Jump to content

Lisias

Members
  • Posts

    7,399
  • Joined

  • Last visited

Everything posted by Lisias

  1. I don't think you will find a bugged Add'On, it's more likely you will find a poor victim walking on the mine field and getting the blame but this is what I do to find the source or the trigger of the problem: * Delete half of the Add'Ons (make sure to do not delete any dependencies from the other half) * Try to reproduce the problem * Put back that half, delete the other half * Try to reproduce the problem. On the half that triggered the problem, split it in half, goto step 1. Do it until you find the thingy that triggers the problem. (We call this a Binary Tree Search). Then post it here, with some luck someone else can recognise the pattern and hint us with some more info.
  2. I used this last year on the previous Fastest Juno, and I can tell you for sure it's highly improbable I had invented it - besides not remembering if I had saw someone else using it before or after I created this stunt. But, yet, swjr-swis had already did it before me on that challenge. But after this one. (and yeah, I learnt a lot on that challenge, it's the reason I have some remembrance of the entries! )
  3. Well, there're kids and perhaps nuns around here. Can't blame Forum about this, otherwise they would need to classify this thing as PG-13 or harsher. (Kraken knows they would need to classify this as XXX if they allow me to unleash my verbosity unchecked - my childhood happened on a very... interestingly impudent times... as it appears. Heavy Metal, MAD and Groo was commonly found on kid's hands on where I had grown!
  4. Well, the technical details (for reference) is that you got an Access Violation after some entering and exiting Scenes (you switched Flight and Trackstation a lot of times, what hints me you were in the middle of some long range mission! Sorry for that....) UnityPlayer.dll caused an Access Violation (0xc0000005) in module UnityPlayer.dll at 0033:1f6921fd. The novelty is that the crash happened on a completely new way (that I had knowledge of): Stack Trace of Crashed Thread 6544: 0x00007FFB1F6921FD (UnityPlayer) UnityMain 0x00007FFB1F68FED5 (UnityPlayer) UnityMain 0x00007FFB1F8B1279 (UnityPlayer) PAL_Thread_YieldExecution 0x00000212D7CE294A (UnityEngine.CoreModule) UnityEngine.Mesh.Internal_Create() 0x00000212D7CE28BB (UnityEngine.CoreModule) UnityEngine.Mesh..ctor() 0x00000212D7CE2323 (Assembly-CSharp) TMPro.TextMeshProUGUI.Awake() <--- Here! 0x00000212BEA0DB70 (mscorlib) System.Object.runtime_invoke_void__this__() 0x00007FFB1E94BC80 (mono-2.0-bdwgc) mono_get_runtime_build_info 0x00007FFB1E8D1D92 (mono-2.0-bdwgc) mono_perfcounters_init 0x00007FFB1E8DAD8F (mono-2.0-bdwgc) mono_runtime_invoke 0x00007FFB1F81B19D (UnityPlayer) PAL_Thread_YieldExecution <.....> TextMeshPro is a tool KSP uses for rendering text on the screen, and it's widely used by Unity projects (and not only KSP). While this doesn't rules out a bug, it makes it somewhat improbable. And stressing again, you entered and exited TrackStation a lot of times, this didn't happened the first scene changing - it's not a trivial or obvious bug. Access Violations can also happens when you have bad memory, or something is corrupting the swapfile on the disk. So, before applying for a bug report on Squad's bug tracker (only them can really understand properly all this stackdump, as only them have access to the full source code!), do as follows: Check your Hard Disk/SDD health There's a thing called SMART, that tells you how the storage device is behaving. Look for ECC/checksum errors Check the remaining lifespan (specially important for SDDs). If you find any errors, replace the storage as fast as your wallet allows, your storage is near end of life. After being assured the storage is healthy, run a chkdsk Check your main memory Use memtest or similar. Check your video card (just in case, these things has memory too) Try OCCT. (I don't know how to use it, I use a Mac) My Mac once crashed on an Access Violation too, I'm using a cheap SDD as boot device and I stupidly used it for the swapfile. After changing the swapfile to the HDD, the machine (obviously) became a little slower, but I had no more crashes by Access Violations since them. It's a long shot, but it may help you on diagnosing your issue - use an HDD for the swapfile if you can and see what happens. If everything above checks OK, you will need help from Squad. Had this ever happened on a unmodded KSP instalment?
