[0.18.1722 Linux x64 + mods] Game lockup / unhandled NULL exception/funlockfile

[2.71828]

I just tried running Rimworld in windowed mode (which for me is actually still fullscreen when it's the only window on my current screen), and it crashed again, same as before: Rimworld crashed to desktop, still blocks about 940 MB of RAM, has 0% CPU load (I had to kill -9 it again), and rimworld_log says:

Code Select Expand

[...]
Platform assembly: /home/<username>/.local/share/Steam/SteamApps/common/RimWorld/RimWorldLinux_Data/Managed/Assembly-CSharp.dll (this message is harmless)
Receiving unhandled NULL exception
#0  0x007ffeb5116c80 in funlockfile
#1  0x007ffeb5117230 in (Unknown)


(I'm not sure whether I'm supposed to laugh about the first line, or not...)

I also just tried the 32-bit executable (so 'RimWorldLinux.x86' instead of 'RimWorldLinux.x86_64'), also in window(ed) mode, but that's not an option, either, because (I think) the textures glitch out. My screen was plastered with giant rat skeletons and similar stuff.

But thank you for the second bug report and your investigations. None of the suggestions in this thread seem to work for me (so no Rimworld until this is fixed, which is really frustrating ), but at least it works for you.

[Supay]

Damn, that's a shame. Maybe I just haven't had it glitch yet then. I played for a couple of hours last night and again this morning with no crash, which is the longest I've been able to.

[Supay]

I've continued playing for up to 3 hours at a time with no crash still. I am using i3wm on Arch in a very minimal setup. I don't know if my specific config avoids the fault by setting to windowed mode, but it seems to work. I tried going back to full screen and it crashed not long in, so definitely something going on with that.

[2.71828]

Thanks for reporting back! While I didn't have any luck with Linux (I'm also using Arch, with spectrwm; I thought the window mode not working could be tied to me using a tiling WM, but if it works for you...), but it worked without any problems on Windows the last two days (with even more mods loaded than before). The only problem was the OS...

I'll try playing on Linux again and report back, because on Windows I played past the point where it crashed before. Maybe it works now.

[2.71828]

Nope, no bueno, as expected. My safe loaded up without any problems, so I paused, spawned a raid, put down some OP turrets in the vicinity, and unpaused. After a few shots: CTD.

Code Select Expand

Receiving unhandled NULL exception
#0  0x007fff739cb380 in funlockfile
#1  0x007fff739cb918 in (Unknown)


[2.71828]

Although apparently no one cares about this other than the people who are affected, I guess I'll keep posting...

Since I got fed up with using Windows today (not only does Rimworld take much longer to start up, the overall performance is also significantly worse, and today the game locked up entirely, after it slowed down to almost a crawl with a lot of stutters on anything above 1× speed...), I tried a couple of things on Linux again:

• Trying Rimworld out with another WM, this time Xfce. I though it might be worth a try, as using window mode seems to work for Supay. For me, it didn't work. Nothing changed.
• I did a clean reinstall of Rimworld (I don't know why I didn't try this earlier. I probably thought letting Steam verify the files would be enough...):
  
  
  • Using no mods and a fresh colony, I spawned a raid, placed a couple of turrets, and watched. Nothing out of the ordinary happened, the raiders died, no crashes, just how it's supposed to be.
  • I activated a single mod, namely ModSwitch, to see whether this problem is caused by just having any kind of mod activated: No problems. Not even a 3000 Pt. raid meeting an anti-grain-IED and a subsequent chain explosion of ≥ 20 turrets and a couple of batteries lead to any crashes.
  • I loaded the mod list of my actual colony (i.e. almost 200 mods), loaded my test colony, spawned a raid and set up defences again. This time, the crash happened just as before: After a couple of shots of who knows how many turrets, Rimworld CTD'd. I guess it happened when one of the raiders died.

This time, at least the 'unhandled NULL exception' was a bit more communicative than before:

Code Select Expand

Receiving unhandled NULL exception
#0  0x007ffe589778c0 in funlockfile
#1  0x007ffe58977e48 in std::vector<core::basic_string<char, core::StringStorageDefault<char> >, std::allocator<core::basic_string<char, core::StringStorageDefault<char> > > >::~vector()
#2  0x007ffe58977e58 in operator new(unsigned long)
#3  0x007ffe58977e78 in operator new(unsigned long)
#4  0x007ffe58977e88 in __run_exit_handlers
#5  0x007ffe58977eb8 in exit
#6  0x007ffe58977ec8 in mono_thread_exit
#7  0x007ffe58977ed8 in mono_method_marked_as_wrapperless
#8  0x007ffe58978198 in mono_amd64_throw_exception
#9  0x007ffe58978358 in (Unknown)

I attached the whole rimworld_log (unedited, except for every instance of my username, which got changed to '<user>'). Maybe this relatively clean log is of any help to anyone who actually wants to help.


