steve_v

You do indeed, but up until fairly recently you needed elevated privileges to make them so they got a lot less use than they do on *nix. Personally, I'd be stoked if an update made that a feature (with a note in the changelog of course). Symlinks are awesome, and they really shouldn't cause problems for a well-written application. Dunno about Windows, but over here symlinks are pretty well transparent to applications as they're resolved at the native syscall level. If you want to know the real path you use readlink() or realpath(). Subnautica is Unity as well AFAIK, so without digging into the API docs I'd guess either Unity or mono is returning the real path by default, and some code in Subnautica didn't make allowances for that.

Tried, tested and very much working as it should.

6.6.0.161 Ed. this problem appeared with CKAN 1.26.8. Previous binaries run fine here. Same results with mono 5.20.1.19. In before you people tell me to update my mono install, 6.6.0.161 is the latest mono release available from the Gentoo repos and all later releases FTBFS for me, probably because mono's build system is insane.

I thoroughly doubt it has anything to do with the amount of system memory available. There's no mention of an allocation failure in the logs, and the system had ~7GB free when the crash occurred. This: D3D11: Failed to create RenderTexture (2560 x 1440 fmt 27 aa 1), error 0x887a0005 Looks very suspicious, and smells like a GPU or GPU driver problem to me. But then I don't know much about Direct3D or it's error codes, and I don't use Windows. Punching that error code into your search engine of choice may yield some things worth trying. Ed. There're a post or two on the Unity forums claiming that error 0x887a0005 is a graphics card driver crash. I have no idea how reliable such comments might be though.

Excellent... Wait, It's not "Hit Steve with a moist haddock for doing silly things with links" by chance, is it?

Alright. The time has come, with some trepidation, to crosspost this one here: FAR alone seems fine and there is little help to be had over there anyway, at least since ferram4 went AWOL. So I now I must humbly poke @sarbian with a pointy stick and ask why adding MechJeb to a happy little FAR install is bringing me such hurt. Any ideas? It looks like an unfortunate interaction to me, rather than solely a FAR bug.

Yurp. I've never used it so I'm not sure how it's supposed to work (I hear Feline Utility rovers can use it, so loaded up some of those parts as a test), but it seems to me like it just flat doesn't, whether GameData is a symlink or not. No shiny shiny, textures look identical to me. What it does do is spam: "OpenGL Error: Invalid texture unit!" in the log incessantly, but it throws no I/O related exceptions. A quick search throws that GL error up as a "Known issue" for some releases of Unity 2019 on GNU/Linux, so I'm guessing it's just the usual Unity borkage on this platform.

I don't see why not. BackgroundResources is a continuation of this mod, and it looks to be as standalone as it ever was, only with more features (mainly to support TACLS and DeepFreeze). Have a look around in the source for an idea of which other mods and stock systems it supports.

Uhh: steve@perdition ~/Games/KSP_Linux_1.8.1 $ mono ./ckan.exe System.IndexOutOfRangeException: Index was outside the bounds of the array. at System.Windows.Forms.ListView.GetItemLocation (System.Int32 index) [0x00018] in <91b18fcffc5e47d3a5d52c9f469a1a52>:0 at System.Windows.Forms.ListView.get_LastVisibleIndex () [0x00043] in <91b18fcffc5e47d3a5d52c9f469a1a52>:0 at (wrapper remoting-invoke-with-check) System.Windows.Forms.ListView.get_LastVisibleIndex() at System.Windows.Forms.ThemeWin32Classic.DrawListViewItems (System.Drawing.Graphics dc, System.Drawing.Rectangle clip, System.Windows.Forms.ListView control) [0x00011] in <91b18fcffc5e47d3a5d52c9f469a1a52>:0 at System.Windows.Forms.ListView+ItemControl.OnPaintInternal (System.Windows.Forms.PaintEventArgs pe) [0x00011] in <91b18fcffc5e47d3a5d52c9f469a1a52>:0 at System.Windows.Forms.Control.WmPaint (System.Windows.Forms.Message& m) [0x0006d] in <91b18fcffc5e47d3a5d52c9f469a1a52>:0 at System.Windows.Forms.Control.WndProc (System.Windows.Forms.Message& m) [0x001a4] in <91b18fcffc5e47d3a5d52c9f469a1a52>:0 at System.Windows.Forms.ListView+ItemControl.WndProc (System.Windows.Forms.Message& m) [0x00071] in <91b18fcffc5e47d3a5d52c9f469a1a52>:0 at System.Windows.Forms.Control+ControlWindowTarget.OnMessage (System.Windows.Forms.Message& m) [0x00000] in <91b18fcffc5e47d3a5d52c9f469a1a52>:0 at System.Windows.Forms.Control+ControlNativeWindow.WndProc (System.Windows.Forms.Message& m) [0x0000b] in <91b18fcffc5e47d3a5d52c9f469a1a52>:0 at System.Windows.Forms.NativeWindow.WndProc (System.IntPtr hWnd, System.Windows.Forms.Msg msg, System.IntPtr wParam, System.IntPtr lParam) [0x00085] in <91b18fcffc5e47d3a5d52c9f469a1a52>:0 Thoughts?

