Jump to content

coconut

Founders [premium]
  • Content Count

    2639
  • Joined

  • Last visited

Community Reputation

1409 Excellent

About coconut

  • Rank
    Founder

Profile Information

  • Gender
    Not Telling
  • Location
    Sweden

Recent Profile Visitors

The recent visitors block is disabled and is not being shown to other users.

  1. Having the ability to use the complex flight model with the AI is nice, but being limited to that is not good. It's especially useless AND costly for bombers, who don't behave one bit realistically anyway (box flights with passive gunners, or disorganized wobbly-tumbly crowd with sniper gunners). Let the handful of AI fighters I'm dogfighting use the complex flight model, have the rest use a cheaper one. Regarding the topic: I think the engine is mostly fine. Eventually it's bound to fall behind engines like Unreal with more money behind, but there's always a risk that Unreal wouldn't meet one of the requirements for simulation. After all FS2020 went for their own engine too, meaning when they started none of the other general-purpose engine were affordable or good enough. The real problem is the limitations that the devs seem to be putting on themselves, like full-fledged flight and damage model for every plane, everywhere, all the time. It's a technical major tour de force to achieve that, but its value to me as an end-user is limited.
  2. You get this error whenever IL2.exe is passed arguments on its command line. In this case, I think it's steam that passes or_enable to enable VR. What you can try is set or_enable=1 in startup.cfg, and convince steam not to emit or_enable=1 (or =0) on the commandline. How to best do this, I'm not sure, but it should be possible by editing the properties of the game in steam.
  3. I second this. If there's something resaver.exe doesn't like in the mission or where it's located, it should say so in an error message.
  4. @GEIER_SnoopyBlack This is not the proper topic for your question, I think. Better to make a new topic. Anyway, you can load the msnbin file even if it's not visible in the file chooser. Just start typing the name in the input field, and auto-completion should show you a file ending in .msnbin. Choose that one, press the "accept" button (or whatever it's called, I forgot), and the mission should load. It's not mean to be edited by hand though, so there's a risk it will look rather confusing.
  5. Getting closer week after week. I don't have as much time to dedicate to this as I used to, which is why it's somewhat slow. But we're getting there.
  6. Flight Simulator isn’t initially by Microsoft. I played version 2 on my father’s Atari ST, and IIRC the producer at the time was called Sublogic. Gaming on IBM PC compatibles was already possible, but it was quite terrible compared to what Atari and Commodore could do. I think that might have been around 1988 or earlier.
  7. The notion of triggering and activating are separate. To use an analogy, if MCUs were guns, then triggering would be pulling the trigger, whereas activating is putting ammo in the gun. For the gun to fire, it needs to be triggered, and be loaded. Activating/deactivating merely puts in and removes ammo from the gun. Alternatively, you could say that activating/deactivating is the same as flipping the safety switch. Now, what I said applies to MCU triggers in the mission editor. Objects such as planes and vehicles are different. Activating or deactivating them will show or hide them. Actually, "spawn" and "despawn" might be better words, but there is also a concept of spawning in the mission editor, which differs slightly in a number of ways I won't go into here. So, to answer your question "Can an activate trigger be used to start a timer?", the simple answer is no. The complex answer I won't go into right now, as that might be too much info, but let's just say there are interesting things you can do with timers by enabling or disabling them after they have been triggered.
  8. airfield objects, I guess. arf_ is the prefix that static blocks at airfields use. As for the question of "increasing the people count", if that means "increasing number of slots in multiplayer", then the answer is no. This is a setting in the SDS config file. Max is 84.
  9. From HMD gdb: My understanding is that they measure the combined FOV of both eyes, whereas the game reports a single eye. The zoom implementation is wrong when PP is off. I think that when they narrow the view cone, they keep it centered on the normal vector to the screen's plane. They should instead center it on the forward vector. DCS used to do it wrong, similarly to IL2 now, but at some point they fixed it.
  10. It’s not the cores you have to look at, as windows moves threads from core to core. Better to look at individual threads in process explorer and see if any threads is at 100% / (# logical cores). But from my experience, 40-50% overall cpu utilization is all I’ll get on my 4-core i5 (idk if that’s what you got too).
  11. I've now updated my area-fitting app: https://1drv.ms/u/s!AtoivprUAORRxx6m_zTqtNZxQC4f?e=Id9EEE It works for the WWII maps, the Aras map wasn't digitized by Kappa. Requires to have .net core 3.1 installed, which I guess can be a problem. @PatrickAWlson let me know if you plan on using this, otherwise I probably won't put more effort into it. My own stuff doesn't use this EXE, it's linked directly with the lib instead.
  12. /Tech jargon warning/ What I've done is build a quad tree using Kappa's digitization, set it so that the max depth leads to a reasonable compromise between space and precision (max depth of 10 seems OK). Then build the complement of that, i.e. retain the empty nodes, throw away the points and save the binary serialization of that. I'd think Java has decent binary serializers (aka picklers). Loading the digitization, building the tree and saving the data can take some time, like 10min to 1hour, but once the result is saved, loading and querying is sub-second. The "complement quad tree" is basically a quad tree where each node is a square that contains no obstacles. Iterating over the squares, starting from the leaves and moving towards the root, I check if a) the square is in the region where I want to place a group of AA (for example), and b) if the bounding convex polygon of the group fits within the box. Build a list of up to e.g. 10 such squares, and pick the center of a random one as the position of the group. The links are to my code. If that's too arcane, you can also simply use a commandline tool I've made that wraps all that (after I update it to work with Kappa's latest digitizations). See this usage info.
  13. IL-2 is DX11. It was 32-bit DX9 when it was released, IIRC.
×
×
  • Create New...