site_archiver@lists.apple.com Delivered-To: darwin-dev@lists.apple.com In addition to the various solutions offered in this thread... 32b -mmacosx-version-min=10.4 64b -mmacosx-version-min=10.5 -- Terry _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev@lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.appl... On Mar 12, 2009, at 6:38 PM, Eric Gouriou wrote: On Mar 12, 2009, at 5:54 AM, Jim Correia wrote: [...] It looks like using posix_spawn solves the problem. (Previously I was using fork/exec because the code was originally written to target 10.4.x.) You can also consider building the various slices of your fat binary with different -mmacos-version-min settings so that, e.g., the 64bit slice will only be running on 10.5 while the 32bit slice is allowed on 10.4 . Adapt to taste. That's the theory at least, please double-check before committing to that solution. This won't work. The only things I look at in the exec code path when grading a binary for preference as to which slice to load is either an explicit preference list from posix_spawn, an override for the sysctl for Rosetta (PPC emulation), or the cpu_type and cpu_subtype in the Mach-o files encapsulated in a fat (Universal) biinary. There's no OS version information in the Mach-o header or the dat container header, so it doesn't matter how you link/compile it. This email sent to site_archiver@lists.apple.com