GuardMalloc vs. auval
GuardMalloc vs. auval
- Subject: GuardMalloc vs. auval
- From: Daniel Jalkut <email@hidden>
- Date: Sun, 14 May 2006 12:11:14 -0400
I like to run my code through GuardMalloc whenever possible, since it
helps me catch some sneaky bugs before shipping. I'm currently
putting my AudioUnit through the paces and am running up against some
stumbling blocks that prevent me from clarifying whether there are
problems in my code or not.
When I debug the "auval" executable from Xcode with "Guard Malloc"
enabled, I seem to invariably run into problems, even if I'm not
targeting my Audio Unit.
As a control case, I debugged auval running with the following
parameters:
-t aufx appl bpas
With GuardMalloc enabled, this consistently crashes in Apple's AU
code, descended from AudioUnitGetProperty.
When I do the same thing targeting my own AudioUnit, I observe a
crash which traces back to what appears to be a bad "outData" pointer
in the params for a host callbacks request.
I have a hunch that auval has some bugs in its params initialization
that show up in Guard Malloc mode, making it impossible to use it as
a test case for finding such bugs in my own code. Does this seem
like a reasonable conclusion? I will report a bug unless somebody has
an obvious reason for why I shouldn't expect this to work. For
instance, does CoreAudio use AltiVec to an extent that I should not
count on Guard Malloc working?
If by any chance this is a known issue and an updated version of
auval is available, I would really appreciate it if somebody would
send me a copy. Thanks a lot for your help!
Daniel
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Coreaudio-api mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden