Wat's your worst Ksp 1.0 glitch?

[mythbusters844]

This:

Javascript is disabled. View full album

Click here to view the full album with bigger screenshots

This has to be the most eccentric bug I've ever run into.

[GudPiggeh]

This:

This has to be the most eccentric bug I've ever run into.

The time warp mod often blows up your ship when time warping at very fast speeds.

[RocketPilot573]

[Canopus]

Spontaneous mid-flight disassembly due to parts clipping though the Service bays. Ruined a whole Spacestation for me.

[Capt. Hunt]

I've been having a bug lately where launch clamps will randomly spawn behind my ship as it is climbing through the atmosphere, usually they just fall back to the ground harmlessly, but occasionally they'll spawn inside the rocket and cause it to CATO.

Another time, I had the big 4 stranded on Duna, I sent a series of resupply ships to help them out, but the first one got attacked by a Kraken a few days out from the Duna encounter. I guess the Kraken likes snacks. They had to starve for another month before the next resupply lander could reach Duna. Of course, at that point Mechjeb malfunctioned somehow and put the lander down a mile off target.

[Zild]

One of my first attempts to do a rescue mission (which was in V1.0) involved a ship design with two capsules, one mounted above the other and upside down. (I mention this because it was the only unusual thing I remember about the mission, up until it ended of course...)

Once back safely on Kerbin, the rescuee Kerne went EVA to collect some simple science. As soon as he opened the hatch she was bounced down into the ground at high speed and promptly ceased whatever bodily functions Kerbals find important.

In for a credit in for a hundred credits, as they say, so the pilot Valentina also went EVA to collect some science. She opened the hatch and vibrated wildly on the ladder for a moment before spontaneously evaporating.

And then, apparently still hungry, the space kraken consumed the entirety of Kerbin as well...

[WWEdeadman]

http://www.youtube.com/watch?v=L1wDCdtTfNc?t=31m49s

This. Spinning that makes no sense, caused by some phantom force, that only sets in when SAS is off. There is literally nothing that should cause this spinning. (I didn't set any trim, and I didn't touch the controls.) Edited June 9, 2015 by WWEdeadman

[Red Dwarf]

Well, I've encountered some more bugs since my last post here. I've been working on a new spaceplane, the Muon, which has some great potential. However, the whole process has been slowed at every turn by bugs. My initial design was way to complex, and it very frequently put stress on the physics engine, especially in the SPH. There were fuel flow problems, symmetry issues, and our dear old friend Crash To Desktop.

I abandoned that design, and moved to something simpler and more efficient on the part count. That design worked a lot better, but the plane had a tendency to snapping in half for no reason. I trimmed up the design some more, and encountered some bugs we haven't seen in ages.

It's actually a combination of two bugs, one old and one new. The old one is the Collided With The KSC bug, which happens anywhere from a hundred meters to a hundred kilometers from the actual phsyical KSC. I was coming in on an approach from the ocean, and was almost touched down on the runway, when the landing pad exploded and some of my control surfaces disappeared. Apparently my wing had clipped one of the fuel tanks near the pad, despite it being hundreds of meters away.

The other bug is something I'm deciding to call the Shockwave bug. I first discovered it when I was doing flybys of the KSC in one of my supersonic jets. On one of my passes, I flew over a Duna base mockup I had been working on, and the whole thing exploded. The mission report didn't read any collisions with the plane itself. Apparently the sheer speed of the jet flying so close to the ground blew it to pieces. I think the Shockwave bug had something to do with the Muon's landing incident, as I had left a test rig on the pad, which was also destroyed. Hopefully I can work out ways to avoid these bugs before I move the Muon over to my career save for official use.

[Popeholden]

the most annoying, but minor, glitch is the heatshield fairings disappearing if i unload and then unload a vessel. i hate having my ships magically linked together with a big space where the fairing is supposed to be.

very minor, now that i type it out

[Brainlord Mesomorph]

Reading all this makes me glad I'm still playing 0.25

I hope they have all this fixed by the time I get to 1.0

[PB666]

The mun disappearing on hard landing or impact. That would be my worse so far.

[Chel]

I suffer from loading screen freeze, every button being unclickable, the game just outright destroying itself, losing every part, craft, flag, even Kerbals, and even losing Minmus at one point. It's SO ANNOYING.

[steve_v]

1.0.x: Temperature gauge memory leaks, random spontaneous overheating, ridiculous parachute bugs, constantly changing aero model. Even lousier performance than alpha versions. Big letdown from 0.90.
1.1.x Improved performance, rendered irrelevant by constant crashing on GNU/Linux. Totally borked wheels. Pretty much didn't play 1.1.x at all.
1.2.x Window resizing madness, crashing of several common window managers. CTD if pulseaudio not installed, though not mentioned in system requirements. (also on GNU/Linux), terrain seams. Despite this, the only post-beta release worthy of the name, IMO.

