I'm pretty sure this is why the bigint unit tests are failing on OSX.
There's a bit of code which deals with position independent code,
storing the address of the function on the stack.
It shouldn't be doing this in a naked function -- there's no stack
frame, so it's randomly corrupting local variables in the calling
function!!!

http://d.puremagic.com/issues/show_bug.cgi?id=4610

I'm guessing this is not too complicated, but since I don't have an OSX box I can't patch it myself. Pretty awful bug, I think it's worth fixing in this release.

http://www.dsource.org/projects/dmd/changeset/612

Don Clugston wrote:
> I'm pretty sure this is why the bigint unit tests are failing on OSX.
> There's a bit of code which deals with position independent code,
> storing the address of the function on the stack.
> It shouldn't be doing this in a naked function -- there's no stack
> frame, so it's randomly corrupting local variables in the calling
> function!!!
>
> http://d.puremagic.com/issues/show_bug.cgi?id=4610
>
> I'm guessing this is not too complicated, but since I don't have an OSX box I can't patch it myself. Pretty awful bug, I think it's worth fixing in this release.
>
>