My game has a mechanism to send in-the-field exceptions to me rather than to Microsoft. Received the following. No additional info known. Dump filename tells me it was Vista, flavor unknown.
FMOD VERSION: 4.14.01
Unhandled exception at 0x1007ab2d in armyofearth_cr_0.518_268938029_OS6.0.6000_589249761209361074.dmp: 0xC0000005: Access violation reading location 0x00e2b000.
Any comment from FMOD? Can you use this address?
- fundictive asked 10 years ago
What version was that? It looks like it is somewhere around the mpeg synth (if using 4.16.00 mapfile is relevant to your version) – are you using mp3 data? If so are you using FMOD_CREATECOMPRESSEDSAMPLE or stream?
There was a change made in the latest versions that fixed seeking into an mpeg stream, tricking fmod’s decoder into thinking random data was actually an mpeg frame header, then decoding rubbish which caused this sort of problem.
Have you only had those 2 reports from the 10th to the 25th?
So, I deployed the FMOD_SYSTEM_CALLBACKTYPE_MEMORYALLOCATIONFAILED callback with FMOD 4.14.04 and my first new-version dmp is:
Unhandled exception at 0x1007b02c in armyofearth_cr_0.519_268939308_OS5.1.2600_312038801211077290.dmp: 0xC0000005: Access violation reading location 0x003789ec.
I.e. not OOM.
Can you help?
These last 2 crashes (0x1004f213 and 0x1007e0f3) are both 4.16.00 and are the only crashes I’ve had on that DLL; so it appears to be more stable than 4.14.04.
My game plays live music via a single createSound call with FMOD_CREATESTREAM | FMOD_NONBLOCKING | FMOD_SOFTWARE. This would be the only source of MP3 playback.
My next version will go out in 2-3 weeks. I’ll deploy the latest stable fmodex again and report back.
Another crash today on 4.16.00 at 1004be82.
Unhandled exception at 0x1004be82 in armyofearth_cr_0.520_268746370_OS5.1.2600_348835441214434896.dmp: 0xC0000005: Access violation writing location 0x3d86cf6c.
Another minidump received today at the same address:
Unhandled exception at 0x1007ab2d in armyofearth_cr_0.518_268938029_OS5.1.2600_146034881210314132.dmp: 0xC0000005: Access violation reading location 0x00373000.
This time it is XP.
FMOD can you comment please?
Army of Earth
Now on 4.16.04. Probably the most stable yet.
However got this user crash:
Unhandled exception at 0x10021b47 in armyofearth_cr_0.521_268573511_OS5.1.2600_228823081215831640.dmp: 0xC0000005: Access violation reading location 0x013f1000.
did the change to the latest version show any effect? We’re having some strange and seemingly random crash within the fmodex.dll as well and I’m still trying to find a way to reproduce it in order to be able to say whether the new version fixes it…
[Frames below may be incorrect and/or missing, no symbols loaded for fmodex.dll]
I am also getting random crashes in fmod recently. The sound seems to be fine, then it just crashes, i’m not sure how to debug to be honest. The dll version is : – 0.4.16.3
Some background: –
I am doing a racing engine – so i have a few wav samples that I change the frequency and volume of according to revs. I combined 2 wavs that were recorded at the RPM nearest to the current RPM. I also have tyre noise depending on the slipratio/angle of the tyre.
Hope this helps debug the problem, if i can provide any more info, please let me know.
All the best,
Unhandled exception at 0x1007aaf3 in armyofearth_cr_0.518_268937971_OS5.1.2600_350443201210392456.dmp: 0xC0000005: Access violation reading location 0x015b1000.
These are all separate computers.
Can I confirm that you are not doing non thread safe fmod calls, ie calling fmod from different threads? Overriding fmod callbacks without having threadsafe functionality in your callback code? Maybe you’re running out of memory? (system out of memory callback can catch that).
No, the game itself uses 1 thread only.
I overload the new operator and generate my own exception for failure, but haven’t seen that come up.
Can you use the addresses to determine the functions that it is crashing in?
Then you are in a better position than I – I’ve never seen it myself and wouldn’t know about the crashes except that I have a mechanism to receive crash dumps from the field.
All the callstacks I received were shallow, i.e. no application calls show.
I suggest you soak with the new version and follow up with the crash address(es) if you get any more…
Please login first to submit.