  5. TweakScale mangles the mass and the attach points... Yeah, something stomping TS toes, an unhandled exception, and anzero mass part someone nullifying any force applied to it. The gantry stretches itself to a point, and by releasing the power, it comes back to the original attitude abruptly, exactly as it would be "welded" to a part by an autostrut. However.... I'm sorry to say, but... "It works for me" ™ I tested it on a near clean test bed, so all I had ruled out is a bug on vanilla IR. It's a interaction between IR and something else. I will inspect the craft file, install the missing Add'Ons based on the missing Modules, and try again. In the mean time, can you publish a "full" KSP.log from your instalment? Delete the MM caches and fire it up, load the craft, execute the action that leads to the problem, quit KSP, and then get the KSP.log? It's plain possible that I could diagnose the problem inspecting the craft file, but if I don't, we would save some time by having your KSP.log by the time I come here to report my findings. --- POST EDIT -- Shoving up the max value for Force, Acceleration and MaxSpeed makes the testing surprisingly entertaining. -- POST POST EDIT -- @[email protected], I checked the "source" for your craft, and installed the Add'Ons that would satisfy that modules: MechJebCore (MechJeb2) USI_ModuleRecycleablePart (UmbraSpaceIndustries) RCSLandingAidModule (RCSLandingAid) FlightEngineerModule (FlightEnginerRedux) NONE of them presented any glitch, everything worked as intended. So I thought if could be KerbalJointReinforcement Continued, at least the last time I checked, it is not compatible with Infernal Robotics. However, my guess was wrong, things were working as expected even with KJR installed. The only way I could kinda reproduce your issue was using AutoStrut to Root. AutoStruting to GrandeParent and to Heaviest just locked the gantry, but using To Root rendered it behaving more or less as your example. So, if you are absolutely sure you are not using AutoStrut on that thing (and the craft file had the autostrut deactivated), I'm pretty sure is some Add'On doing something like the autostrut (perhaps an old version of KJR/N? I know of one release that borked, but it was quickly fixed). In order to keeping pursuing this bug, I will need that KSP.log I asked above. Follows a video with the test run (video is terrible, Apple cripped QuickTime on Mojave! My older MacMini, still on Sierra, recorded videos better than this piece of <insert your favorite non forum compliant adjective here>.
  6. At some point, we need to extend this consideration to the modders that made the mods at first place. Their dedication is equally commendable.
  7. I hate it in a way that forum rules prevent me to express myself. But I still love this freaking game - so I'm still willing to withhold some level of abuse in order to help keeping this thing healthy. But, granted, I don't need to be here on Forum in order to to that. It's only, yet, the best way to share my work, but I don't need this to carry it on.
  8. @steve_v has an excellent point here. And he is not the first one to talk about it. And, again, it's not a surprise software developers are choosing to gather on private Discord servers or anywhere else but here to get the problems solved. (emphasis is mine) Quod Erat Demonstrandum. Collaboration is happening on private messages and personal emails nowadays. I took some time, but now I get it. On FOSS, the binary is irrelevant, it's just a bunch of bytecodes you are forced to generate using a tool (compiler) in order to the get the code (the only thing that really matters) running on the metal. However, on non FOSS licenses [and I think it's the reason that on the beginning (at least as I was told), it was demanded that code should be published here using a FOSS approved license] binaries are copyrighted material. Everything that is not source code are also copyrighted, and this includes the DLL. So recompiling the DLL is a derivative work (from the legal point of view!), something yet more relevant on CC licenses that blatantly doesn't demands the sharing of the Source code - and even Creative Commons advises against such a practice, by the way. The guy that waved the rule to license the source code on a FOSS approved license to something be published here made a really, really bad move.
  9. This appears to be a somewhat drastic reaction in order to cope with some asinine copyright laws from some countries. Perhaps an overreaction, but it's not too far from what I would do myself. What does not means that this cannot be used, also, on what you are complaining. The weapon that protects is the same that kills, it's the hand that hold it that defines the role. Follow the money. it's always about the money.
  10. Looks like an autostrut trolling you. Can you publish the craft file? It would make easier to check it!
  11. I agree. But with the robotics, this part lost priority on the queue. You can easily reproduce the desired feature with Breaking Ground as well with IR/Next. I'm way more interested on resurrecting that gorgeous landing gears.
  12. Nope, it was Work in Progress when I take this Add'On, and I didn't took it down to prevent breakage on some artistic work that could be using it. It needs some work before being really useable on a Career or something.
  13. Moved from that thread again. I'll not create a new thread just for this post, and since the guy was asking about what happened, moving the conversation to where it happened made sense to me. We started to discuss something completely unrelated to the thread, besides being triggered by the something on the thread. So instead of keeping derailing it until someone report the posts and so the moderatos have to move it themselves, I choose to cut some corners and do it myself. At least I realised the problem after my first post on the mess.
