On Fri, Sep 5, 2008 at 7:13 PM, Daniel Dunbar <[EMAIL PROTECTED]> wrote:
> darwin x86-32 is '_', but it might be worth just consulting the gcc source.

I didn't bother trying previously, but since you asked... it's just as
difficult as I figured it would be. I can't sort out how the heck the
target definitions work, and I don't particularly feel like investing
a ton of time into it. It's a lot easier to just check.the compiled
gcc.

> Would it make more sense to have this embedded in TargetIt probably doesn't 
> make much of a difference; it's easy to change if we ever need it.Info? It 
> looks
> like it is something that is always set by GCC but the definition varies
> based on the target.

It probably doesn't make much of a difference; it's easy to change if
we ever need it.

> We may also need this information anyway in other
> places, for example if we were to support -fno-underscores, but I haven't
> investigated this.

If we wanted to support -fno-underscores, we'd have to extend LLVM
somehow... it would have to decorate the symbols differently.

-Eli
_______________________________________________
cfe-commits mailing list
[email protected]
http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits

Reply via email to