not been worked on yet. The details of how it will work and what zombies has not been decided.

Crawlers currently path on 2m heights just like all the rest of the AI.

An airport does not have to be huge. It could be whatever size you wanted to make it. Like a small airport with one to two (parallel) 100m runways.

Steve. He says it is mostly done, but probably a few bugs lurking as is typical in new stuff, but I've not looked at it recently. He actually has a new task improving dismemberment features that one of our artists asked for.

Not yet, but we do have some stuff in the works.

I've been doing a lot of cleanup in the chunk system, so some moderate gains there and fixed some mesh/material leaks, but there is plenty more to do. Today I worked on improving the chunk mesh LOD update (really just hides meshes). Chunk decals/models now dither out, so I can reduce the ranges they get hidden at while having no popping.

Yesterday I added the xml for the burnt feral zombie, which has the flames. He is tougher and faster than the normal one.

How their weapons will work has not be determined.

I think many of the settings could use more choices, so I could see more being added, but possibly after a20.

I have several months of work left for A20 exp, so it is not soon.

I would guess yes. Not sure new game would matter. A bandit POI could just not spawn them if the option was turned off in an existing game.

We are pretty much just trying to get tasks done. Dev talk should pick up as we get closer to A20 exp.

Clouds/sky/weather are legacy systems. They work how previous devs made them and then we refactor them to better meet the final design of the game. The why of designs is often lost and in the end I don't care, we just need to make it work to meet the goals of shipping the game. I've been cleaning up and rewriting those systems as time permits, but it is a big WIP.

Some of what you describe are just side effects of a combination of things and it may change by the time we go gold.

Test features are usually for debugging and generally don't become normal features. I'm not sure w

Correct. That list of features is just a list of what we plan to do. It may be months before someone works on it and often features will get pushed back, because it can't be done in time or at all due to unforeseen circumstances. As the alpha progresses, the list tends to get updated to reflect where we are at, but is no guarantee of anything.

We have bug tickets for such things, so eventually they should get looked at.

Basically yes. There are and may still be some global parts, but it is all WIP.

Not planned.

No plans, but it could happen months from now.

Well that is your typical problem with having standard vertical character controller cylinders with something that is drawn mostly horizontally (like many animals). Basically it needs to be offset back, so the head is more towards the center, but is probably not a trivial change. It should get "better" at some point.

There is nothing natural about those vultures. They are Undead.

I have not, but it would make sense if heat was a worse effect, that you would want things to counter it.

It will be a big can of worms, but no work has been done on it yet, so can't say what w

You are blaming SI, when I don't think that is the main overhead. Profiling clearly shows chunk mesh and collider updates as the big overhead, which is primarily internal to Unity with data copying and PhysX updates. There are multiple chunk layers and their mesh and colliders update in the same frame, so hopefully I can slice that over multiple frames to reduces spikes. Have not tried changing it yet, since I tend to bounce between features, optimizations and bugs. There are some other mesh settings to try, plus some of our mesh building code could be faster. Our mesh building is on a separat

Before gold? I've got a never ending list of things I'm working on, so hard to say when it percolates to the top.

The main functional change is biome weather parameters are in groups now, so something like a storm can change all parameters to ranges a storm should have.

Biome temperatures are in absolutes now and global temp is just a random +/-5.

Grace period is same.

There probably won't be many more changes (for a20) beyond that.

Probably not. If you profile the game, it shows neither one of those things as issues. We plan on continuing to optimize the slower parts of the game, not parts that already run fast enough.