[patch] boehm-gc darwin backport 4.2.*

[Andreas Tobler]

Mike Stump wrote:
> On May 18, 2007, at 11:29 AM, Andreas Tobler wrote:
>> upon request I backported r120684, 120801, 120853, 120874 and 120977 
>> from main to 4.2.1.
>>
>> Ok to apply once my build & test finished?
> 
> As I read it, the code only consists of darwin bits...  If I'm wrong, 
> the below only applies to the darwin only bits.

Yep, only darwin bits.

> 
> Ok, but one thing to think about....
> 
>> Index: include/private/gcconfig.h
>> ===================================================================
>> --- include/private/gcconfig.h    (revision 124828)
>> +++ include/private/gcconfig.h    (working copy)
>> @@ -791,7 +794,7 @@
>>  #     define DATAEND (_end)
>>  #   endif
>>  #   ifdef DARWIN
>> -#     if defined(__ppc64__) || defined(__x86_64)
>> +#     if defined(__ppc64__)
>>  #       define ALIGNMENT 8
>>  #       define CPP_WORDSZ 64
>>  #     else
> 
> You meant to do this, as some other code does what needs doing for 
> x86_64, right?  This is mainly a rhetorical question as I suspect the 
> answer is yes, but I have to ask anyway.

Yes, it's in here:

# if defined(POWERPC)
#   define MACH_TYPE "POWERPC"
...

So I do not need to check for _x86_64, or do I?

> I'll try and build and fix any darwin9 building problems after you check 
> it in.  Since it is 4.2 we're talking about, please try and watch for 
> any fallout from this...

I do, building and testing on ppc/i?86-darwin right now.

Thanks.

Andreas