Edit:
Because I'm apparently very bored/a masochist, I thought I'd try removing some of my mods to see whether something changes, starting with any mod that I added around the time this topic here was posted. I selected 'Path Avoid' (one of the reasons was that the OP also uses this mod, and that it's probably more error-prone than some kind of xml-only mod), and while without the mod I still get the same crash as always, I don't even get an 'unhandled NULL exception' anymore and rimworld_log just ends after some (probably) completely harmless error message from 'ArchitectIcons'...

[attachment deleted by admin: too old]

[Supay]

I have Path Avoid running and no issue.  I also have over 130 mods or so, and still no issue, so long as run in windowed mode under i3wm.  Could you try installing i3 with a default config file and seeing if Rimworld works fine for you that way as well?  My PC is also ancient, a CPU/mobo nearly 11 years old with only 8GB RAM and a 2011 Nvidia 560Ti.  So I don't think it is system capability/performance related, particularly as now my colony is large I am seeing it slow down performance wise yet now with no crash, whereas a handful of mods and no more than a few rooms built with 3 colonists was crashing constantly.

[2.71828]

Thank you for your report and for being so persistent about the window manager!  I was pretty much ready to install i3 and give it another go, but wanted to test out the last of the 3 WM I have already installed, i.e. xmonad (I used to use xmonad and then switched to spectrwm, because I wanted to get rid of all the Haskell dependencies; Xfce was a fallback WM). I didn't think it would work now, but I just tested my test safe, and got through two dev spawned raids without a crash (once with only vanilla turrets, once with turrets from 'More vanilla turrets' [I thought that might make a difference]). What I did get this time, though, were new error messages of the following kind:

Code Select Expand

Exception ticking Clark: System.ExecutionEngineException: SIGILL
at (wrapper dynamic-method) RimWorld.InteractionWorker.Interacted_Patch2 (object,Verse.Pawn,Verse.Pawn,System.Collections.Generic.List`1<Verse.RulePackDef>) <0x00000>
at (wrapper dynamic-method) RimWorld.InteractionWorker.Interacted_Patch2 (object,Verse.Pawn,Verse.Pawn,System.Collections.Generic.List`1<Verse.RulePackDef>) <0x00027>
at (wrapper dynamic-method) RimWorld.Pawn_InteractionsTracker.TryInteractWith_Patch1 (object,Verse.Pawn,RimWorld.InteractionDef) <0x004ae>
at Hospitality.Harmony.Pawn_InteractionsTracker_Patch/TryInteractRandomly.Replacement (RimWorld.Pawn_InteractionsTracker,bool&) <0x003f6>
at (wrapper dynamic-method) RimWorld.Pawn_InteractionsTracker.TryInteractRandomly_Patch1 (object) <0x00046>
at RimWorld.Pawn_InteractionsTracker.InteractionsTrackerTick () <0x00151>
at Verse.Pawn.Tick () <0x00372>
at Verse.TickList.Tick () <0x003b8>
(I think 'Clark' might be the name of some raider that got killed?)

I'm now guessing that Rimworld crashed with spectrwm and Xfce, because those WMs react differently to SIGILLs, but I have no idea if that even makes sense. Do you also get these kind of errors?



I guess this means that I can now play on Linux again, but it certainly doesn't mean that there are no problems anymore. Is the error message above of any more use to the devs than what we have posted before? Could we get some kind of feedback?

[2.71828]

Ok, it seems I celebrated too early. I just loaded my normal save and unpaused. After a few in-game hours, without any kind of raid happening, Rimworld crashed again, without writing anything useful to rimworld_log, again. The only thing that was at least a tiny bit out of the ordinary was Alpha beavers appearing, which I then tagged for hunting, that's it. I don't think that any of my colonists were even up, because I saved in the middle of the night, and it wasn't even morning yet when the crash happened.

What the flying f* makes Rimworld crash all the time? Why did my test save work all of a sudden? 



@Supay: I just thought about something: Are you running 'steam-runtime' or 'steam-native'?

