I found and fixed the problem in a 3rd party library (not in URL Access
Manager). I didn't see it at first because the debug version of the library
didn't have Traceback Tables enabled. The library was repeatedly creating
UPP's (ThreadEntryUPP, ThreadSwitchUPP, ThreadTerminationUPP) instead of
reusing a statically initialized UPP.
These UPP's which consume only 32 bytes on Power Macs, somehow consume much
more on an Intel Mac.
The problem remains that there are no memory debugging tools that work for
PowerPC apps on Intel Macs. MallocDebug.app doesn't work and setting some
Malloc Debug Environment Variables in the Terminal will cause LaunchCFMApp
to immediately or eventually return a segmentation fault.
> From: Joe van Tunen <email@hidden>
> Date: Tue, 13 Mar 2007 15:41:45 -0700
>
> I have a PowerPC CFM application (not a bundle - just a single file with a
> resource fork and a data fork) which does not leak on PowerPC Macs but will
> crash after leaking 3.49 GB of virtual memory over several hours on an Intel
> Mac.
>
> Sampler.app from the Developer Tools Performance Tools only shows a few K
> being leaked by some URL Access Manager calls but BigTop.app shows that my
> app is leaking about 5 MB per minute. I am unable to get MallocDebug.app to
> work with my app on an Intel Mac.
>
> Is there a way to make MallocDebug.app work with my app on an Intel Mac?
>
> What else should I try?
>
****************************************************************************
This email may contain confidential material.
If you were not an intended recipient,
Please notify the sender and delete all copies.
We may monitor email to and from our network.
****************************************************************************
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Carbon-dev mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
http://lists.apple.com/mailman/options/carbon-dev/email@hidden
This email sent to email@hidden