My everyday language should probably be AO at the very least. There are still a good many word the filters don't like that I'd put squarely in the All Ages category though, which is my point... Relatively inoffensive synonyms for buttocks which are used as common parlance in many parts of the English speaking world, for example. It's another facet of the conservative to the point of paranoia cheeks-covering that goes on here, and that's why I mention it. Would that last sentence really have been any more or less suitable for children if it contained a word rhyming with "grass" rather than the inane substitution? Those were part of my childhood also. Amazingly (with liberal helping of sarcasm), they neither turned me into a psychopath nor got anyone sued.

Thanks dude. To be fair, that kind of assumption is probably perfectly safe on Windows, It's just us *nixers who like to throw symlinks around like confetti. I really can't blame you for not anticipating it. I suspect it's actually easy enough to fix for even a complete .NET noob like myself to sort out if he could find the motivation, but it's your baby so it's probably better if you tickle it.

That is another thing I don't like about this forum TBH. IRL I swear like a pirate. I tone it down to levels I would normally reserve for nuns and small children, yet the filters still mange my meaning regularly. But I digress. My thoughts on this FUBAR licencing and publication situation and my thoughts on the rules imposed on my printing of those thoughts are both unprintable... Spontaneous combustion of publication medium levels of unprintable. I guess I'll withhold them too, lest the database server takes up smoking.

I have, repeatedly. The problem is not that it's missing some files or in the wrong directory, it's that it can't find its own stuff if the GameData directory is redirected with a symbolic link. All the other mods in there are perfectly happy with this situation. The game itself found and loaded scatterer through that same link, and from my admittedly very casual look at the source it appears scatterer is just using the standard methods shipped with the game for file I/O. Odd. I'm missing something obvious, but a reinstall it is not.

Indeed. The source code is the IP, a binary is just a side-effect, something produced by mechanical means with no creative input. Since source is distributed freely and anyone can create a binary from it whenever they please, nobody cares. It would be like copyrighting the ink rather than the story. Everywhere else I collaborate everything being FOSS is a given, so arguments like this never arise. And so... I took some time, but now I get it. I still don't like it though, that code doesn't run at all on my brain. Frankly, the more times I read it, the more I think it's just a terrible choice of licence for software. Aside from being incredibly vague in general, it talks only about "the work", which in FOSS copyleft terms is understood to be the source code, but in non-FOSS could mean anything at all - and there's not a less-vague clarification of this point anywhere I can find. Or Mailing lists and IRC, as it has since the dawn of time. Absolutely agreed. I thought it was a silly thing to do at the time, but I had no idea at all as to the the depth of that error or it's ramifications.

