View Issue Details
| ID | Project | Category | Date Submitted | Last Update | |
|---|---|---|---|---|---|
| 0005376 | AI War 1 / Classic | Crash/Exception | Dec 21, 2011 6:35 pm | Dec 22, 2011 8:54 am | |
| Reporter | metavital | Assigned To | |||
| Severity | crash | ||||
| Status | closed | Resolution | no change required | ||
| Product Version | 5.020 | ||||
| Summary | 0005376: Game crashes when launched from Steam, but runs successfully when AIWar.exe is run directly | ||||
| Description | The game repeatedly crashes at "GameInit (4 - [Timestamp])" when I launch it from Steam, but when I run "C:\Program Files (x86)\Steam\steamapps\common\ai war fleet command\AIWar.exe" directly it launches successfully. I do not use Kaspersky 2010 but the latest version of Microsoft Security Essentials. I turned off MSE's real-time monitoring and tried again several times but to no avail; so I don't think the antivirus program is the cause for crash. To the best of my knowledge, I do not have a program like D3D Overrider that forces vSync and triple buffering which might crash the game. I have checked the driver settings (ATI Catalyst 11.12), but every relevant option is set at "Use Application Settings". | ||||
| Tags | No tags attached. | ||||
| Internal Weight | New | ||||
|
|
It's actually getting to the 9th step, just not drawing it yet, based on the log you uploaded. So that makes it seem really likely that something is killing it based on it doing an update check, which would make me think of the Microsoft security stuff. I don't know of any problems with their suite, but there's always a first report. My suggestion is to launch the game from outside steam, then go into settings and disable automatic update checks for the game, then try launching it from steam again. My best guess is that it would work, given the log above. |
|
|
No, it still does not work. I had MSE's real-time scan turned off. I launched the game outside Steam, turned off automatic update check, I launched the game within Steam: still crashes at GameInit 4. I changed System Locale and Display Language to U.S. English (since I run a non-English Windows 7 x64), rebooted the system and tried again: still crashes. I reverted back, this time opened up the Task Manager, killed off any potentially interfering processes one by one: including Dropbox, Apple's MobileDeviceService and BonJour, Google's crash handler and Google Talk plugin, Office 2010 IME, PnkBstrA.exe (Valve's PunkBuster, I assume) and whatnot. It still crashes. Given the fact that AIWar.exe only crashes when run within Steam, I suspect it may have something with Steam. But even when I launched AIWar.exe outside Steam despite Steam.exe running at the same time, it ran without a glitch. So there may be something when the exe is run inside Steam browser. I can't really figure it out. |
|
|
I tried to kill any "suspicious" processes that may interfere with AIWar.exe, and tried running it again inside Steam with processes as few as possible (as most of the remaining processes were Windows' internal services or related to AMD's drivers and such). Still crashes inside Steam, runs fine outside Steam, with the same log showing crash at GameInit 4. |
|
|
Hmmm, that is really super bizzare. Closing in on 200k copies of this out there now, over two and a half years, and this is literally the first time I've ever heard this. If you could post an updated output.log that would be helpful -- what you see on the screen has a slight lag behind what is actually happening in those unit steps, because they happen so fast. So most likely it is still dying on step 9, as it was from the first log you posted. When steam runs, it launches a number of threads inside the context of our application. I'm not sure how many it is on windows, but on OSX it was around 20 (OSX counts file handles the same as threads, though, in a lot of respects -- so that's inflated). AI War uses background threads to load assets for the game, meanwhile -- or to do things like asynchronously check for updates. There is a finite number of threads that any process can have, but that varies by OS version and normally even with steam running it's overly we're nowhere near that cap. And given what the first log you posted showed, it was dying on the first spawned thread (which was a call to a website, hence that being the thing that antivirus can catch erroneously). That makes me think that the thread limit really couldn't be the issue, unless some other program besides steam is inserting even more threads into our process. Not sure if any antivirus or antispyware programs do that or not. But obviously you've disabled all that and more. The only thing that really comes to mind at the moment is some experiences I used to have with games 10 years ago when I ran the zone alarm firewall. I would shut off zone alarm and expect my games to work like there was no filtering. And indeed, warnings would no longer pop up. But zone alarm back then actually replaced a driver or something, and so my ports would still be blocked despite the stupid program being disabled. I actually wound up having to run the program, then