As always, I'm hoping for a 1.3 that contains no serious bugs, and doesn't require 4 hotfixes to earn the label "release". With history as my guide, I am not optimistic. C'mon QA, prove me wrong.
 

Edited April 19, 2017 by steve_v

[Numerlor]

Completely wrong velocity while quick loading 100m above surface

[LoSBoL]

Just one bug is haunting me. I can't select an planted flag at the KSC as target, every time I try to do so KSP just hangs on me.

Could be an mod though.

I think KSP is pretty stable, as long as you have a descent and stable computer and keep it clean from to much clutter.

[steve_v]

31 minutes ago, LoSBoL said:

I think KSP is pretty stable, as long as you have a descent and stable computer and keep it clean from to much clutter.

Nothin' wrong with clutter. I got ~12TB of clutter and it don't destabilize me... much *mutter mutter* . Too many cats is what makes computers unstable, and mighty indecent too.*
All my clutter makes my desk super stable, holds it from floating away and that. My computer is on my desk, so it too must be stable.**

*A picture of a cat is worth 6.73 live cats. Watch out for that, it sneaks up on you.
** 1 logic is worth -13.49 cats, you too can offset your cat count.

[sal_vager]

3 hours ago, steve_v said:

1.2.x Window resizing madness, crashing of several common window managers.

Which is (still) a Unity 3D issue, try Sunless Sea 64bit in a window and you get the same result, not known if fixed in Unity 5.6

 

3 hours ago, steve_v said:

CTD if pulseaudio not installed, though not mentioned in system requirements.

Unity bugs are system requirements now? While KSP has its faults this is not one of them (also, fixed in Unity).

 

3 hours ago, steve_v said:

C'mon QA, prove me wrong.

Oh sure, after accounting for every possible permutation in players hardware, software, middleware, malware, firmware, drivers, 'stock' installs, 'near stock' installs, "I consider mod X stock' installs, network drive installs, protected folder installs, restricted permission installs, non-target platform installs, corrupt installs, 32bit Windows installs, overclocking, drive fragmentation, Incorrect-Unity-values-in-registry, cpu resources eaten by Chrome etc, insufficient memory and just plain good old fashioned broken operating systems so the engine can't even start.

And after all that it still requires programmer hours to fix any actual KSP bugs.

I recommend a nice clean supported OS install on known good hardware with no unnecessary software running, if you have an issue try a second PC to rule out local issues.

KSP itself is remarkably stable and low on bugs in 1.2.2, most issues I see are either caused by mods or by the system KSP is being run on.

But I'll bite, the building tier bug in 1.2.2 is a shame (it's a regression), though easily worked around and it's fixed in the pre-release

[steve_v]

50 minutes ago, sal_vager said:

Which is (still) a Unity 3D issue, try Sunless Sea 64bit in a window and you get the same result, not known if fixed in Unity 5.6

So what? The offending binary that is passing garbage to the window manager is named KSP.x86_64. KSP stands for Kerbal Space Program, right?
Also, thanks for that bit of information, I was contemplating buying Sunless Sea. Now I won't, at least not until it's fixed.
 

50 minutes ago, sal_vager said:

Unity bugs are system requirements now?

If the game will not run without certain additional software installed, it's a system requirement and should be listed as such - just like, say, DirectX 9.0. As for why, or who we're passing the buck to, I don't really care.

 

50 minutes ago, sal_vager said:

Oh sure, after accounting for <snip>

None of which can be blamed for any of the issues I mentioned, and none of which should affect QA... Unless your QA testers are all using unstable, misconfigured, malware ridden overclocked machines? What's your point?
Was stock parachutes being non-functional on launch day due to overclocking?  Or perhaps the infamous temperature gauge memory leak was caused by drive fragmentation? The mind boggles.
 

50 minutes ago, sal_vager said:

KSP itself is remarkably stable and low on bugs in 1.2.2, most issues I see are either caused by mods or by the system KSP is being run on.

This I will agree on, it is pretty stable... Once engine regressions have been worked around.
Here's what my first impression of the 1.2 prerelease looked like (after applying community fixes for pulseaudio):


It still does this, but hey, at least it doesn't CTD at random like 1.1 did.

Edited April 19, 2017 by steve_v

[sal_vager]

1 hour ago, steve_v said:

So what? The offending binary that is passing garbage to the window manager is named KSP.x86_64. KSP stands for Kerbal Space Program, right?

A binary that is the same for any Unity game, it's just renamed, nothing else in there but Unity code no matter what it might be called.

But it's only worked around instead of really fixed, but that is out of Squads hands so citing this as a KSP glitch, as per the thread title, is a little unfair.

I'll just quote this guy.

 

1 hour ago, steve_v said:

If the game will not run without certain additional software installed, it's a system requirement and should be listed as such - just like, say, DirectX 9.0. As for why, or who we're passing the buck to, I don't really care.