I have no difficulty whatsoever understanding the reasons as you state them. What I do have difficulty with is agreeing with them... That and refraining from replying to such a condescending tone in a snarky and/or combative one. You're right, I don't author mods here, I won't release any of the little bits and pieces I wrote for my own use (though I was contemplating getting around to that before this discussion happened), and I'd never want to moderate because inflicting rules and regulations on others is anathema to me. If you're having difficulty understanding the reasons and would like to learn more, you're welcome to read this thread from the beginning. Despite people moving posts around, it's probably still in something close to chronological order.

It's much the same in every scene where kerbin is visible from space. But scratch all that, I have just discovered that the two problems are not related at all. Actually scatterer related problem report follows: It would appear that scatterer is failing to resolve symbolic links properly when loading some of it's files. Why I don't know, Neither the base game nor any other mod I have ever used has this problem. No revealing exception is generated until I load into the game proper, which I ran out of time to do earlier as a result of being called out to an instrumentation fault in the middle of a storm (yay). So here is the first and most handy clue: [Scatterer] Custom sunflare cannot be added to Sun System.IO.DirectoryNotFoundException: Could not find a part of the path "/home/steve/Games/KSP_Linux_1.8.1/GameData_scatterer/config/Sunflares/Sun/sunSpikes.png". at System.IO.FileStream..ctor (System.String path, System.IO.FileMode mode, System.IO.FileAccess access, System.IO.FileShare share, System.Int32 bufferSize, System.Boolean anonymous, System.IO.FileOptions options) [0x00164] in <ad04dee02e7e4a85a1299c7ee81c79f6>:0 at System.IO.FileStream..ctor (System.String path, System.IO.FileMode mode, System.IO.FileAccess access, System.IO.FileShare share) [0x00000] in <ad04dee02e7e4a85a1299c7ee81c79f6>:0 at (wrapper remoting-invoke-with-check) System.IO.FileStream..ctor(string,System.IO.FileMode,System.IO.FileAccess,System.IO.FileShare) at System.IO.File.OpenRead (System.String path) [0x00000] in <ad04dee02e7e4a85a1299c7ee81c79f6>:0 at System.IO.File.ReadAllBytes (System.String path) [0x00000] in <ad04dee02e7e4a85a1299c7ee81c79f6>:0 at scatterer.SunFlare.start () [0x000ba] in <b83aac02f0a14d468c952212b6223826>:0 at scatterer.Core.Update () [0x001d5] in <b83aac02f0a14d468c952212b6223826>:0 I don't know where it's getting "KSP_Linux_1.8.1/GameData_scatterer", the real path is "KSP_Linux_1.8.1/GameData_Testing/scatterer", where GameData -> GameData_Testing is a symlink: readlink -e /home/steve/Games/KSP_Linux_1.8.1/GameData/scatterer /home/steve/Games/KSP_Linux_1.8.1/GameData_Testing/scatterer So, the pertinent question is: Why does every other plugin in the ~85 mods I'm running find it's files just fine in this situation, while scatter comes up with a mangled path if a symlink is involved? What's different in the way scatterer resolves canonical file paths? This is the only time I have ever seen a problem like this with any application whatsoever, and that's why it didn't occur to me to test without the link in the first place. Most curious. As for the original question - why the atmosphere halo in the stock game is messed up and what that new noise about an unsupported shader implies - that I'd also like some input on. Now technically off-topic for this thread, but still... And this time I swear it's got nothing to do with symlinks or strange file paths.

I'm reasonably sure this is a base game problem, but as it affects scatterer as well and all the shader gurus seem to hang out in this thread, I'm posting it here too in the hopes of a nibble. KSP 1.7.3 looks just fine. KSP 1.7.3 + Scatterer 0.0540 looks even better, just how it should. KSP 1.8.1 appears to have some acquired some funkiness, I swear the atmosphere shouldn't be shades of grey when seen from space: KSP 1.8.1 + scatterer 0.0541 (or 0.0540) is even funkier: Here be KSP.log. Here be Player.log. And here be a snippet from Player.log which appears in KSP 1.8.1, but not in 1.7.3: WARNING: Shader Unsupported: 'Hidden/Nature/Terrain/Utilities' - All passes removed WARNING: Shader Did you use #pragma only_renderers and omit this platform? ERROR: Shader Shader is not supported on this GPU (none of subshaders/fallbacks are suitable)WARNING: Shader Unsupported: 'Hidden/Nature/Terrain/Utilities' - Setting to default shader. Any thoughts, oh graphics gurus? KSP looking terrible I can handle, but scatter being upset is utterly depressing.

