Jump to content


  • Posts

  • Joined

  • Last visited

Posts posted by godarklight

  1. I knew the positioning system had problems, but now I know for sure how bad the problem is with numbers, and in which order to fix them!

    a) It seems like HackyInAtmoLoader is broken, it should be moving the vessel when it's physics are disabled

    b) Because HackyLoaded vessels are "landed" in the air, their unpack (physics enabled) distance is 200m.

    c) Angular velocity, I knew it was borked, I didn't realise it was this borked.

    d) Flappy wing syndrome. I suspect saving the part bending and re-applying it after the pos update will fix this.

  2. This is probably the most unadvertised part of DMP - it downloads all the flags from the server

    In order to upload one you dump it in dmp's flags folder and switch to it by clicking the flagpole in front of the astronaut complex (it's tiny but clickable).

    I originally made this after the "game crashes when the flag is missing bug" reared its head, thought it would be nice to see what flag they actually use instead of the default one.

    Those flags are definitely not in base DMP, but to be fair we've all definitely left a green goo on the launch pad, and I have it on good authority that Wernher von Braun *really* likes drinking fanta and may have snuck one of those flags into the kerbal space center :wink:

  3. @Tazius I have thought about that before, but I don't think it really makes all that much sense

    If I was doing anything with connections, it'd be CKAN integration. Connect to a server, when it detects ckan support have a box saying "do you want to sync mods to this server", and opening CKAN/closing KSP, seeing as mods are a major pain :P

  4. @TwinKerbal it looks like DMP is failing to delete something in the cache, which is quite odd.. You can try deleting the KSP/GameData/DarkMultiPlayer/Cache folder.

    That problem normally comes up though because DMP is currently broken on 32bit, which I've fixed for KSP 1.3.0 (dev version currently, I'll push it tomorrow after I play it some more and sort out anything I should really get out of the way)

    @Nailed it! DMP should work everywhere, I'm on linux and the fact it works on windows too is nice :P

    I'll hop on DMP on sarcasma's server tomorrow and get the feel for KSP and DMP again, I might be tempted to fix something.

  5. Woah, I hadn't been keeping track here.

    There's a lot of HAYO things about DMP, and I'm not sure if I ever should have even tried to support career mode as it's not something I play, and I'd much rather spend my time working on the general jankiness of DMP itself that effects everyone. I can't think of how to prevent the sfs cheat without disabling reverting / quickloading / quicksaving entirely...

    @JamesL86 in case you're still wondering, it's in the Settings folder. The file is only created when you set the server to custom.

    @ChainZ186 it works on 64bit, but I've addressed this in dev (supporting 1.2.2 still) and unstable (1.2.9). We initialised DMP differently to what's normal without realising it until now :P

    @megaspeeder two people in the same ship will put one of them into spectate mode, as for dual control, it's not something I've thought out all that much, and the only thing I can think of is allowing a pilot to give control to a spectator

    @VolcanoINC career mode was a mistake.


    On 09/05/2017 at 4:22 PM, Varsi said:

    KSP isn't a multiplayer game tho.

    DMP is a single player game where roughly the same thing happens :)

    @damian0816 DMP supports 1.2.2, so does "dev" currently, "unstable" currently supports 1.2.9.

    My development branch will always support the latest version of KSP, it's essentially "release with hotfixes" when I actually work on it. Unstable is for things I think aren't ready, or for prereleases.

  6. I finally have power back :D


    @Planetworks I think RockyTV plays with the career mode thing more than I do, but I'll take a poke at it, and add it to my "check DMP button" that's in the debug builds to see if I have to do anything during updates.

    @ajud On the modded client, go to options -> generate blacklist DMPModControl.txt and move it to the server. The server doesn't actually load mods (although it has a plugin system for the public server list and random things like that), it just tells the clients they are ok to use.

  7. Hrmm, seeing as how I've got a bit of time off I'll look into why DMP refuses to run on windows 32 bit, I believe rockytv's playing around with the new graphical system backfired - there's something different between 32bit and 64bit KSP.


    @moritz31 DMP is built on linux with mono (edit, run with mono DMPServer.exe), it's just nice that it works on windows too. Another recommendation I have is running it inside of 'screen' (or tmux if you're one of those heathens) so you don't have to have a console constantly open.


