Random crashes with eufloria 2.02 -- linux + crossover-games

[Don]

Hi,

    Yes....I know I'm on linux, using crossover-games, and as such it's tarred with that brush...however...

  http://www.dyson-game.com/smf/index.php?topic=407.0

   ...is looking to me like exactly the same issue - maybe I can chip in on this front and help isolate
   this pea in a pod .... my hidden agenda, is to try and get this fixed so I can progress past level_10  B-)


Obviously, there's going to be differences to how windows reacts when it triggers this sort of thing,
and how crossover's wine engine is going to react to the same situation. The link above, the various
descriptions of that problem, fit the example of what I'm seeing here wrt the full 2.02 version and my
drunken_penguin situation....here's the comparo ;

* Eufloria will often fail to start with crossover/linux (be that crossover-games or crossover-office). It
initializes the main wine window, displays background color and the loading progress bar, which
then will stop on the first dot of the progress indicator. I've got some trace on this..enough to suspect
it's a .NET initialization error, and this *may* be related to the crash here, I'm not sure...but....the effective
workaround for this startup glitche *is* something worth mentioning -- if this happens, and Eufloria fails
to start (repeatedly) like this, what I've discovered works here is...if you have firefox running, shut it
down...90% of the time, Eufloria will then start properly - conversely, if you don't have firefox running
at the time, start it up...again,  90% of the time, Eufloria will then start properly. Same is true for some
other applications as well, not just firefox can be used to effect this...pretty strange, mentioned here
only for completeness - if you want any trace, let me know...

* In the thread linked above, people report the windows case as "the game will spontaneously give me
an error and say it must close" -- this is not what I see in crossover/linux....ie; I see no error message.
I can probably explain that, with the following snippet of trace...

Code: [Select]

fixme:advapi:RegisterEventSourceW ((null),L".NET Runtime 2.0 Error Reporting"): stub                                                                                                                           
fixme:advapi:ReportEventW
(0xcafe4242,0x0001,0x0000,0x00001388,(nil),0x000b,0x000000e0,0x3009a1b4,0x7dc5e4): stub                                                                                               
err:eventlog:ReportEventW L"clr20r3"                                                                                                             
err:eventlog:ReportEventW L"eufloria.exe"                                                                                                       
err:eventlog:ReportEventW L"2.0.0.0"                                                                                                             
err:eventlog:ReportEventW L"4b12e52b"                                                                                                           
err:eventlog:ReportEventW L"eufloria"                                                                                                           
err:eventlog:ReportEventW L"2.0.0.0"                                                                                                             
err:eventlog:ReportEventW L"4b12e52b"                                                                                                           
err:eventlog:ReportEventW L"733"                                                                                                                 
err:eventlog:ReportEventW L"5e2"                                                                                                                 
err:eventlog:ReportEventW L"system.typeinitialization"                                                                                           
err:eventlog:ReportEventW L"NIL"                                                                                                                 
fixme:advapi:DeregisterEventSource (0xcafe4242) stub                                                                                             
fixme:thread:NtQueryInformationThread info class 9 not supported yet
For me, the issue presents itself as follows - start the game, start level_10 (note: level_6
was doing the same thing for a week before it managed to somehow not bug), the game
proceeds as per normal, and will just simply 'stop' dead in it's tracks....ie; all my seedlings
stop moving, mouse input no longer recognized...but if I look in processes list, it appears
as though Eufloria.exe is still happily running. When then I click the close window widget
on the main wine window, it's actually my window_manager that comes back to tell me
the process has stopped responding, not wine itself...so it *appears* like Eufloria.exe has
totally detached itself from the real world when this happens....which is sort of symbolic,
I agree, but not the desired behavior 

This can (and does) occur regardless of user input...I can start level_10, populate the next
nearest asteroid and walk away and just wait...it'll happen. The last thing my winedebug runs
see at the time of this incident is ;

Code: [Select]

err:seh:setup_exception_record nested exception on signal stack in thread   
0032 eip 7bc69360 esp 7ffdbc7c stack 0x242000-0x340000

Although I did catch one trace going down were the error was reported as ;

Code: [Select]

fixme:advapi:RegisterEventSourceW ((null),L".NET Runtime"): stub                                                                                 
fixme:advapi:ReportEventW (0xcafe4242,0x0001,0x0000,0x000003ff,(nil),0x0001,0x00000000,0x33e170,(nil)): stub                                                                                                   
err:eventlog:ReportEventW L".NET Runtime version 2.0.50727.42 - Fatal Execution Engine Error (79F97075) (80131506)"                                                                                             
fixme:advapi:DeregisterEventSource (0xcafe4242) stub                                                                                             
trace:seh:raise_exception code=80000003 flags=0 addr=0x7b83cce3 ip=0x7b83cce4 tid=0022                                                                                                                         
trace:seh:raise_exception  eax=00000001 ebx=00000000 ecx=0033e1c4 edx=0033e228 esi=00000000 edi=79f97075                                                                                                       
trace:seh:raise_exception  ebp=0033e1bc esp=0033e1bc cs=0023 ds=002b es=002b fs=0063 gs=006b flags=00200202

Asiding the fact that it's wine debug, I'm hoping this might be of some help
tracking this down....otherwise level_10 is a brick wall for me and I'll never
find out about the secret weapon...