Sorry bud, you're on your own. I was planning on looking into this myself at some point, as I kinda like seaplanes and all, but now I'm outta here. If I were you, I'd check out the repo from github and have a poke around the math involving waterDragForce and hackWaterDragVal. It looks like there may have been some fudges installed to make water behaviour reasonable, by ferram''s definition of reasonable, and as much as that is based on real physics it doesn't always balance nicely with the rest of the unrealistic physics in KSP... Such as it being nearly impossible to build an aircraft as light and buoyant as the real thing. Parts in this game are apparently made of depleted uranium. The code is GPL so you can muck with it all you like, but you'll need to learn a bit of C# to make sense of it.

Perhaps, but when the reply to my "If you tell me what is wrong I might be able to fix it" in that instance was "where's your licence and source code", I feel it's rather closely related... Source code for a for a unified diff? Really? All those diffs over on the Debian bug tracker would like to have a word... The ultra-conservative interpretations of derivative work and the knee-jerk paranoia over copyright, to the point of piling additional rules on top of perfectly good licences or in some cases claiming violation before even ascertaining what the material in question is, that is exactly what this is about. If you're driving development and discussion to another platform, there's something horribly wrong with the way your community is run.

@Lisias, @Deddly We clearly aren't going to agree, and it's after midnight here so I'm not going to write another novel, as tempting as it is. Before the inevitable "agree to disagree" bit, as I call it a night, I'll leave you with something: This is not an isolated incident, and that's part of why it irritates me so much. The last time I tried to contribute on this forum I was blasted by another member of staff for posting a .diff, for his perusal, before he even looked to see what it was. He didn't open it, didn't even check the file extension, just went direct to "there's so much wrong with that thing I didn't look at, you don't know what you're doing, go away". That wasn't the first time, and it probably won't be the last. Anywhere else, I wouldn't even think twice about posting a patch file or a test build, and nobody would bat an eyelid. Here it's just not worth the angst. This place is so bizarre, so touchy, and so tangled in aggrandizing self-interest and but-covering legalese that I wonder if a collaborative spirit can even exist. Keep your possessive attitude, your paranoia and your strange rules. I'm out. I'm not helping anyone again, not here. So, let's agree to disagree. I'll stay for the bugs, but others can fix them. I'm done.