Well the system requirements do say you need Ubuntu or SteamOS if you're using Linux, both of which include Pulseaudio by default so it's expected that this is available, a user would have to deliberately uninstall Pulse to run into this bug, or should Squad also add the X server to the requirements page? Unity doesn't explicitly state you need xorg so it should work without it, right?

 

1 hour ago, steve_v said:

None of which can be blamed for any of the issues I mentioned, and none of which should affect QA... Unless your QA testers are all using unstable, misconfigured, malware ridden overclocked machines? What's your point?
Was stock parachutes being non-functional on launch day due to overclocking?  Or perhaps the infamous temperature gauge memory leak was caused by drive fragmentation? The mind boggles.

Well, except this one...

1 hour ago, sal_vager said:

Oh sure, after accounting for every possible permutation in players hardware, software, middleware, [snip]

 

Fact is we see a lot of issues that can't be reproduced, are local to one or a few players and turn out to be something external to KSP making tech support much harder these days, so I hope to see people like yourself who have some savvy with PC's come to others aid, but that becomes harder when people write off KSP as bug ridden rather than try to diagnose the problem, that's a meme that really needs to die.

A recent example, Avast! on Windows is now interfering with KSP.

If an issue isn't in Squads code it shouldn't be stated as such.

Edited April 19, 2017 by sal_vager

[sal_vager]

Oh I better add...

53 minutes ago, steve_v said:

As for why, or who we're passing the buck to, I don't really care.

The buck for issues like these may rest with Unity Technologies, but those of us who can figure these out or simply keep an eye out for fixes by others should do what we can to let others know how to deal with them, just as we did and do for issues in KSP itself.

[steve_v]

17 hours ago, sal_vager said:

Squad also add the X server to the requirements page?

Might be a good idea, Ubuntard is planning on switching to Wayland soon...

 

17 hours ago, sal_vager said:

Fact is we see a lot of issues that can't be reproduced, are local to one or a few players and turn out to be something external to KSP making tech support much harder these days

True, however again this has no real bearing on QA, it's simply out of scope to test on misconfigured setups.
As for the pulseaudio issue, Unity (and by extension KSP) is advertised as being able to run without it, and hence lacking it is not a misconfiguration of the users machine.
This previously working functionality clearly wasn't tested at all, yet QA is supposed to test for regressions, is it not?
 

17 hours ago, sal_vager said:

I hope to see people like yourself who have some savvy with PC's come to others aid

And I do. However I am pretty tired of situations where the only valid reply is "known bug, several releases old now".
 

17 hours ago, sal_vager said:

that becomes harder when people write off KSP as bug ridden rather than try to diagnose the problem, that's a meme that really needs to die.

Releasing with known bugs and leaving serious engine problems unfixed for several patches is not helping here. They may be Unity bugs, but they're going to reflect badly on KSP anyway - particularly for those who lack the motivation or knowledge to dig  around on the bugtracker.
KSP being bug-ridden is a meme because history says so. Name one post-beta major release that didn't need immediate hotfixes.
That this happens again and again can only be described as a QA failure, or releases marching to a marketing-department tune. And it's clearly built up a reputation.

Take #11382, for example: It has now been worked around for 1.3, without upstream code fixes. Yet for the entire 1.2 cycle it was written off as "Unity bug, only Unity can fix it, case closed".
Or the pulseaudio crash: I don't see an official SQUAD workaround anywhere, rather it was the community that put in the effort to get the game running. Where's the effort from SQUAD on this one? This doesn't reflect well on them at all.

Don't even get me started on how "double free or corruption" memory mismanagement is not a user-end misconfiguration. And this was, again, known about on release day and left unfixed for an entire cycle.
Was there an official memo in the release notes? Any kind of warning for those not following the bug tracker or keeping an eagle-eye on the forums?
Ditto for having a standards-compliant window manager that respects EWM hints...

How about save-destroying console port bugs? We're at, what, 7 months without a patch now? Maybe it's FTEs fault, but have a wild guess who looks bad here... Yup, it's those guys with a big logo on the loading screen.

 If you keep tossing game crashing/breaking bugs in the too hard basket and releasing regardless, is it surprising that people write off KSP as bug ridden?

Edited April 20, 2017 by steve_v

[sal_vager]

On 4/20/2017 at 7:39 AM, steve_v said:

Might be a good idea, Ubuntard is planning on switching to Wayland soon...

...

Got a little issue with attitude there (emphasis mine), and I'm sure when Wayland is more established Unity will build a version of their engine for it.

Just a thought on #11382, we were all running round the houses on that one, Squad staff and players alike, sure it turns out it was a setting bu no one knew that, aren't you glad it's finally no longer going to be an issue?