Any clues?, advice?, ...slaps in the face for posting winedebug?  B-)

Cheers!

Don

[annikk.exe]

I can't help with the problem, but can I just say, that is a beautifully crafted bug report..

[Don]

I can't help with the problem, but can I just say, that is a beautifully crafted bug report..

Hi,
      Sure...you can just say that - I'll even say thank you for saying that, how's that!! (;

Essentially, it's a hard topic to broach - one is already 'behind the 8 ball' when it comes
to running into something like this...ie; a piece of software designed to run on windows, and
me using the crossover/wine to approximate the same runtime conditions. This is to say, I
*expect* problems, I expect insurmountable problems at times...but when it becomes peculiar,
is when the wine engine starts tripping the same issue -windows- users are running into....

...I'm going to need do the same thing really...ie; run the demo version, and see if I get any
ooopsing in level_5 like other windows users report experiencing...and if so, that'll give the
tangent a firmer meaning. I do a lot of time helping out over at Codeweaver's website as an
advocate (for this and other game titles), so I'm very used to debugging that wine engine
and trying to get things to work. When it comes to issues like this...and it's by no means the
first time....when the wine engine experiences exactly the same problem a 'real' windows
kernel/environment ...one tries to look for clues or solutions to the problem in the windows
tomes, and then try to transpose those into the wine environment - sometimes it works, a lot
of the time it doesn't. and then sometimes no solution has yet been found....

....to give you an example of just how funky this can get (from the wine angle), is a recent
boofle I had with a Source engine based game ; in this case, it was -much- more specific..ie;
the actual software function call/error message was being hit by windows users and wine engine
users alike. Not all windows users, just a notable number of them. When I went to debug the
issue, in the act of issuing a debug commandline that wrote a logfile to disk, that in itself was
enough to 'fix' the problem - for some strange reason, sniffing debug channels and generating
disk I/O was enough to change the timing of the program's execution/runtime environment enough
to allow it to work properly.....so, right, if that works for the wine engine to get around the same
problem, how does that apply to the windows situation - does it help explain anything - this is
the way drunken penguins think about it, there's something to learn there somewhere...

Eufloria is not so disposed, but it's not being very forthcoming either. I did some more sniffing,
and was watching the dsound channel, and what it shows is what I suspected - it doesn't
actually crash - it hits the error, the sound layer keeps running after the error for a bit (emptying
buffers?)..then it too stops, but the process doesn't crash..it just switches into a catatonic state...
...my *guess* is it's waiting for user input on the crash dialog that the wine engine isn't displaying...

Cheers!

[annikk.exe]

I run Windows XP and experience no problems.  Wine is some kind of emulator I'm guessing?  Like a virtual machine?  Could it be the emulator that needs an update, rather than the game?  </educated guesses>

[Don]

I run Windows XP and experience no problems.  Wine is some kind of emulator I'm guessing?  Like a virtual machine?  Could it be the emulator that needs an update, rather than the game?  </educated guesses>

No,  Wine Is Not an Emulator as we say, nor is it a VM -- wine is a reimplementation
of the windows API in native code that runs on linux/Mac/*BSD operating systems...

I know you're trying to help, but the logic goes like this -- it works fine for you in XP ;
what about those people running XP (in the thread I initially linked to above) for whom it
is *not* working fine and seem to be hitting the same error I am with the wine engine?
What does your educated guess say to these folks?...upgrade to vista/7? Reinstall XP?
My educated guess (30+ years of computing experience) says it's a game glitche that
can affect both winXP installations, as well as the wine engine...

For the sake of thread_subject, I did download the demo version of Eufloria (although I
believe in the post I linked to, they were probably discussing the pre-2.02 demo release).
Regardless, the 2.02 demo release is *much more* fractious to get running (as per what
I've already described with the full 2.02 version being difficult to start at time) -- the demo
is difficult to start all the time -- this also concurs with other windowsXP users' experience.
The demo version also has the same issue as winXP users wrt to the problem with level_5
-- level_5 of the demo does the same "Ooops, I'm outta here.." thang.....I get the same trace
and the same fault/error ..ie  0037:err:seh:setup_exception_record nested exception on signal
stack in thread 0037 eip 7bc6cb00 esp 7ffdbc7c stack 0x242000-0x340000 ..so it would
indeed appear I'm tripping over the same issue the windows users describe...