Trolls file against open source all the time, and they get thrown out all the time. I'm not particularly concerned, and I'm even less concerned in the context of a small and relatively obscure piece of code for a game. And what exactly is stopping me from running my own mercurial server? I'm sure people could clone that to wherever they wanted, if they wanted. People host all sorts of stuff on platforms that don't support cloning at all, and they're not getting copyright strikes every day either. My personal hosting has been up, running from my desk in my house, for almost exactly as long as github... Admittedly I only got me a valid SSL cert recently, but that's only because people are getting all paranoid about that sort of stuff these days. From a defending copyright point of view. From a computing freedom point of view, not so much. Also worth note is that copyright law where I live is marginally less screwed up than it is in the US, which may well be a point in favour of local hosting. I am extremely sceptical that I would ever need to defend myself from a copyright infringement for recompiling some open source code such as the one in question anyway. Big popular project? It's a consideration. KSP plugin? I think not. Oh, I see. That'll be that corporate bureaucracy vs old school hacker culture bit again. How I can use my skills to help others is usually of greater concern than some remote and unlikely threat to myself. That's just how I think. As with any other hosting that allows people to open accounts without a background check. The forum allows links to dropbox / onedrive etc. does it not? Those are not cloned by large numbers of people, there's no way of knowing who really owns the code there, and I know of several mods that host downloads on them. How is a storage space on someone else's server any different from one on my own in this regard? Sure. I have no problem whatsoever with making sources available. But so long as the source is provided, what difference does it make who signed up for the account that hosts it, or what machine the disks are on? Even if I had linked directly to a forked repo, there's no evidence to prove that I didn't just use one of my many internet-identities to open a repo, clone the authors code, and put an unrelated binary with a convincing name up there. How would that go for trust in licence compliance? Linking upstream source is a thing all over the place, I sure see links to upstream source in the forum thread in question. You mean aside from a fairly cursory inspection of the binary revealing that it's been built against .net assemblies several versions newer than the one on the github repo? Or the way I specifically mentioned the original author in the forum post, which happens to have been the only extant link to the file in question? The licence in question is quite clear in that one may provide attribution and indication of modification in "any reasonable manner based on the medium, means, and context" Context : Good faith, without any claim of rights. Medium: posted on forum. Means: text in said forum post. Sounds pretty reasonable to me. The post in which my link was embedded stated the identity of the original author, made it clear that I am not he, and defined the modification I had performed. Perhaps, though I would still argue that this is both implied given the context and reasonable obvious from examining the binary. As far as I am concerned, compiling some freely available open-licenced code is neither an original nor a derivative work, it's a public service. I didn't post an addon, I ran someone else's addon through a standard compiler to save another user a few minutes of their time. If squad wants to twist definitions and equate it to burning kittens, that's their loss. If I were to make code changes I would provide those changes, as the licence requires. I made no code changes, ergo there is nothing to provide, as links to source and licence are already readily available at the top of the same forum thread in which I provided the binary. Duplicating identical source-code repos is a waste of my time and bandwidth. If wasting time and bandwith are rules around here, I want none of it. In the context of overall FOSS, I can think of an awful large number of backport repositories that provide recompiled binaries without duplicating the source repos. Strangely, I've never heard of any of them encountering legal trouble. If we're so hell-bent on covering everybody in the case of legal dispute, perhaps we should reopen that "GPL code dependent on and linked against non-GPL code" can of worms? Seems to me there's a lot more Squad than everybody else in that protection plan. In the context of a simple recompile, they are exceedingly onerous. The requirements beyond those in the relevant software licences provide no tangible benefit to me or the people I try to help with such things, and involve more time and effort than a recompile itself. I'll gladly donate my time to helping other individuals, but I'll not waste it wantonly to satisfy some corporations byzantine leagaleese. I'm simply not going there. Sure, and anecdotes are not evidence. For something like this, I'm not particularly interested in having a bunch of witnesses to say "oh yeah, I've see [anonymous internet handle] around a bunch on github, he's probably legit". I trust myself more than either, so I store my private "cloud" instance on disks I control, with software I vetted and configured. I don't have or need accounts on github, dropbox, or any of the plethora of hosters hawking storage I don't trust on machines I can't find.

Everywhere sane, republishing is either piracy or mirroring for the good of all, depending on the licence. Apparently, in this twisted reality, republishing in a separate thread is seen as a hostile takeover attempt, and posting a quick fix in the existing ones is grounds for having posts deleted. How exactly are we supposed to go about this kind of thing then? You're not exactly leaving many options, as far as I can see it's a choice between annoying the modders and annoying the mods... I'll leave it to your imagination as to which one gets my gratitude and respect in this situation. Strange is an understatement, I'm going with demented. Recompiling a piece of code does not change it's intent, function, licence or ownership. This is true by convention, common-sense, and everywhere open-source software is found bar here. People come and go, they have holidays, sometimes a full-blown fork is unwarranted or met with hostility, and sometimes all that is needed is a quick rebuild. This is why we can't keep our nice things. I see. By which you mean making your job easier at the expense of those without the knowledge to rebuild plugins themselves, and making things gratuitously tiresome for those who want to throw them a bone? This is why the less computer-savvy members of this community can't have nice things. You'll note that the mod in question is CC licenced, it hasn't been updated since March 2018, and end-users were asking after it. I'm still not seeing how helping them out retarded free development, degraded end-user experience (ate your cat, anyone?), stepped on any author's toes, or created any strife on the forum... The only strife I see here was instead caused by the reaction of a certain powerful few to my donation of free-time. I think I'll spend my effort hacking on open source code in communities that respect the spirit as well as the letter of open software licences from now on, this place does nothing but harsh my buzz and spit in my face. This place is why trying to mix proprietary software bureaucracy with open-source hobby hacker culture is a horrifying idea. They seem to have a lot of blind and unwarranted trust in random github repos and whatever else google throws their way, I don't see how a private cloud instance is any different. I'd much sooner trust an individual than a corporation. Most individuals are basically decent.

