Team Fortress 2 Forum

I added +file, ran, tried to join a server, and uploaded.

In case it helps, the error I got was:

CAudioSourceMemWave(player/pl_scout_dodge_can_open.wav):
GetDataPointer() failed.

for some reason that uploaded file was corrupted and I cannot gunzip it.

try this instead:

cxgames/bin/wine --bottle <bottlename> --cx-log ~/tf2crash.log --debugmsg +wave,+tid,+seh,+winmm,+file --cx-app steam

then
gzip ~/tf2crash.log

maybe the gziping part of the logging causes problems.

-aric

I re-ran and gzipped after - did a quick zcat to make sure that it could be read. Let me know if you're still having problems.

Matt

thanks,

ok nothing obvious... Could you try this?

cxgames/bin/wine --bottle <bottlename> --cx-log ~/tf2crash-relay.log --debugmsg +wave,+tid,+seh,+winmm,+relay --cx-app steam

The resulting log will be big, very big.

hopefully gziped it will be small enough to still be able to be upload (we have about a 100mb limit for uploads) If it is too large then I can get you ftp access to upload to.

Also if you could tell me the text of the error message when it crashes that would be helpful.

I tried that this morning - I'm guessing it was the extensive logging, but everything was going really slowly, and I wasn't able to get a game started before I had to leave. I'll run that tonight, and get back to you.

Matt

I tried again this morning - it waited forever at the main TF2 loading screen, and by the time I quit and exited out, the log file was 8GB or so. Not sure how manageable the filesize will be when I'm able to get TF2 to actually start a game.

Matt

Does this problem also occur when you start your own server? In that case you can probably try my trick to start TF2 manually, in which case you can avoid logging Steam, which probably produces most of the lines in the log.

I've got the same problem as everyone else, is there a fix yet?

Stefan Dösinger wrote:

Does this problem also occur when you start your own server? In that
case you can probably try my trick to start TF2 manually, in which
case you can avoid logging Steam, which probably produces most of
the lines in the log.

I can confirm that this does happen when creating a server, I just got this error on a fresh install of TF2 using Crossover.

Corey Quilliam wrote:

Well, for me, I solved the issue by completely replacing the Steam
directory created by installing it using Crossover with the Steam
directory from my actual Windows install.

This leads me to believe that Wine is somehow corrupting these wav
files or something along those lines. I don't think its sound
driver related at all.

Note: Even creating a backup of my TF2 install in windows and
restoring it in a Crossover-installed Steam does not solve the
issue. I have to completely replace the Steam directory in
~/.cxgames with my Steam directory from Windows and all is well.

BIG THANKS!!! The above worked for me after many hours of messing about with Pulseaudio:

I was getting the CAudioSourceMemWave (xxx): GetDataPointer() failed error everytime it tried to load a map.

Pulseaudio removal/purge/pasuspender didn't solve it for me. The error kept occuring with only ALSA installed and selected.

Disabling sound completely in wine allowed me to run the game. At least it worked but with no sound!!

Then I stumbled across your thread on codeweavers where the above fix worked for me.

Ubuntu Jaunty 9.04 x64
Wine 1.127
OpenGL 2.1.8870
ATI 4890 Catalyst 9.8

The error occured over & over again, no matter what, fresh OS installs without Pulseaudio, new Steam installs, re-downloading TF2 from Steam.

I happened to have a backup of TF2 taken using the Steam backup app from my Vista 64 install taken just a day ago. The error occured again after restoring this backup into a fresh copy of Steam when trying to load a map.

The only thing that's worked for me is renaming the Steam directory in wine to Steam-bak, then copying the Steam directory from my windows installation to wine. Works like a charm.

I still have the old directory that causes the error, so let me know if want any log files from it and I'll see what I can do, as I'm not actually using crossover software.

Corey Quilliam wrote:

Well, for me, I solved the issue by completely replacing the Steam
directory created by installing it using Crossover with the Steam
directory from my actual Windows install.

This leads me to believe that Wine is somehow corrupting these wav
files or something along those lines. I don't think its sound
driver related at all.

Note: Even creating a backup of my TF2 install in windows and
restoring it in a Crossover-installed Steam does not solve the
issue. I have to completely replace the Steam directory in
~/.cxgames with my Steam directory from Windows and all is well.

BIG THANKS!!! The above worked for me after many hours of messing about with Pulseaudio:

I was getting the CAudioSourceMemWave (xxx): GetDataPointer() failed error everytime it tried to load a map.