[2.71828]

Another (and probably final) report:

I started another colony with a slightly different set of mods (not really less, though), 'steam-runtime' instead of 'steam-native' (which I used before), and xmonad. Yesterday, I could play this colony for a couple of ingame days, without any problems, so I thought using 'steam-runtime' (and its out-of-date libraries) must have done it. Well, of course it hasn't. I just tried to play my old colony again, and after I unpaused: CTD. Then I thought: maybe this save is just fucked, so I reloaded my other, new save, that was working flawlessly yesterday. Guess what: CTD after I loaded the save, paused immediatly, went afk for a couple of minutes because I had to do something, and then unpaused. So this time, the CTD happened immediately after I unpaused. No shooting, no one dying, nothing. Just ye olde 'unhandled NULL exception'

Code Select Expand

Setting breakpad minidump AppID = 294100
Steam_SetMinidumpSteamID:  Caching Steam ID:  76561197990898994 [API loaded no]
The class Vampire.VampireUtility could not be loaded, used in Vampire, Version=1.18.0.0, Culture=neutral, PublicKeyToken=null
The class Werewolf.WerewolfUtility could not be loaded, used in Werewolf, Version=1.18.0.0, Culture=neutral, PublicKeyToken=null
The class Vampire.CompVampire could not be loaded, used in Vampire, Version=1.18.0.0, Culture=neutral, PublicKeyToken=null
The class Werewolf.CompWerewolf could not be loaded, used in Werewolf, Version=1.18.0.0, Culture=neutral, PublicKeyToken=null
The class Werewolf.CompWerewolf could not be loaded, used in Werewolf, Version=1.18.0.0, Culture=neutral, PublicKeyToken=null
Stacktrace:


Native stacktrace:

/home/<username>/.local/share/Steam/SteamApps/common/RimWorld/./RimWorldLinux_Data/Mono/x86_64/libmono.so(+0x91c4d) [0x7f52cc0a4c4d]
/usr/lib/libpthread.so.0(+0x11d9f) [0x7f52da700d9f]
/usr/lib/libc.so.6(gsignal+0x110) [0x7f52d9e17860]
/usr/lib/libc.so.6(abort+0x1c8) [0x7f52d9e18ec8]
/home/<username>/.local/share/Steam/SteamApps/common/RimWorld/./RimWorldLinux.x86_64() [0x9f67cb]
/home/<username>/.local/share/Steam/SteamApps/common/RimWorld/./RimWorldLinux_Data/Mono/x86_64/libmono.so(+0xd0cb2) [0x7f52cc0e3cb2]
/home/<username>/.local/share/Steam/SteamApps/common/RimWorld/./RimWorldLinux_Data/Mono/x86_64/libmono.so(+0x34a7d) [0x7f52cc047a7d]
/usr/lib/libpthread.so.0(+0x11d9f) [0x7f52da700d9f]
[0x405d5d8d]

Debug info from gdb:


=================================================================
Got a SIGABRT while executing native code. This usually indicates
a fatal error in the mono runtime or one of the native libraries
used by your application.
=================================================================

[...]

Receiving unhandled NULL exception
#0  0x007fff040c7cc0 in funlockfile
#1  0x007fff040c8250 in (Unknown)




I think I'm giving up. None of the people who could probably do something about this seem to give a fuck (for example, the only one who even looked at the last rimworld_log I attached was me, unless the statistics aren't correct). At this time, I'd actually be happy with a 'We'll look into it', 'It's a mod problem, play it without mods' or even 'We don't care, play it on Windows or GTFO', but this is worse.

If I hadn't already sunk so many hours into A17 in the short time that it was possible for me (I bought RimWorld on 9th November, didn't get the memo about B18 approaching and of course neglected to backup any of my mods), I'd get a refund. For me, (at the moment) RimWorld B18 is broken and unplayable.

[jooe]

I'd just like to point out that several people have been reporting this bug which appears to be a base game or even unity engine bug - mods should not be able to cause unhandled exceptions in any software.

We haven't received any response in more than 3 weeks now, which is kind of dissapointing considering the amount of information delivered. At least a one-liner like "seen it, other stuff is more important, on the to do list for Beta 21" would be nice, because it looks like as if noone cares

Myself, I'm currently not playing Rimworld because of this bug ...

[2.71828]

I'm pretty sure now that the devs really don't care one bit about this and that this won't be fixed. They do seem to look at the bug section of the forum at least occasionally: Example; someone posted a (possible) bug, and it was acknowledged as a bug only a day later by Tynan himself, and then another dev posted two hours later that the bug was already fixed. When I saw this, I thought: 'Wow, that's great support. I bet our issue will be fixed/looked at in now time.' Still, our/your thread is pretty much a support group of 4 people talking among themselves for more than 3 weeks.