  14. Ideally, one by one. But, boy, this is boring... one way to speed up things is kinda do a binary tree approach: remove half the add-ons (but not the dependencies from then the other half). see what happens. put them back, remove the other half. See what happens. Take that half that gave you the behaviour under research. Split it in half, go to step 1. This is pretty good to speed up looking for a single trigger for the problem, but it's not bullet proof. Now and then you end up having to do the hard way on a second attempt. On some weird bugs as this one, flipping a coin can be the wisest choice sometimes...
  15. IMHO, this is another issue. And yeah, I agree with you on this. People are using private Discord servers for a reason.
  16. It's the other way around. The smaller the community, better the troll~s chances to get something. Bigger FOSS projects have too much people involved, trolls are overwhelmed by the backslash. Small projects (FOSS or not) don't have so much people caring for it, so chances are that the troll will push his stunt unchecked. On proprietary software it's yet more probable, as usually the guy that is paying for the party wants things settled up fast and quietly - and this gives an edge to copyright trollers. Nothing. But so, way less people will clone the stuff. And, so, way less people would testify for you. But some of them do. If you are willing to take your chances, it's up to you. Please don't criticize people that choose not to do so. Chain of Responsibility. There's no one overlooking your servers but yourself. Weak defense on a litigation. It's not about you. Of course you trust yourself. The key point is: why anyone else would do it? Believe it or not, github has way more to lose than you by failing to comply with a copyright law. This make them more trustworthy on some situations than you. Three different forks, each one stating the same license file, and I'm covered. A single link to a fork of yours, on your server, not that much. What's reasonable to an American may put an Australian in jail. It's not about you, it's about everybody else - there're no Good Faith (or even Fair Use) about Copyrights on some countries, and yet their citizens are users of this Forum. We fail to do things right for them, we (theoretically) jeopardize their lives. I agree there're not much of chance for this to happen - but not much is not none, and the bad press could be bad enough to justify simply ruling out the possibility by "playing safe" on the matter. What we are telling you is that there's a need for the stunt - and we are explaining the reasons, and the failing to understand it doesn't nullifies such a need. Keep in mind that we are not asking you to like this stunt (I hate it, by the way), nor criticizing you by not being willing to contribute under such conditions.
  17. Yep. But until there, you need a workaround to at least have something to play until the fix is not applied. One thing that may help is to delete a memory intensive Add'On to see what happens. If nothing changes, put them back and delete something else. Rinse, repeat. And yeah, it sucks - I know it. I have some KSP 1.8.x installments at hand for doing tests, and besides had shoving a huge amount of Add'Ons (some of them not "updated" yet), I never got this one. But I'm using a Mac, this could be something specific to Windows….
  18. Curiously, I know of some people that do the same to Software Developers! BY THE KRAKENS! I'M A KERBAL!!!!!!
  19. That random github repos are, statistically, way more cloned by individuals than private cloud instances. And so, they have more people in a position to testify for them. You see, there's no digital proof for anything. All we have are evidences and testimonials. Me too. Being the reason I trust people that clone public repos from a large corporation, and not the large corporation that stores a private cloud instance! You have a point, but not exactly by this. FOSS has its share of asinine legalese meant to protect themselves too. Blame the game, not the gamers.
