On 6/1/11 11:44 AM, asiga wrote: > Alexandre Julliard wrote: >> "asiga" <wineforum-user@xxxxxxxxxx> writes: >> >> >>> Hi, >>> >>> I know Wine64 is still experimental, and I've read threads from last Summer stating that Wine64 won't build nor run on OSX at this time. >>> >>> However, Wine64 would be really useful for me on OSX, because I build small simple apps for both win32 and win64 mode, and being able to crosscompile and run them from OSX would be a great time saver for me. >>> >>> So, what's the status of Wine64 on OSX? Will it run soon? Is there somebody working at it? >>> >> >> It's not going to happen anytime soon, if ever. It requires support in >> the kernel and the compiler, and there's no evidence that Apple is going >> to do this. There are also some tricky ABI incompatibilities. >> >> -- >> Alexandre Julliard >> julliard@xxxxxxxxxx > > Well, I don't think those are the real causes, as it's possible to virtualize 64 bit OSs on OSX. I know WINE is not a virtualized OS, but a binary compatibility layer, but, however, I'm pretty sure that if Wine was a Mac-centered project, Wine64 would build and run. Note that I'm not blaming anybody, I'm just saying that the real cause is not the OSX kernel, but the lack of user interest in Wine64 for Mac. And I say it *is* a lack of kernel support. I would know, because I looked into this. These are the exact problems that prevent Wine64 from running on Mac OS X: - There is no way to set the GS base MSR. Wine needs it for the TEB. - Speaking of GS, it's already being used by the System Library for TLS--a direct violation of the System V ABI (to which Apple claims Mac OS X conforms). - Clang (which is already beginning to replace GCC as the compiler of choice on Mac) does not yet support the ms_abi and sysv_abi attributes. I filed http://llvm.org/bugs/show_bug.cgi?id=9148 to keep track of that. - Apple GCC is stuck at version 4.2.1, because Apple fears/hates GPLv3. So Apple GCC is no longer supported, and will thus never support the ms_abi and sysv_abi attributes. I filed Radar 9269783 (note: Apple radar bugs aren't public), and Apple has not responded yet. I don't think they're interested in this right now. Chip
