[Mono-dev] Status of non-full AOT on x86 mac

[Elijah Taylor]

Hi, saw this thread from a few days ago and wanted to clarify:

On Sun, Oct 31, 2010 at 9:35 AM, Brian Luczkiewicz <brian at sooloos.com>wrote:

>
>>
> And are you sure that __native_client_codegen__ on OSX should influence the
>> file name extension for shared libraries?
>
>
> Since __native_client_codegen__ is the only pre-existing case where we do
> aot on x86 mac, I was being extra careful to make sure that when it's
> defined, the pre-existing behavior is unaffected. I'm unsure whether
> __native_client_codegen__ depends on filename extension and am not in a
> position to test it.
>
> If you happen to know better then that condition is unnecessary.
>
>
It looks like your changes wrt __native_client_codegen__ are good, though I
didn't look at them in fine detail or compile it.  The main thing here is
that __native_client_codegen__ asm output is going to target a completely
different toolchain than the standard mac toolchain.  It's more similar to a
linux gcc/binutils toolchain.

Really the __native_client_codegen__ in AOT is a cross-compiler but we
didn't go through the motions of splitting out the arch specific files
because they're 99% similar to existing codegens.  We also wanted to benefit
from future changes to the arch specific files.  But we needed it to work on
Macs too so that's why we have this define sprinkled in around APPLE
defines.

-Elijah
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.ximian.com/pipermail/mono-devel-list/attachments/20101101/2c84c946/attachment.html