  20. And then the linked repo changes name, or is deleted. And a copyright troll claims you are stealing this code. Good luck. The trust is on how many eyes are looking into the repository and cloning it. One can rebase the repository, "falsifying" timestamps on the commits (it's really easy on GIT, and one of the reasons I prefer Mercurial on professional works - you just can't do it on Mercurial, the older branch is archived and will be always available for inspection on the server!). The only defense you have against this is people cloning the repo, and you have way better chances to get the thing forked on a public repository. Large organisations, usually, have a greater lifespan than personal hosting. So you are better served by cloning from there. You must keep in mind: if someone challenge you on a copyright strike, it's yours the burden to prove you are right. It's way easier this way than your way. You are getting it wrong. You are not protecting the downloader by doing it. You are protecting yourself. It's your the SAS on the line, sir, not the downloader - he's getting what he had paid for, it's free as in beer, no guarantees. All this stuff is needed by when people strikes you with copyright claims. So Forum can't allow such a link to be published, as they would be liable in the case the download it's illegal. Are you getting the point now? So do as PGP says, send the DLL in person instead of using Forum! Forum is a man in the middle. It's like a Key Repository for PGP, everybody must trust each other, or the thing just doesn't works. Forum needs to trust you that the DLL is license compliant, and the way to do that is forcing you to publish the code you used so anyone can inspect it. Make no mistake, Forum is covering their SAS on it - not yours or mine. They just can't allow copyright infringement material being linked here - the costs of defending themselves from such a stunt would easily prompt TTI to shutdown this thing as the cheaper way out of the mess. Agreed. But it's what they can do to protect themselves from Ridiculous and Insane Legislation™ the best (or less worst) they can. Not exactly. One thing that every license explicitly forbids is impersonating. By recompiling the DLL and posting only the link to the DLL recompiled, there's not evidence that you had recompiled the thing other than the post on the forum. Had you changed the copyright claims on the AssemblyInfo.cs? No? So how one will tell your recompiled from the original author's one? If there's no way to tell the difference, you can be accused of impersonating the original author. Do you want to take this risk? Forum doesn't. If you ask me about my personals feelings about the need of doing that, I would promptly stay silent in order to avoid infringing the Rules 2.2 itens A, C, D, E and above all, G. But yet, I can't blame them for doing that (besides cursing silently... hehehe). Nope. You need to state what was changed ("I targetted 4.5, and linked against 1.8 DLLs - no source code was changed") too. And, ideally, changing the AssemblyInfo.cs so the DLL itself is marked somehow as being from yours, and not from the original author. And so, Forum needs to trust you in order to accept your link in the hope there's nothing malicious on your file. You see, Forum needs to have in you the same trust you need to have on Microsoft, Dropbox and Google.
  21. From the source: You can download and rebuild the thing. YOU can't prove nothing by itself, you can provide evidences and allow someone else to check it. Even PGP works this way. All you can prove is that you used one specific private key to sign the code, someone else that both sides trust must tell that you are the owner of the respective public key. FOSS works on a chain of trust: you fork the repo, you compile the thing and you publish the DLL in a place both sides trust. Then anyone can double check your claims. But they provide a like link to download their patches and another link telling you from where they are downloading the third-parties material, right? Under FOSS rules, you are usually obliged to provide all the source you used to compile thing, and/or inform from where the user can do it. But if you are using GPL, you have yet another problem: you still need to provide the source in the case the link you used to provide it goes down - and it's simple like that. You provide a DLL and the source you linked goes down, you are in copyright infringement (I think I had read something about this requirement being waved after 5 years or something like that). On the specific case of the Contract Configuration, it's on the CC BY-NC-SA 4.0 and one of the clausules says: So you are still obliged to acknowledge the source - but granted, you do not need to provide the source yourself. But you are still obliged to tell you had changed the thing (or not). The easiest way out is to fork the damn thing and let GitHub do the hard lift for you. And forum rules says that you are obliged to provide the source nevertheless to have it linked here. Nope. Forum is private land, they rule what can or cannot be published here. Forum is not restricting the right to compile and publish the material. Forum is restricting the material from being linked here. The dude are still entitled to publish the material anywhere else. And keep in mind that besides CC doesn't provide a term demanding the publishing of the Source Code, some Software Licenses does that - and by failing to provide it the guy would be in license infringement (that so implies in copyright infringement). And since Law says that linking to copyright infringement material is an infringement itself, the safer way out of the mess is to demand publishing the source in a way that would prevent liability to Forum. Some time ago people used to publish the source code on the same package that the binary, instead of pinpoint a public repository. That would do too, as it will both satisfies the worst case licensing scenario as well the Forum Rules. But, yet, forking the thing and linking to it is way more practical - and you earn a place to publish the binary too instead having to use your private quotas from DropBox or Goggle Drive. But I digressed.
  22. Well... I inadvertently derailed a thread, so I choose to move my post to a new thread here and clean up the OP. Source. -- -- -- -- -- The whole Copyright Law is ridiculous. But yet, it's the Law. Our current Copyright Legislators are insane. We are being ruled by STEM illiterates that are abusing their position for funds from some big players of the industry. It's the exact same code? You need to prove it - otherwise you risk a copyright strike. The law is cristal clear, it's up to the content distributor to prove he/she is following the license terms, and the only way to prove that is... Forking the repo and publishing the compiled DLL from there (or from a place where the you can link to the repository, so anyone can verify your copyright claims). It's not their rules. It's the Copyright Law. It was ruled that LINKING to a copyright infringement material is a infringement itself. Forum need to cover their SAS. Uh... But it essentially what's happen? Gentoo, Debian, Fedora, OpenSUSE, all of them has their own repositories publishes from where they compile their code. -- -- -- On a side note, perhaps you are being sarcastic and I missed the point?
×
×
  • Create New...