On Fri, Nov 11, 2011 at 5:25 PM, Greg Parker <gpar...@apple.com> wrote: > Did you actually write the minimal test you described? If you had, you would > have discovered that setrlimit(RLIMIT_DATA) does not limit malloc memory. > > That's one reason why we ask bug reports to include test cases.
To call setrlimit() was not the cause of the Simulator bug, but my suggested fix for it. My reading of the setrlimit() man page suggested it would fix the bug, not allow it to be reproduced. OK, I'll take your bet and raise you: Have the Simulator instead allocate a single fixed-sized buffer that is more or less the size of the virtual memory that would be available to Apps on real devices, then have the Simulator's malloc() pass out chunks of just that one buffer, without ever asking Mac OS X for more virtual memory. > armv6 or armv7? There's a known LLVM 3.0 codegen bug for armv6 Thumb. > Workaround is to > disable Thumb. Both of the devices I reproduced it with have ARM Cortex A8 CPUs, so armv7 - the first-generation iPad and iPhone 4. I also have a first-gen iPhone, which is armv6, but have not yet tested it. I did come across some reports that Thumb for armv6 is broken, and figured that even though this was armv7, but I have not yet been able to figure out how to disable Thumb in Apple's LLVM 3.0 compiler for Xcode 4.2. The way one does it for GCC has no effect on LLVM. If you can enlighten me, I'll give it a try right after I get back from supper. -- Don Quixote de la Mancha Dulcinea Technologies Corporation Software of Elegance and Beauty http://www.dulcineatech.com quix...@dulcineatech.com _______________________________________________ Cocoa-dev mailing list (Cocoa-dev@lists.apple.com) Please do not post admin requests or moderator comments to the list. Contact the moderators at cocoa-dev-admins(at)lists.apple.com Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/cocoa-dev/archive%40mail-archive.com This email sent to arch...@mail-archive.com