Pulseaudio removal/purge/pasuspender didn't solve it for me. The error kept occuring with only ALSA installed and selected.

Disabling sound completely in wine allowed me to run the game. At least it worked but with no sound!!

Then I stumbled across your thread on codeweavers where the above fix worked for me.

Ubuntu Jaunty 9.04 x64
Wine 1.127
OpenGL 2.1.8870
ATI 4890 Catalyst 9.8

The error occured over & over again, no matter what, fresh OS installs without Pulseaudio, new Steam installs, re-downloading TF2 from Steam.

I happened to have a backup of TF2 taken using the Steam backup app from my Vista 64 install taken just a day ago. The error occured again after restoring this backup into a fresh copy of Steam when trying to load a map.

The only thing that's worked for me is renaming the Steam directory in wine to Steam-bak, then copying the Steam directory from my windows installation to wine. Works like a charm.

I still have the old directory that causes the error, so let me know if want any log files from it and I'll see what I can do, as I'm not actually using crossover software.

I would also like to add that replacing the directory with my windows directory worked like a charm...

this is still an unacceptable solution however as many of us do not have any window directories available!

I'm just wondering, replacing it with a Windows directory doesn't fix showing Steam messages in-game, does it?

nope ..Crossover or wine would be unable to run the code that steam injects for the ingame messages.

Oh, right. :(

-bump-<<

Oh goody, I was just able to trigger this error with the following title ;

http://www.codeweavers.com/compatibility/browse/name/?app_id=6930

Being a HL-2 mod, my guess is we're looking at the same problem..

I don't use pulseaudio -at all-, never been installed, never will be installed,
I just use the native alsa drivers...so pulseaudio may be a problem in some
places, but you don't need to have pulseaudio around to hit this error.

I've only just discovered this - I'll do some logging/triage later and
see if I discover anything of interest...

Cheers!

Here kitty kitty kitty....

,,,,no, I haven't lost the plot - I just got around to looking at this issue.
I've got a wonderful pattern of behavior happening with this and D.I.P.R.I.P

~/cxgames/bin/wine --bottle steam-mixed --cx-log ~/diprip_crash2.log --debugmsg +wave,+tid,+seh,+winmm,+file --cx-app steam

If I launch Steam like that, and then launch D.I.P.R.I.P from within the Steam GUI,
the CAudioSourceMemWave (xxx): GetDataPointer() failed doesn't happen and the game
goes on to start correctly...

~/cxgames/bin/wine --bottle steam-mixed --cx-log ~/diprip_crash_no_channels.log --cx-app steam

If I launch Steam like this, D.I.P.R.I.P crashes every time -- sometimes it crashes without
hitting the CAudioSourceMemWave (xxx): GetDataPointer() error message and dies quietly, the rest
of the time it -does- hit the error and dies as I noted before...

..slightly dazed and confused, I continue on ....

Cheers!

Hi everyone ,
i have been one of the fortunate/unfortunate few who have yet to experience this error..

However , i would like to try and help
First of by trying to recreate the error..(on TF2 and Diprip ..or what ever hl2 mod games i can get my hand on:)

Reading the thread ... some people commented that by replacing the steam directory in the bottle from the Windows directory might solve the problem.
However , i find it rather unlikely that the problem is really 'Fixed ' or if there was a problem to start with.
This error most likely occurs from the HL2 mod game being unable to open a sound file it requires..
This could mean a few things...But the most possible and direct idea is that the sound file that was required is corrupted..
Therefore , replacing the bottle steam directory with the windows steam directory worked as the corrupted file is replaced...with an uncorrupted one.
That would perhaps be case scenario 1.

Another case , would most likely be pulseaudio causing it ...However , with Pulseaudio removed , some users still reported that the error still persist...That could mean two things...Either pulseaudio was not cleanly uninstalled ..or there are other factors causing the problem..
try locating for pulseaudio libs if you suspect if its uninstalled not cleanly..This is so as Pulseaudio hogs the sound streams and does not give Cxgames the right to use it ...

Third case , would be the audio driver causing the error ...either that or Cxgames /wine is unable to interact/utilize with the audio driver..
Therefore , this could be a rather generic case...If a sound file was corrupted ... the sound clip would be unable to be played.
Therefore , resulting in the error ..It also could be a buggy sound driver..leading to Cxgames being unable to communicate with the Audio driver and sound card..
I am not to sure how extensive the current list of supported sound cards that Linux currently is..but..
In any case , If you are using linux , make sure ALSA / OSS ( which ever 1 you are using ...try configuring it via sudo alsaconf) is properly configured. After that you could try and use winecfg to configure and test the sound.