The double free issue is legit, your first one so far, what you don't know is unfortunately covered by the NDA between Squad and myself, which leaves me at a disadvantage when speaking with you.

As for FTE? I can't say a bad word about them, not because of the NDA, but because I respect their team.

And I think you have a very different idea of what Squads QA does compared to the reality, they (we) don't have the power to delay a release, or force development changes much as I'd like to, you can read more about how it works here but the producer, lead programmer, designer et al call (most of) the shots on what the game is, which to my understanding is the same elsewhere in the games industry, serious production code may well be different.

Actually all this reminds me of something, I'm an old gamer, I started with 8-bit machines, moved onto Dos gaming etc, I've seen my fare share of buggy games, and I've seen the tolerance from gamers go down just as the complexity of games has increased.

Is this fair?, right? should we be accepting of poor code? should we be so quick to judge and find people to blame?, some would say yes, some would demand and expect nothing but (their idea of) perfection, others jump on the hate bandwagons, that has consequences though, look at Phil Fish and Fez for an example from the gaming industry, and from outside of gaming we have Ubuntu dropping Mir and Unity (both good, well programmed software) because of hate.

Back in my early gaming days we were pretty happy just to get games, the really bad ones didn't sell as well due to word of mouth but everyone has different tastes, your crap game could be on my top shelf.

You obviously love KSP, you wouldn't still be here otherwise, and you want KSP to be the best it can be, that's different to every player and what we forget is that there's real living human beings behind every game, poring their love into their vision of what that game should be.

Sure some are buggy, some are not your cup of tea, but we still have fun or we can appreciate that others enjoy a game we'd not touch with a barge pole (usually because we didn't like what we heard about it).

KSP's pretty special as it hold your attention for so long, and we all want it to be the best so everyone else sees it the way we do, including the QA team who've seen things in KSP far worse than the occasional memory crash or a bit of slowdown.

I think when we find a bug in a released game we should remember all the rest of the game that's brought us so much fun, and remember that programmers have as much if not more love for the game than you do, they are the ones working on it after all.

People make mistakes, but they also make the games we love.

[steve_v]

14 hours ago, sal_vager said:

Just a thought on #11382, we were all running round the houses on that one, Squad staff and players alike, sure it turns out it was a setting bu no one knew that, aren't you glad it's finally no longer going to be an issue?

Sure am. But my point was, if it's sorted now, why have I been hearing "Unity bug, can't fix" for so long?
 

14 hours ago, sal_vager said:

The double free issue is legit, your first one so far, what you don't know is unfortunately covered by the NDA between Squad and myself, which leaves me at a disadvantage when speaking with you.

Ahh, so the cries of "Unity Unity Unity" from official channels was just a smokescreen? That figures.
 

13 hours ago, sal_vager said:

And I think you have a very different idea of what Squads QA does compared to the reality, they (we) don't have the power to delay a release, or force development changes much as I'd like to, you can read more about how it works here but the producer, lead programmer, designer et al call (most of) the shots on what the game is, which to my understanding is the same elsewhere in the games industry, serious production code may well be different.

This may be so for KSP, but all the projects (not games, generally) I have been involved with see bugs that crash the application or prevent it from running on supported configurations as extremely good reasons to delay release.
Honestly, this is probably the bit that yanks my chain the most: Issues are raised in pre-releases, by people testing in their own time, then they're not addressed for release. Then they're not addressed for patches either.
Kinda feels like SQUAD asking for people to find bugs, then only listening to them when it suits the marketing department.

 

14 hours ago, sal_vager said:

Actually all this reminds me of something, I'm an old gamer, I started with 8-bit machines, moved onto Dos gaming etc, I've seen my fare share of buggy games, and I've seen the tolerance from gamers go down just as the complexity of games has increased.

So am I, and back when games were distributed on floppy disk or CD, needing 4 hotfixes after release (which many would never get) would be a death-knell. And very good cause for a acquiring a "bug-ridden" reputation.
Yes, complexity has increased, but the tools to deal with this complexity have also improved, and I can't help thinking that a lot of the release-day bugs are there because "Yeah, we know, but if it becomes a problem we can just patch it later".

I'm not annoyed that there are bugs, there will always be bugs. My issue is known bugs that make it into releases and then go unfixed for an inordinate length of time.
"It's a new bug, they're working on it" paints a rather different picture of the game than "It's a known bug from the prerelease, 3 or 4 patches ago".

 

14 hours ago, sal_vager said:

You obviously love KSP, you wouldn't still be here otherwise, and you want KSP to be the best it can be, that's different to every player and what we forget is that there's real living human beings behind every game, poring their love into their vision of what that game should be.

Indeed, and I dislike the "KSP is bug-ridden" meme as much as anyone else. But the only way to kill it is to stop making buggy releases.

Anyway, this has been off topic for a while now. If you want the last word, it's yours. I've said my bit.