Cheers!

[Don]

The other thing I should say here I suppose, is that it's not just my interests I have at
heart here, but also those of other wine engine users (not just the crossover wine engine),
and seeing as Eufloria is now a steamapp as well, my guess is the game's exposure to
the world is going to take a little increment - if that is so and there is indeed a problem here
somewhere, the number of people encountering it should probably increment as well...

The only possible pattern I've observed, is that it does tend to be level specific - I enabled
dark_matter mode, and it gets through the first level there no problem - we hit the same error
again on the next level. I've tried disabling sound altogether on the just in case principle (as
I'd noted the sound overrunning the fault trigger)....and I have seen the sdl sound layer have
a spit about something in a 'real' crash situation....cite;

Code: [Select]

Unhandled Exception: System.TypeInitializationException: The type initializer for 'SdlDotNet.Audio.Mixer' threw an exception. --
System.Runtime.InteropServices.SEHException: External component has thrown an exception.
   at Tao.Sdl.SdlMixer.Mix_OpenAudio(Int32 frequency, Int16 format, Int32 channels, Int32 chunksize)                           
   at SdlDotNet.Audio.Mixer.OpenInternal()                                 
   at SdlDotNet.Audio.Mixer.Initialize()                                     
   at SdlDotNet.Audio.Mixer..cctor()                                                                                           
   --- End of inner exception stack trace ---                                                                                 
   at SdlDotNet.Audio.Mixer.Initialize()                                                                                       
   at Eufloria.Game.Go()                                                                                                       
   at Eufloria.Game.Main()                                                                                                     
wine: Unhandled exception 0xe0434f4d at address 0x7b840bb0 (thread 0022), starting debugger...                                                                                                           
trace:seh:start_debugger Starting debugger "winedbg --auto 51 576"         
trace:seh:call_stack_handlers handler at 0x7b8705f0 returned 1               
Unhandled exception: 0xe0434f4d in 32-bit code (0x7b840c39).                                                                   
Register dump:                                                                                                                 
 CS:0023 SS:002b DS:002b ES:002b FS:0063 GS:006b                                                                               
 EIP:7b840c39 ESP:0033e628 EBP:0033e68c EFLAGS:00200212(   - 00      - -IA1)                                                   
 EAX:7b82c311 EBX:7b8ae814 ECX:00000000 EDX:00000004                                                                           
 ESI:0033e6c8 EDI:0033e648                                                                                                     
Stack dump:                                                                                                                   
0x0033e628:  00000036 790c20c8 e0434f4d 00000001                                                                               
0x0033e638:  00000000 7b840bb0 00000001 80131534                                                                               
0x0033e648:  79e7be3f e0434f4d 0033e6c4 790c2000                                                                               
0x0033e658:  02000036 0033e670 79e814da 0033e67c                                                                               
0x0033e668:  02000036 00000001 0033e6ec 79e87ff4                                                                               
0x0033e678:  0000012c 790fabcc 0012d850 0033e6c4                                                                               
Backtrace:                                                                                                                     
=>0 0x7b840c39 RaiseException+0x89() in kernel32 (0x0033e68c)                                                                 
  1 0x79f97065 in mscorwks (+0x127065) (0x0033e6ec)                                                                           
  2 0x7a05b942 in mscorwks (+0x1eb942) (0x0033e728)                                                                           
  3 0x7a05b9eb in mscorwks (+0x1eb9eb) (0x0033e754)                                                                           
  4 0x7a05b9f6 in mscorwks (+0x1eb9f6) (0x0033e848)                                                                           
  5 0x00000400 (0x0000002c)                                                                                                   
  6 0x00000000 (0x00000000)                                                                                                   
0x7b840c39 RaiseException+0x89 in kernel32: movl 0xfffffff4(%ebp),%ebx

That may however be extraneous - typically speaking, I don't see this report
when things mysteriously 'lock up' ...that said, it's always best to mention all
rather than not when a process launches off into limbo_mode (zombie ..ie; maybe
an error is being thrown, but the main process has detached before the error reporting
can complete....sometimes you catch something you don't normally see...

...I'll have to try the Steam version of the demo and see what happens there...

Cheers!

[Don]

The Steam version of the Eufloria demo reliably locks-up in the same fashion, with
the same error trace I'm seeing - I don't have to wait until level_5 to hit it, level_3
does it here for me...

[Alex]

Hey guys. It's a file not found error (clr20r3) I think, seems like the game's searching for a music file it can't find or something. I'll run a stress test overnight and see if it crashes.

[Don]

...I've also been able to recreate the issue reported by other (windows) users wrt
to the dark matter mode lockup...cite:

http://www.dyson-game.com/smf/index.php?topic=456.0

For the record, same level (4)....but it did it on the penultimate asteroid to be conquered,
not the last asteroid (and I thought for just a millisecond I was gunna complete the level
...when suddenly "ding"....seedlings freeze....