add an exception for my games, THEN shut it down. I've never used the security software you're using, so I can't even begin to comment on how it works. But it's a thought, and knowing how integrated MS likes to make everything on their OS, it wouldn't surprise me if it's still partly active even while disabled. What to do about that... I am less sure, not knowing the program. It could also be antispyware programs, but it sounds like you thought of that already. I wish I had an easy fix here, but this is really a first if the antivirus isn't the culprit. Then again, the fact that it works when run outside the steam console is already inconsistent with the kaspersky cases, come to think of it, so that makes it even more of a mystery at the moment. What I can say is that, to my knowledge, steam has not changed anything that would cause this (or we'd have a lot of people reporting it by now, given the winter achievement). And we haven't changed anything that would be remotely related to this (indeed, though we've done many beta updates, the last official version that steam delivers is from 11 months back, so really no surprises there). You could try verifying the steam cache of the game files to make sure nothing was corrupted, but I would think it would not work when run as a standalone if that was the problem. So that pretty much leaves some sort of external program messing with the process in some way, is all I can think. Other ideas: You could run steam in offline mode, and see if that helps. If you're feeling really adventurous you could run in safe mode and see if you can launch it that way; fewer background programs get launched that way. You could do some malware scans with something like spybot search and destroy or ad aware, to see if there is something infecting your machine and messing with programs like this. I'm not really suggesting you have malware, but in the interest of brainstorming, those are the ideas that occur to me. You've already hit huge numbers of other likely culprits in your own testing of it. Wish I had something more definitive, but hopefully that helps! |
|
|
Oh my goodness. I found the problem, but I grow more puzzled. It's literally the same problem with Zeraxeal's comments: http://www.arcengames.com/forums/index.php/topic,7429.msg61810.html#msg61810 I also had Microsoft Office 2010 installed, which had Microsoft's IME 2010 (Input Method Editor) along with the main suite. I changed my keyboard settings in the control panel to, uh, what seems like Microsoft's default Korean input configuration in Windows. Then voila, AIWar.exe runs beautifully inside Steam. I did not bother to think of IME as the culprit because: (1) I had already disabled an exe process that governs update of Office IME modules, and (2) I paid several visits to other bulletin boards in which other Korean users seemed to enjoy AIWar.exe launched from Steam without any problems (maybe they did not have Office in their gaming PCs at all, but who knows). What still remains the mystery is how Office IME 2010 seems to interrupt with, say, the way Steamworks handled AIWar.exe, but I guess that's the problem of Microsoft and Valve, not you. I feel sorry for the mess. This is the kind of embarrassment which happens from time to time when one grows up in a 2-byte-language world. Next time, I'll make sure to look up the forums more vigorously (didn't find that thread the first time, sorry). |
|
|
Really glad to hear you got this working. Strange cause indeed! |
|
|
I'd forgotten all about that other thread, as that was the sole other case of this sort of issue. I'm glad you figured it out at any rate, and I'll definitely have to keep that in my mind if someone else runs into this! Thanks for confirming what the fix was, and sorry for all the hassle you ran through for this -- I hope you enjoy the game. :) |
| Date Modified | Username | Field | Change |
|---|---|---|---|
| Dec 21, 2011 6:35 pm | metavital | New Issue | |
| Dec 21, 2011 6:35 pm | metavital | File Added: output_log.txt | |
| Dec 21, 2011 6:37 pm | metavital | Description Updated | |
| Dec 21, 2011 8:57 pm | Chris_McElligottPark | Note Added: 0017920 | |
| Dec 21, 2011 10:04 pm | metavital | Note Added: 0017921 | |
| Dec 21, 2011 10:04 pm | metavital | File Added: remaining_processes.png | |
| Dec 21, 2011 10:07 pm | metavital | Note Added: 0017922 | |
| Dec 21, 2011 11:25 pm | Chris_McElligottPark | Note Added: 0017923 | |
| Dec 22, 2011 4:56 am | metavital | File Added: output_log_updated.txt | |
| Dec 22, 2011 5:37 am | metavital | Note Added: 0017924 | |
| Dec 22, 2011 5:37 am | metavital | Note Edited: 0017924 | |
| Dec 22, 2011 5:38 am | metavital | Note Edited: 0017924 | |
| Dec 22, 2011 6:20 am | tigersfan | Internal Weight | => New |
| Dec 22, 2011 6:20 am | tigersfan | Note Added: 0017925 | |
| Dec 22, 2011 6:20 am | tigersfan | Status | new => closed |
| Dec 22, 2011 6:20 am | tigersfan | Resolution | open => no change required |
| Dec 22, 2011 8:54 am | Chris_McElligottPark | Note Added: 0017926 | |
| Apr 14, 2014 9:28 am | Chris_McElligottPark | Category | Bug - Crash or Exception => Crash/Exception |
