17:08 <GH> [Principia] DaMichel opened issue #1616: Crash on loading savegame https://git.io/vFL4Z

18:05 <GH> [Principia] eggrobin commented on issue #1616: > I removed Principia a couple of times during the game.... https://git.io/vFL06

19:06 <lamont> why would people think that they can add and removing principia from a savegame and everything would stay sane?

19:08 <egg> lamont: because they can do that with many other mods, and they haven't quite realised just how much principia has its own version of *everything*

19:08 <lamont> yeah, but just think about it for a second...

19:08 <lamont> its N-body physics that replaces how all the planets move

19:09 <egg> lamont: well unless you understand how we do the integration and what is stored by stock, you could imaging doing that by just perturbing the stock game's orbit and keeping those as the authoritative source

19:10 <egg> but instead we have giant protos that we dump into the savefile :D

19:21 <awang> Does Principia not like HyperEdit messing with things?

19:34 <egg> awang: well it can't distinguish it from ksp being inaccurate, so it will simply contemptuously ignore whatever hyperedit may do :D

19:44 <awang> Ah

19:44 <awang> That'd explain a lot

19:44 <awang> I assume that also applies to the "set orbit" part of the debug/cheat menu?

19:48 <egg> awang: yup

19:49 <awang> Hmmm

19:49 <awang> Does that also apply to the "set orbit" option of KRASH?

19:50 <egg> probably? I don't know KRASH, but anything that preserves the identity of the vessel, doesn't make it vanish from principia's realm for at least a frame, and moves it will be ignored

19:51 <awang> That could be problematic for testing

19:52 <awang> And once you start managing vessels in-atmosphere everything will be in principia's realm, right?

19:54 <egg> awang: not if you're landed :-p

19:54 <awang> Oh right

19:55 <awang> Any plans on changing that?

19:56 <egg> awang: so it's kind of fundamentally impossible to change that per se (KSP wrongness is indistinguishable from intentional modification), we could have some way of saying "declare this vessel unmanageable for testing" hidden in an obscure menu I guess

20:10 <awang> Er, I was intending to ask about changing managing landed vessels

20:10 <awang> But that answer works too

20:10 <egg> awang: not much we can do about landed vessels

20:11 <egg> they collide with the ground, we don't do collision physics

20:11 <egg> if we did there would be nothing left for the physics engine to do :-p

20:12 <awang> Hey, I'm sure replacing the physics engine entirely would make some people quite happy

20:16 <egg> awang: at this point could we just replace the game itself

20:17 <egg> awang: I cannot into graphics but surely we can make KSP a command-line game

20:17 <egg> > activate engine

20:17 <egg> *automated abort*

20:20 <awang> Oh god

20:21 <awang> That'd be hilarious

20:21 <awang> Also, ASCII art orbits

21:34 <lamont> your rocket has been eaten by a grue

21:46 <lamont> well i just wrote my own rkf45 ODE integrator for systems of linear equations and plugged in the equations of motion for a rocket in a central gravitational field around the earth, and integrated a hohmann transfer burn from a 185x185 to a 185x1000 in C on the command line.

21:46 <lamont> i did not get eaten by too many grues

21:46 <lamont> although i did have one sign error

21:50 <egg> lamont: you could use the principia libs! :D

21:51 <lamont> but i wanted to write it myself =)

21:52 <lamont> yours are all C++ software devved up...

21:53 <egg> :-p

21:53 <egg> but strongly-typed quantities and reference frames \o/

21:54 <lamont> but spraying pointers around with no idea how long of a vector they point at and ECI

22:07 <lamont> the rkf45 integrator ultimately winds up as C# code in MJ, and possibly plugged into PEG to entirely replace its predictor and all the approximations it does

22:08 <lamont> and then rub some calculus of variations on it and newtons method and it turns into a pretty robust trajectory optimizer