I'd like to know what that 'ding' sound is...level complete? (it isn't)...

[annikk.exe]

My educated guess (30+ years of computing experience) says it's a game glitche that can affect both winXP installations, as well as the wine engine...

It's clear you have thought and researched more about this than I have.  Also your 30 years experience owns my paltry 22.  I'll leave you to it :>

[Don]

My educated guess (30+ years of computing experience) says it's a game glitche that can affect both winXP installations, as well as the wine engine...

It's clear you have thought and researched more about this than I have.  Also your 30 years experience owns my paltry 22.  I'll leave you to it :>

Ohh!!...take no offense or nothing.....when you're half a century old, you'll know how I feel  B-)

I've been chasing this for about 6weeks now -- the most of that, is typically making sure the wine
engine isn't primarily to blame ; this is how we tend to feel when using wine and a windows exe
has an oops...ie; it's wine's fault. When the occurrence of software behavior is noted in wine, which
is also reflected in the exact same fashion on the OS the software was designed to run on, you have
to start looking somewhere else.....


For the sake of the thread, I went back and recreated the level_4 dark matter "ding" freeze again, this
time with debug running.  I hit the ding/freeze problem with 4 asteroids to conquer this time around.
Checking thru the (2.8gb small..) logfile, I see no error trace at all like with locking-up thang in other
levels. Indeed, when this happens in dark matter mode, sound seems to keep running indefinitely despite
the visual lockup...and eufloria.exe itself is still responsive (attached to wine) and I don't have to reach
for the smoking gun of killall -9  ....


 

[Don]

Hey guys. It's a file not found error (clr20r3) I think, seems like the game's searching for a music file it can't find or something. I'll run a stress test overnight and see if it crashes.

Hi Alex,

In the trace where I reported ;

err:eventlog:ReportEventW L"clr20r3"


Typically speaking, this is thrown at program start (or failed start which hangs and is thrown a ^C)
Very rarely, this will happen immediately after a seemingly successful startup, when one goes to
load an actual game (maingame)....after loadup, when you click the OK on the level descriptor
GUI, my guess is it throws to start the sound replay, but goes for a nose dive instead. If this error
code is 'file not found' then it must be a bad pointer or it can't get disk I/O when it wants to - as in
a subsequent restart will not hit this error - the file is there...

The dark matter level_4 thing is interesting -- ostensibly, you get the exact same symptoms...seedings
stop moving, no user input...but the trace I consistently get in the main game(s) associated with this
behavior..ie;    err:seh:setup_exception_record nested exception on signal stack in thread   
0032 eip 7bc69360 esp 7ffdbc7c stack 0x242000-0x340000  ...is nowhere to be seen in the dark
matter case.

Cheers!

[Don]

..By doing truly horrendous things with wine/wine-debug flags, I've been able
to continue up to level_6 of dark matter mode.  Basically, this entails loading
the debug flags so thickly, the log reporting actually causes the game to jerk
and stutter during play and things run a little slowly. The logfiles created getting
through levels 4&5 were in the order of some 20gb per level - when I say truly
horrendous, I mean it.....but....it doesn't hit the 'ding'/lockup error like this...

[annikk.exe]

Very well.

I know you're trying to help, but the logic goes like this -- it works fine for you in XP ;
what about those people running XP (in the thread I initially linked to above) for whom it
is *not* working fine and seem to be hitting the same error I am with the wine engine?
What does your educated guess say to these folks?

If one group of people are experiencing the error, and the rest aren't, I usually assume that there is a common link between the computers that do not play the game as expected.

Since you are running the game in Wine and others experience the same error in plain WinXP, it's unlikely to be Wine causing the problem.

It could potentially be something like missing or corrupted .NET components.... Microsoft .NET is needed to run Eufloria, if I recall correctly.  So my next step would be to look at that angle, which (correct me if I'm wrong) you have already tried to no avail.

So failing that, I would then try comparing hardware.  For example, what sort of graphics card do you have, and what driver version, and how does that compare with others with similar problems, etc.  If it turns out all of the people experiencing the problem use the same or very similar graphics cards, it may be related to that particular chipset.


If I was unable to find any common links like those described in the previous paragraph, my next step would be to try to generate error logs to see if something obvious is going on.  I would also be sure to check the Windows Event Viewer to see if any unusual messages appeared at the time of the crash.  As you have said, you were unable to find any errors in the log.



If I reached that stage and could safely say I had exhausted all the options above, at that point I would probably roll a joint and sit and think about the problem for a while until inspiration strikes.  This approach has worked for me a lot in the past, when all the usual procedures have failed.  YMMV.

[Alex]

Interesting stuff, thanks Don, not sure quite what to make of it all