Another thing i noticed ...when during sound intensive scenes..or repetitive sounds ( like a guy burning in CSS)
The game would crash too...with that error..( in the log )
Lol...i am not too sure bout this..

@ Don : do you notice which /what sound file was mentioned in the error ? ...Or was there no sound file .
Apart from that .. Don ..would it be possible if i could have a package list of your installed software/apps ?
I would like to compare and see the difference ...and to try and isolate the package(s) that is/are causing the error.
And at the diprip error ...do you think that by giving Cxgames too much proccess / commands to handle ...Cxgames would fail to work ? :o
( stupid idea lol..)

oh yes.. i would be trying drprip out..:)
(hmm...evidently 6 hours of sleep per day is bad for my health...cough cough...sniff sniffle...:( )

cheers :D

Sam! wrote:

@ Don : do you notice which /what sound file was mentioned in the
error ? ...Or was there no sound file .
Apart from that .. Don ..would it be possible if i could have a
package list of your installed software/apps ?
I would like to compare and see the difference ...and to try and
isolate the package(s) that is/are causing the error.
And at the diprip error ...do you think that by giving Cxgames too
much proccess / commands to handle ...Cxgames would fail to work ?
:o
( stupid idea lol..)

oh yes.. i would be trying drprip out..:)
(hmm...evidently 6 hours of sleep per day is bad for my
health...cough cough...sniff sniffle...:( )

cheers :D

...goodgrief man, sleep gives you cancer...everyone knows that...<grin>...

Well, now that diprip is up and running, it hasn't encountered this error again...yet.
It's hard to -really- tell when you live in .AU and the only servers found are 400+ms
away (disconnect threshold) but the one I did have a run on (at around 380ms lag) the
game seemed stable enough. I have sent a more extensive blather on the subject to info@
so they know this has resurfaced, and I've got enough debug logs here to build a house...

Regarding your question...it would/could be various files...but I don't think the files
are the issue. I'm also starting to question whether it has anything to do with audio
hardware/drivers either...not going by what I just went through here with this...

...I think it has to do with the creation of these files...;

~/.cxgames/[bottle_name]/drive_c/Program Files/Steam/steamapps/[steamusername]/diprip warm up/diprip/maps/soundcache/

The game engine (Source) is actually -processing- the audio files at
the time of the error - it's not the file going wrong, it's the process.
This crash will usually end up creating a .mdmp crashlog...utter binary
jibberish for the most part, but what I think I see there is the something
to do with files under the above directory, and/or something going awry when
the cache creation process passes the handle onto the main exe...something
like this....

By "package list of your installed software/apps"...what exactly do you mean?
My Steam bottle makeup?...or?...

Cheers!

oops...sorry bout that..
( i meant the lack of sleep is bad for my health :S ...>.< )

hm...oh ya i was referring to the stuff installed in your linux system :)
i dont really think it has anything much to do with conflicting software /services...but its always nice to see and check :)

Sam! wrote:

oops...sorry bout that..
( i meant the lack of sleep is bad for my health :S ...>.< )

hm...oh ya i was referring to the stuff installed in your linux
system :)
i dont really think it has anything much to do with conflicting
software /services...but its always nice to see and check :)

Well...you're out of luck wrt to this linux instance, because it doesn't
use a package manager (or packages)...a Debian instance (on this same hardware)
and another Debian installation in the house are Debian 5.0.4 + updates...

Now having seen this issue 'first hand', and in consideration of all those
other (Windows) users out there hitting the same issue, it really starts to
look like a Source engine glitche...

Cheers!

Just thought I would pass along my findings for this particular issue. I know most of the problems reported pertain to the Linux version of Crossover, but I was able to "resolve" this under Mac OS X and wanted to share for anyone else using Mac.

I installed Crossover Games, Steam and TF2 clean from scratch and received the CAudioSourceMemWave error before joining or starting a game after all the loading completed. I performed the following:

• Clicked Configure -> Manage Bottles
• Clicked the Control Panel tab
• Opened the winecfg option to enter the wine configuration
• Opened the Audio tab
  • At this point it warned me that there was no audio driver configured and that it would use a default. A ha!
• I clicked Apply after the default was chosen for me (CoreAudio) and tested by clicking the Test button. So far so good.
• Loaded up TF2. Created a server to test and was able to join the game with no crashing.