The only trust involved is the word of the hoster and hostee. You take J. Random coder's word that the binary he provides was compiled from the source repo he claims it was, not just swiped from some unrelated website and uploaded to his account. My word is that the binary I linked was compiled from the source repo specified a few pages prior in the same document. I can't prove it any more than he can, and nobody can double check anything without verifiable builds or some fancy cryptographic signature that ties binaries to the source code and the coders DNA, none of which we have. So you're left with trust in human decency. What you really appear to be claiming here is that there is some inherent trustworthiness conferred by having a repository hosted by a large organisation, and frankly I don't see it. As far as I am concerned content hosted on a private server is every bit as trustworthy or not as something hosted on github/bitbucket/whatever corporate poison you choose. Github provides no evidence that a binary was compiled from particular source, or where the uploader got it. It'll tell you who uploaded it, insofar as logins can be trusted, but that's it. The forum software here told you a login called steve_v linked a file, and my nextcloud server will tell you a login called steve uploaded it. Who actually signed up for any of these accounts or how a file came about in the first place is completely unknown, the "evidence" is entirely comparable, and there is no chain of trust whatsoever. Perhaps you trust [insert well-known hoster] more than the box on my desk, but neither can prove that my name isn't really Muck the Terrible, and that I didn't steal someone else's work/implant a rat/cover it in cat poison. PGP suggests that you exchange keys in person, and provides a verifiable cryptographic chain of trust from then on. KSP addons are unsigned, unverifiable, and nobody goes to check the security of a modders github login in person... so where were you going with this comparison again? ... As the licence, attribution and link to source are available in the same forum thread, along with the rest of the components required to make the file I uploaded function, I assumed that would suffice for "any reasonable manner,". If I was wrong on that score, I'll be sure to make such information more explicit in future. Hence the use of the word "recompile" (which I'll be sure not to use in future as the mods clearly have an alert on it), rather than "modify". Indeed, thus "risk" not "are". It's still draconian, unreasonable, and apt to stifle development. The "process" that "includes an updated licence" and "sets up a public repository of your forked code", in the post snark linked bears no resemblance to either the official addon posting rules or the licence of the mod in question. In fact it appears to be little more than someone's personal and imposing interpretation, designed to discourage such activity. As I read it, all I actually needed to do according to the addon posting rules was include the licence file in the archive, and perhaps change "don't bug nightingale about it" to "all credit nightingale". Unless the definition of "every location you offer the plugin for download" WRT source code is individual posts rather than threads of course, in which case I should have linked to the original repo (as there are no code changes)... And mod authors can't post quick testing versions or recompiles in their threads without a new source link either - something which happens regularly. I'll use equipment I own thanks, that way I can set up any hosting I like for free, without using any quota at all, and without dealing with filth like Microsoft, Dropbox or Google. But I also digress.

Since we're playing hide-and seek now:

That is, frankly, ridiculous. Forking the repo / changing the license etc. etc. etc. for a compiled dll with zero code changes is insane. It's the exact same code as in the OP, so the same licences apply. Why I need to take ownership for clicking "build project" is beyond me. If this is what your rules imply, then your rules need a rethink lest we loose a bunch of perfectly functional but unmaintained mods. Imagine if FOSS development worked this way... every single linux distro forking upstream repos just to be allowed to build them from source... madness.