For normal testing I build an application using Xcode and selecting "Build for Profiling" from the Product menu. For production I do "xcodebuild clean build -configuration Release ......." I notice a big performance difference. In my case the XCode profiling build runs in under a minute, the xcodebuild version takes over 4 minutes. The XCode profiling build uses the Release configuration, the xcodebuild is also using the Release configuration. What additional configuration options are being set/used when "Building for Profiling"? I'm having a hard time finding an answer to this question.

The following code sequence works on Catalina and on Intel Big Sur devices, it does not work on the new Apple M1 devices running Big Sur. CFMutableDictionaryRef matching = IOServiceMatching("IODisplayConnect"); io_iterator_t iter = 99; kern_return_t err = IOServiceGetMatchingServices( kIOMasterPortDefault, matching, &iter ); The "matching" variable seems valid: matching CFMutableDictionaryRef 1 key/value pair [0] "IOProviderClass" : "IODisplayConnect" 		key __NSCFConstantString * "IOProviderClass" 		value __NSCFString * "IODisplayConnect" The call to "IOServiceGetMatchingServices" sets the "iter" value to 0. If I use another class name in the "IOServiceMatching" call, say "IOMediaBSDClient", "IOServiceGetMatchingServices" returns a valid iterator. Would Big Sur on Apple M1 devices need additional calls, or has this service class been deprecated? Using "IODisplayConnect" allows me to get display information via "IODisplayCreateInfoDictionary". Suggestions...