Maybe I'm doing the devs an injustice and they're already working on it, but there has literally been zero feedback, so what else am I supposed to think.

[Calahan]

We haven't received any response in more than 3 weeks now, which is kind of dissapointing considering the amount of information delivered. At least a one-liner like "seen it, other stuff is more important, on the to do list for Beta 21" would be nice, because it looks like as if noone cares

I'm pretty sure now that the devs really don't care one bit about this and that this won't be fixed.

The devs certainly do care about this bug, along with all the other bugs that players have gone out of their way to report. RimWorld wouldn't be anywhere near as polished as it is today without the thousands of bug reports that players have filed. But it's not just this bug that the developers have not responded to recently, as of you take a quick look around the bugs forum you will see that there are dozens of open bug threads that are awaiting input from the devs.

Why is this you might ask?...

Well, and please don't quote me on this, or take this a gospel, as I could be 100% wrong. But my general observation from helping organise the bug forum for the past ~18 months is that the devs seem to work in fairly strict development stages.

After a new alpha beta is released, ison (and sometimes Tynan) will for a short while go through the bugs being reported from the new beta, fixing them (for their development build) as they go. My hunch is they are just checking in case any game-breaking bugs or issues exist that need an emergency patch. And by way of example, if you currently look back 7-10 pages in the bugs forum you will see dozens of closed threads where ison is the last poster. These threads being B18 bugs that ison has addressed, and then closed the thread to mark it as resolved.

Once they are sure there are no game-breaking bugs, the devs then seem to move onto a heavy development stage where they put their heads down and concentrate entirely on developing whatever features are planned for the next update. And during this particular stage it's rare to see the devs post in the bugs forum, and basically because they are not on the bug fixing stage at this point. I'm pretty sure this is the stage the devs are currently on right now. Which can be seen by the lack of any recent posts by the devs in the bugs forum, and from the limited number of recently closed threads (with those that are closed likely done so by myself because they are non-bugs or duplicate reports etc.).

Then, and usually about a 4-6 weeks before internal testing on the new beta begins, you will see ison methodically working his way through all the open bug threads that have been posted (although they usually start with the bugs on the Mantis first https://ludeon.com/mantis ). It is during this stage that one of the devs, and most likely ison, will almost certainly comment on this bug, and provide an update on its current status and any ETA on a fix. This bug fixing stage then continues until the next update is released, after which the cycle starts again from the top (although I don't know if the shift from Alpha to Beta will change any of this).


Although as I said please don't quote me on the above as it's not official in any way, and is nothing more than my own observation of how the developers allocate their time between updates (my moderator position gives me absolutely zero additional insight or info into the development of the game. As I said, it's purely my own observations as a regular forumite and from regularly manning the bugs forum for the past ~18 months).


So this issue will certainly be addressed, but my prediction is that that won't be until the devs (or possibly just ison) shift to the bug fixing stage of their development cycle. Which is an unknown period of time away from now. Plus due to the nature of this bug, and if it is strictly a Unity problem, then it is likely that the devs will be unable to do anything about it apart from updating Unity for B19 (assuming that's what the next update will be called) and hoping that that will fix the problem. Although that's also assuming there has been a Unity update during this period, which there might not have been (and even if there has been, it would only fix this bug if Unity Technologies fixed this bug in their engine on their end).

So I guess this post by me will have to do in lieu of a post from the devs, and until they move onto the bug fixing stage. And I haven't commented before now because there's basically nothing about this issue that I can comment on, other than to say "thank you for the bug report, the developers will address this when they move onto the bug fixing stage of their development cycle".

[jooe]

Thank you Calahan for taking the time to write this.

Personally, I would have felt much better if you had just written this one line "thank you for the bug report, the developers will address this when they move onto the bug fixing stage of their development cycle". Because now it feels like someone is listening and I can relax and be patient for the time being. Well, relaxing is maybe not the right word as I'll be working night shifts during the holidays...

[Roolo]

I'm posting this since it might help some of you. While I'm also convinced it's a base game issue, my mod RunAndGun does seem to trigger it (maybe indirectly). If you have RunAndGun installed, I advice you to put it as high as possible in the load order, as some users reported that this helped. It won't completely solve it, as some people seem to have the issue without any mods loaded, but it might help at least.