    Odd's are I'll also do some yak shaving and fix a few other things around the place as well :P

  8. @Not Sure Just heard of that in IRC, it was a bug that affected everyone. I've also done some other little fixes too!

    DMP v0.2.4.2 is now available for KSP 1.2.2!


    Server admins might want to notice this:

    Changed the "sendPlayerToLatestSubspace" server option and removed "keepTickingWhileOffline" (as it no longer makes sense).

    When set to "false", you will be sent back to the exact time you disconnected.

    When set to "true", you will be sent to the latest universe time (default, and old default behaviour).

    As "pausing" the latest time while the server is offline / no players are connected, and remaining locked to a subspace while offline made little sense - I've decided to remove the option which should have never existed in the first place, the option should be a lot more intuitive now, before it depended if a player was online, now it doesn't. It'll also keep you paused even if you turn off the server.

  9. Heh, ever since KSP 1.1 came out I forgot I could just run 1.0.5. But in any case this applies:

    5m offset:


    You can see some places where the terrain goes 5m higher than the PQS

    Side view:


    Top view with 0m offset:


    Red areas mean PQS is higher, grassy terrain is where obviously the terrain is higher :wink:
    And because someone probably tried to drag their mouse around to rotate this picture, have a video instead:






    Have fun @Lothsahn :3

  10. Hence why I tell RockyTV my hunches and he's been doing the git committing and replying here.

    I think my ultimate hope is that LMP kicks DMP's behind, as an active mod developed by someone who actually plays KSP has at least a chance of fixing problems that arise. I nearly didn't want to update to 1.2 and just do what KMP did, by locking and forwarding - but they've rewritten some parts that need a bit more work from what I've heard. Surely with time though they'll be better, if they aren't already :)

    EDIT: The procedural bit is completely deterministic, as in y=x^2, but if you draw a line from 0-1 you get a different line then if you draw a line from 0.5 to 1.5, despite coming from the same function. If the points could be binned somehow, I suspect that is what would fix it, but I know almost nothing about how KSP samples the ground and draws the mesh. The solution shouldn't require network transfers, just some type of way of making the ground completely deterministic, such as binning the origin of the ground mesh to the nearest 2km grid or something along those lines (although I suspect this isn't how KSP does the ground, so that might be impossible)

    I might actually make a quick little mod that someone else can video to demonstrate what the PQS actually looks like around the vessel, if I can get someone else to record it to show you what I mean.

  11. Thought I might check up on the forum thread... :P

    @Redneck from what I managed to gather when working on the raytrace hack, KSP appears to use some type of wave function for the ground, and polling the PQS actually returns the same values no matter the quality setting. I believe the problem is where the verticies are placed on the ground, because if they aren't in the same position, you get a problem that looks like this (imagine this as a 2D cross section):



    The black line is the wave function, the red line is an example of "client A", the blue line is an example of "client B". They both have the same number of sample points but the blue line's sample points are just slightly offset to the right compared to the red.

    If I could chose to magically make a problem disappear out of KSP, this is the one I'd pick first if they are serious about tackling multiplayer. In a non-multiplayer context though, the one I'd want fixed is the ability to start the game at all. Thanks Unity 5.

  12. I've done the bare minimum (merged whatever rocky did, and then fixed a few compile errors for the unstable branch to 1.2), but it has had 0 testing, and KSP is now insanely crashy on linux so I cannot test it at all, so I highly recommend using forks. This might kick your dog and kill your cat, but the build is on my build server in my sig (unstable branch)


    I've also been working quite a lot lately, and have moved onto another game so I've not had much time for DMP. When I do get some time I'll let anyone who has forked pick my brain for a bit, but I apologize in advance for the mess that is DMPClient :P

  13. Err.... I can't even? I can't even even. I'm all odd now.

    The method TimeWarp.SetRate is *definitely* there in both the linux and windows Assembly-CSharp.dll files. It compiles without error (and I've forced a rebuild of the development and unstable builds against 1.1.3, which appear to work on linux here). But that error pops up on windows somehow anyway?


    (More seriously, I'll see if I can convince RockyTV to build it for me and I'll provide rocky's build on the build server if it works).


    EDIT: I've had one person confirm that the latest dev build against 1.1.3 works, if a few other people tell me it works (one that had problems 1.1.3), I'll tag a release.

  • Create New...