[reSIProcate] Access Violation in Data.cxx

[Jeremy Geras]

Turns out I can repro this with basicCall...  Just created a batch file
that keeps running basicCall.exe, and after 10 or 15 tries, I got the
error...

 

________________________________

From: resiprocate-devel-bounces at list.resiprocate.org
[mailto:resiprocate-devel-bounces at list.resiprocate.org] On Behalf Of
Jeremy Geras
Sent: Thursday, November 16, 2006 2:53 PM
To: Scott Godin; Byron Campen
Cc: resiprocate-devel at list.sipfoundry.org
Subject: Re: [reSIProcate] Access Violation in Data.cxx

 

Not at the moment, but I'll work on it...

 

Using MS Visual Studio 2003 SP1 and .NET 1.1 SP1.

 

________________________________

From: Scott Godin [mailto:slgodin at icescape.com] 
Sent: Thursday, November 16, 2006 2:30 PM
To: Byron Campen; Jeremy Geras
Cc: resiprocate-devel at list.sipfoundry.org
Subject: RE: [reSIProcate] Access Violation in Data.cxx

 

Do you have test program that can post that will duplicate this error?

 

From: resiprocate-devel-bounces at list.resiprocate.org
[mailto:resiprocate-devel-bounces at list.resiprocate.org] On Behalf Of
Byron Campen
Sent: Thursday, November 16, 2006 5:25 PM
To: Jeremy Geras
Cc: resiprocate-devel at list.sipfoundry.org
Subject: Re: [reSIProcate] Access Violation in Data.cxx

 

          So ApiCheckList didn't complain about anything? Hmm. I would
ask what version of Visual Studio or .NET you're using, but frankly I
know very little about working with resip on Windows. Scott Godin might
have a better idea of what you're dealing with here. The members of the
Data seem ok; is this from a crash?

 

Best regards,

Byron Campen

 

Including ApiCheckList.hxx didn't help - same problem.

 

I've attached a screen shot showing the QuickWatch of the tid variable
at TransActionMap::add

 

I'm not 100% sure any more that reverting 6504 completely fixes the
problem (though it does seem to make it happen a lot less frequently)...
I.e. I've experienced a crash with that revision reverted, but
unfortunately I wasn't attached to the debugger so I'm not sure it was
the same crash.

 

________________________________

From: Byron Campen [mailto:bcampen at estacado.net] 
Sent: Thursday, November 16, 2006 12:26 PM
To: Jeremy Geras
Cc: resiprocate-devel at list.sipfoundry.org
Subject: Re: [reSIProcate] Access Violation in Data.cxx

 

            Ok, after looking at the stack trace in your mail, it
appears that stuff is hosed well before we get into Data::rawHash()
(unless 0xfeeefeee is a valid pointer, and you have a transaction id
that is 427075694 characters long, both of which are doubtful). It is
possible you're running into one of the known API/ABI mismatch issues
that pop up when you compile against the resip headers with different
flags than the libs were compiled with. Try including
resip/stack/ApiCheckList.hxx in whatever app you're building, and see if
anything pops up. If that doesn't solve anything, let's have a closer
look at what state the members of that Data are in when things go south.

 

Best regards,

Byron Campen

Hi,

 

I'm using the latest revision from the SVN head, and I've got an access
violation occurring consistently in Data.cxx.  I've attached a screen
shot of the debugger at the point of the crash.

 

All I do to generate the crash is make/answer/hang up a few calls.

 

As you'll see in the screen shot, I'm using Windows.  I'm also using
DUM.

 

I didn't have this problem until I updated a day or so ago...  I think
prior to this update I was using a pretty old revision (probably from
late August).

 

>From the svn log it looks like there were some signed/unsigned changes
to the hash stuff in Data...  wonder if that could be the cause?

 

Thanks,

 - Jeremy -

<resip_crash.PNG>

_______________________________________________

resiprocate-devel mailing list

resiprocate-devel at list.sipfoundry.org

https://list.sipfoundry.org/mailman/listinfo/resiprocate-devel

 

<tid_quickwatch.PNG>

 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://list.resiprocate.org/pipermail/resiprocate-devel/attachments/20061116/f28afff7/attachment.htm>