[Mono-osx] Building mono-1.1.18 20061018
allan at counterpop.net
Tue Oct 31 02:48:32 EST 2006
I already sent in the following patch last week to the mono-devel
list, as well as libgc upstream. I added a #define to gc_priv.h
called GC_MACH_THREAD_STATE_FLAVOR and replaced usages of
MACHINE_THREAD_STATE with it. Apologies for not CCing this list.
On Oct 30, 2006, at 10:46 PM, Diane Trout wrote:
> On Oct 18, 2006, at 17:18, Allan Hsu wrote:
>> One thing to note is that current Mono SVN will not build properly
>> under XCode 2.4 for Intel due to changes to the value of
>> MACHINE_THREAD_STATE in the new 10.4u SDK. Mono will compile under
>> 2.4, but the resulting binaries will not work properly. The
>> headers that come with XCode 2.3 should work fine.
> Ok I spent some time trying to figure out what was going on with this.
> It appears that in MachOSX10.4u.sdk/usr/include/mach/i386/
> MACHINE_THREAD_STATE is defined to be x86_THREAD_STATE (7) which
> seems strangely unrelated to x86_THREAD_STATE32 (1) or
> x86_THREAD_STATE64 (4). I tried running mono with the
> MACHINE_THREAD_STATE set to 5 but that didn't seem to work.
> I made an alternate patch that is only triggered on darwin/x86. I
> hope its a bit closer to being acceptable for SVN. Basically in
> mono/libgc/include/private/darwin_stop_world.h I set
> MACHINE_THREAD_STATE to x86_THREAD_STATE32 if x86_THREAD_STATE32
> and __i386__ were defined.
> Hopefully the test for x86_THREAD_STATE32 will make the include a
> non-op on the 10.3.9 SDK
> I tested the patch on intel/os x by managing to finish my build
> without hanging and by running the ironpython interpreter.
> Mono-osx mailing list
> Mono-osx at lists.ximian.com
Allan Hsu <allan at counterpop dot net>
1E64 E20F 34D9 CBA7 1300 1457 AC37 CBBB 0E92 C779
More information about the Mono-osx