Hope this helps someone.

CaptainPoopsock wrote:

Just thought I would pass along my findings for this particular
issue. I know most of the problems reported pertain to the Linux
version of Crossover, but I was able to "resolve" this under Mac OS
X and wanted to share for anyone else using Mac.

I installed Crossover Games, Steam and TF2 clean from scratch and
received the CAudioSourceMemWave error before joining or starting a
game after all the loading completed. I performed the following:

• Clicked Configure -> Manage Bottles
• Clicked the Control Panel tab
• Opened the winecfg option to enter the wine configuration
• Opened the Audio tab
  • At this point it warned me that there was no audio driver
    configured and that it would use a default. A ha!
• I clicked Apply after the default was chosen for me (CoreAudio)
  and tested by clicking the Test button. So far so good.
• Loaded up TF2. Created a server to test and was able to join the
  game with no crashing.

Hope this helps someone.

Hi,
Yeah, this is the idea here I think...ie; report/record "what worked for you" on your
platform, so that others can try the same idea. When you look at this issue on a broad,
global basis..ie; stuff the search string "CAudioSourceMemWave : GetDataPointer() failed"
into your favorite search-engine and then spend an hour or two reading all about it (and
then from the Windows users angle), you cannot see a clean and consistent 'fix' that works
for all Windows users/computers. I am of the mind that I would rather chew broken glass
than actually use Windows (or files copied from same environment) to fix -any- crossover/wine
problem -- if we can discover effective means (plural, multiple solutions) to try and resolve
this natively using crossover/wine/OS of your choice, I'd rather head in that direction, n'est par?

I received an email back from Balfour - he agrees with my own suspicions that this is a
critical timing issue wrt to what the software is doing on the hardware involved. My example,
wherein DIPRIP 'magically' got over it, once (and only if) a --debugmsg +bla,+blah,+blablah
commandline was thrown at it, is a very good example of this apparent timing critical operation
being at play -- imagine if you will, the debug commandline generates extra disk I/O interrupts,
and this will change the timing of when a process is actually able to access disk data...

If you read enough winehq, you'll come across more than a few cases, where that's actually
the way people get an app/game running in mainstream wine -- prepend the commandline with
WINEDEBUG=+bla,+blah,+blablah and the title runs ; if you don't do this (normal command
or desktop/menu shortcut), the title reliably crashes...

If you look at the example of your solution, you'd be forgiven for thinking "Ah-HA! So it
is audio related" ....but I don't think so. Normally speaking, you don't have to set
the Audio tab/winecfg stuff...I think normally speaking, if that configuration is untouched,
wine goes looking for what it can use and decides for you. If however, you set that winecfg
configuration section, there is no need to go look for what audio driver wine is going to
use -- this has already been decided by you setting the winecfg configuration here. I'll
hedge a bet, this also affects the relative timings of the game engine accessing the audio
device...it'll only be a fraction of a moment, but sometimes that's all it takes...

This, however, is the point of it all...your solution, my solution, some one else's solution,
are likely to work in some cases and not others, just depending of the relative timings in
each machine instance..so the more ideas people have to try, the more likely it is they'll
find a solution that works in their situation...

Here's a quote from Balfour's reply that draws a picture about this;

"I've seen this behavior before, where debugging magically cures the problem. Talking to the devs they
call those 'timing bugs', where by turning on debugging causes the program to run its operations at a
different speed, and works around the issue. Something is probably trying to load or execute before other
parts that it relies on are ready to go."

Cheers!

I've been having a bit of a HL-2 modfest of late - I stumbled across this;

"
Q: When I launch the game the first time, it takes a couple of minutes compiling something. What's that?
A: In order to keep the installer file size as low as possible, we didn't include the sound cache files.
Instead, all you get is a bunch of text files to tell the engine how to create them the best way possible
so to avoid the dang stuttering problem. For further information, refer to Valve Developer Community.
"

cite: http://uoz.multiplayer.it/ra/index.php?content=about

If you follow the the link on that page to the Valve Dev wiki, you end up here..;

http://developer.valvesoftware.com/wiki/Soundcache

...going by that description, if anything goes wrong with the soundcache file generation,
it might end up throwing a handle to the game engine process about a file location on disc
(the .wav file) that isn't where it's supposed to be (corrupted data in the cachefile perhaps).
Furthering the description, it seems like you can head into salt-water crocodile territory
and try futzing with the soundcache files and perhaps break it - theory is, deleting the
soundcache files outside should invoke the generation process...